bibtex_url
stringlengths 41
53
| proceedings
stringlengths 38
50
| bibtext
stringlengths 528
3.02k
| abstract
stringlengths 17
2.35k
| authors
sequencelengths 1
44
| title
stringlengths 18
190
| id
stringlengths 7
19
| arxiv_id
stringlengths 0
10
| GitHub
sequencelengths 1
1
| paper_page
stringclasses 528
values | n_linked_authors
int64 -1
15
| upvotes
int64 -1
77
| num_comments
int64 -1
10
| n_authors
int64 -1
52
| Models
sequencelengths 0
100
| Datasets
sequencelengths 0
15
| Spaces
sequencelengths 0
46
| paper_page_exists_pre_conf
int64 0
1
| type
stringclasses 2
values |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
https://aclanthology.org/2023.findings-emnlp.620.bib | https://aclanthology.org/2023.findings-emnlp.620/ | @inproceedings{shao-etal-2023-enhancing,
title = "Enhancing Retrieval-Augmented Large Language Models with Iterative Retrieval-Generation Synergy",
author = "Shao, Zhihong and
Gong, Yeyun and
Shen, Yelong and
Huang, Minlie and
Duan, Nan and
Chen, Weizhu",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.620",
doi = "10.18653/v1/2023.findings-emnlp.620",
pages = "9248--9274",
abstract = "Retrieval-augmented generation has raise extensive attention as it is promising to address the limitations of large language models including outdated knowledge and hallucinations. However, retrievers struggle to capture relevance, especially for queries with complex information needs. Recent work has proposed to improve relevance modeling by having large language models actively involved in retrieval, i.e., to guide retrieval with generation. In this paper, we show that strong performance can be achieved by a method we call Iter-RetGen, which synergizes retrieval and generation in an iterative manner: a model{'}s response to a task input shows what might be needed to finish the task, and thus can serve as an informative context for retrieving more relevant knowledge which in turn helps generate a better response in another iteration. Compared with recent work which interleaves retrieval with generation when completing a single output, Iter-RetGen processes all retrieved knowledge as a whole and largely preserves the flexibility in generation without structural constraints. We evaluate Iter-RetGen on multi-hop question answering, fact verification, and commonsense reasoning, and show that it can flexibly leverage parametric knowledge and non-parametric knowledge, and is superior to or competitive with state-of-the-art retrieval-augmented baselines while causing fewer overheads of retrieval and generation. We can further improve performance via generation-augmented retrieval adaptation.",
}
| Retrieval-augmented generation has raise extensive attention as it is promising to address the limitations of large language models including outdated knowledge and hallucinations. However, retrievers struggle to capture relevance, especially for queries with complex information needs. Recent work has proposed to improve relevance modeling by having large language models actively involved in retrieval, i.e., to guide retrieval with generation. In this paper, we show that strong performance can be achieved by a method we call Iter-RetGen, which synergizes retrieval and generation in an iterative manner: a model{'}s response to a task input shows what might be needed to finish the task, and thus can serve as an informative context for retrieving more relevant knowledge which in turn helps generate a better response in another iteration. Compared with recent work which interleaves retrieval with generation when completing a single output, Iter-RetGen processes all retrieved knowledge as a whole and largely preserves the flexibility in generation without structural constraints. We evaluate Iter-RetGen on multi-hop question answering, fact verification, and commonsense reasoning, and show that it can flexibly leverage parametric knowledge and non-parametric knowledge, and is superior to or competitive with state-of-the-art retrieval-augmented baselines while causing fewer overheads of retrieval and generation. We can further improve performance via generation-augmented retrieval adaptation. | [
"Shao, Zhihong",
"Gong, Yeyun",
"Shen, Yelong",
"Huang, Minlie",
"Duan, Nan",
"Chen, Weizhu"
] | Enhancing Retrieval-Augmented Large Language Models with Iterative Retrieval-Generation Synergy | findings-emnlp.620 | 2305.15294 | [
""
] | https://huggingface.co/papers/2305.15294 | 1 | 1 | 0 | 6 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.621.bib | https://aclanthology.org/2023.findings-emnlp.621/ | @inproceedings{hua-etal-2023-dynamic,
title = "Dynamic Low-rank Estimation for Transformer-based Language Models",
author = "Hua, Ting and
Li, Xiao and
Gao, Shangqian and
Hsu, Yen-Chang and
Shen, Yilin and
Jin, Hongxia",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.621",
doi = "10.18653/v1/2023.findings-emnlp.621",
pages = "9275--9287",
abstract = "Matrix decomposition methods, such as Singular Value Decomposition (SVD) and its importance-weighted variants, have been widely used for compressing Transformer-based language models. While importance-weighted decomposition methods alleviate the strong assumption of equal importance for each parameter in SVD, they still rely on two fundamental assumptions: 1) unchanged importance distribution during further fine-tuning, 2) equal importance across weight matrices in different layers. Furthermore, these methods necessitate a well-trained task-specific model as the starting point and require additional fine-tuning after compression. In this work, we proposed RankDyna, a matrix decomposition method that enables dynamic rank resource allocation among matrices across different layers during the training process. Starting from a general pre-trained model, RankDyna accomplishes the dual goals of compression and adaptation to the downstream task, all within a single round of fine-tuning. The extensive evaluations demonstrate that RankDyna can outperform current SOTA methods under various parameter budget levels, and the advantage of RankDyna is further enhanced with higher compression rates.",
}
| Matrix decomposition methods, such as Singular Value Decomposition (SVD) and its importance-weighted variants, have been widely used for compressing Transformer-based language models. While importance-weighted decomposition methods alleviate the strong assumption of equal importance for each parameter in SVD, they still rely on two fundamental assumptions: 1) unchanged importance distribution during further fine-tuning, 2) equal importance across weight matrices in different layers. Furthermore, these methods necessitate a well-trained task-specific model as the starting point and require additional fine-tuning after compression. In this work, we proposed RankDyna, a matrix decomposition method that enables dynamic rank resource allocation among matrices across different layers during the training process. Starting from a general pre-trained model, RankDyna accomplishes the dual goals of compression and adaptation to the downstream task, all within a single round of fine-tuning. The extensive evaluations demonstrate that RankDyna can outperform current SOTA methods under various parameter budget levels, and the advantage of RankDyna is further enhanced with higher compression rates. | [
"Hua, Ting",
"Li, Xiao",
"Gao, Shangqian",
"Hsu, Yen-Chang",
"Shen, Yilin",
"Jin, Hongxia"
] | Dynamic Low-rank Estimation for Transformer-based Language Models | findings-emnlp.621 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.622.bib | https://aclanthology.org/2023.findings-emnlp.622/ | @inproceedings{wang-etal-2023-non,
title = "Non-parallel Accent Transfer based on Fine-grained Controllable Accent Modelling",
author = "Wang, Linqin and
Yu, Zhengtao and
Yang, Yuanzhang and
Gao, Shengxiang and
Mao, Cunli and
Huang, Yuxin",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.622",
doi = "10.18653/v1/2023.findings-emnlp.622",
pages = "9288--9298",
abstract = "Existing accent transfer works rely on parallel data or speech recognition models. This paper focuses on the practical application of accent transfer and aims to implement accent transfer using non-parallel datasets. The study has encountered the challenge of speech representation disentanglement and modeling accents. In our accent modeling transfer framework, we manage to solve these problems by two proposed methods. First, we learn the suprasegmental information associated with tone to finely model the accents in terms of tone and rhythm. Second, we propose to use mutual information learning to disentangle the accent features and control the accent of the generated speech during the inference time. Experiments show that the proposed framework attains superior performance to the baseline models in terms of accentedness and audio quality.",
}
| Existing accent transfer works rely on parallel data or speech recognition models. This paper focuses on the practical application of accent transfer and aims to implement accent transfer using non-parallel datasets. The study has encountered the challenge of speech representation disentanglement and modeling accents. In our accent modeling transfer framework, we manage to solve these problems by two proposed methods. First, we learn the suprasegmental information associated with tone to finely model the accents in terms of tone and rhythm. Second, we propose to use mutual information learning to disentangle the accent features and control the accent of the generated speech during the inference time. Experiments show that the proposed framework attains superior performance to the baseline models in terms of accentedness and audio quality. | [
"Wang, Linqin",
"Yu, Zhengtao",
"Yang, Yuanzhang",
"Gao, Shengxiang",
"Mao, Cunli",
"Huang, Yuxin"
] | Non-parallel Accent Transfer based on Fine-grained Controllable Accent Modelling | findings-emnlp.622 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.623.bib | https://aclanthology.org/2023.findings-emnlp.623/ | @inproceedings{xu-etal-2023-compositional,
title = "Compositional Generalization for Data-to-Text Generation",
author = "Xu, Xinnuo and
Titov, Ivan and
Lapata, Mirella",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.623",
doi = "10.18653/v1/2023.findings-emnlp.623",
pages = "9299--9317",
abstract = "Data-to-text generation involves transforming structured data, often represented as predicate-argument tuples, into coherent textual descriptions. Despite recent advances, systems still struggle when confronted with unseen combinations of predicates, producing unfaithful descriptions (e.g.,hallucinations or omissions). We refer to this issue as compositional generalisation, and it encouraged us to create a benchmark for assessing the performance of different approaches on this specific problem. Furthermore, we propose a novel model that addresses compositional generalization by clustering predicates into groups. Our model generates text in a sentence-by-sentence manner, relying on one cluster of predicates at a time. This approach significantly outperforms T5-baselines across all evaluation metrics. Notably, it achieved a 31{\%} improvement over T5 in terms of a metric focused on maintaining faithfulness to the input.",
}
| Data-to-text generation involves transforming structured data, often represented as predicate-argument tuples, into coherent textual descriptions. Despite recent advances, systems still struggle when confronted with unseen combinations of predicates, producing unfaithful descriptions (e.g.,hallucinations or omissions). We refer to this issue as compositional generalisation, and it encouraged us to create a benchmark for assessing the performance of different approaches on this specific problem. Furthermore, we propose a novel model that addresses compositional generalization by clustering predicates into groups. Our model generates text in a sentence-by-sentence manner, relying on one cluster of predicates at a time. This approach significantly outperforms T5-baselines across all evaluation metrics. Notably, it achieved a 31{\%} improvement over T5 in terms of a metric focused on maintaining faithfulness to the input. | [
"Xu, Xinnuo",
"Titov, Ivan",
"Lapata, Mirella"
] | Compositional Generalization for Data-to-Text Generation | findings-emnlp.623 | 2312.02748 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.624.bib | https://aclanthology.org/2023.findings-emnlp.624/ | @inproceedings{hendel-etal-2023-context,
title = "In-Context Learning Creates Task Vectors",
author = "Hendel, Roee and
Geva, Mor and
Globerson, Amir",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.624",
doi = "10.18653/v1/2023.findings-emnlp.624",
pages = "9318--9333",
abstract = "In-context learning (ICL) in Large Language Models (LLMs) has emerged as a powerful new learning paradigm. However, its underlying mechanism is still not well understood. In particular, it is challenging to map it to the {``}standard{'} machine learning framework, where one uses a training set $S$ to find a best-fitting function $f(x)$ in some hypothesis class. Here we make progress on this problem by showing that the functions learned by ICL often have a very simple structure: they correspond to the transformer LLM whose only inputs are the query $x$ and a single {``}task vector{'} calculated from the training set. Thus, ICL can be seen as compressing $S$ into a single task vector $\boldsymbol{\theta}(S)$ and then using this task vector to modulate the transformer to produce the output. We support the above claim via comprehensive experiments across a range of models and tasks.",
}
| In-context learning (ICL) in Large Language Models (LLMs) has emerged as a powerful new learning paradigm. However, its underlying mechanism is still not well understood. In particular, it is challenging to map it to the {``}standard{'} machine learning framework, where one uses a training set $S$ to find a best-fitting function $f(x)$ in some hypothesis class. Here we make progress on this problem by showing that the functions learned by ICL often have a very simple structure: they correspond to the transformer LLM whose only inputs are the query $x$ and a single {``}task vector{'} calculated from the training set. Thus, ICL can be seen as compressing $S$ into a single task vector $\boldsymbol{\theta}(S)$ and then using this task vector to modulate the transformer to produce the output. We support the above claim via comprehensive experiments across a range of models and tasks. | [
"Hendel, Roee",
"Geva, Mor",
"Globerson, Amir"
] | In-Context Learning Creates Task Vectors | findings-emnlp.624 | 2310.15916 | [
"https://github.com/roeehendel/icl_task_vectors"
] | https://huggingface.co/papers/2310.15916 | 3 | 40 | 8 | 3 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.625.bib | https://aclanthology.org/2023.findings-emnlp.625/ | @inproceedings{njoo-etal-2023-talkup,
title = "{T}alk{U}p: Paving the Way for Understanding Empowering Language",
author = "Njoo, Lucille and
Park, Chan and
Stappart, Octavia and
Thielk, Marvin and
Chu, Yi and
Tsvetkov, Yulia",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.625",
doi = "10.18653/v1/2023.findings-emnlp.625",
pages = "9334--9354",
abstract = "Empowering language is important in many real-world contexts, from education to workplace dynamics to healthcare. Though language technologies are growing more prevalent in these contexts, empowerment has seldom been studied in NLP, and moreover, it is inherently challenging to operationalize because of its implicit nature. This work builds from linguistic and social psychology literature to explore what characterizes empowering language. We then crowdsource a novel dataset of Reddit posts labeled for empowerment, reasons why these posts are empowering to readers, and the social relationships between posters and readers. Our preliminary analyses show that this dataset, which we call TalkUp, can be used to train language models that capture empowering and disempowering language. More broadly, TalkUp provides an avenue to explore implication, presuppositions, and how social context influences the meaning of language.",
}
| Empowering language is important in many real-world contexts, from education to workplace dynamics to healthcare. Though language technologies are growing more prevalent in these contexts, empowerment has seldom been studied in NLP, and moreover, it is inherently challenging to operationalize because of its implicit nature. This work builds from linguistic and social psychology literature to explore what characterizes empowering language. We then crowdsource a novel dataset of Reddit posts labeled for empowerment, reasons why these posts are empowering to readers, and the social relationships between posters and readers. Our preliminary analyses show that this dataset, which we call TalkUp, can be used to train language models that capture empowering and disempowering language. More broadly, TalkUp provides an avenue to explore implication, presuppositions, and how social context influences the meaning of language. | [
"Njoo, Lucille",
"Park, Chan",
"Stappart, Octavia",
"Thielk, Marvin",
"Chu, Yi",
"Tsvetkov, Yulia"
] | TalkUp: Paving the Way for Understanding Empowering Language | findings-emnlp.625 | 2305.14326 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.626.bib | https://aclanthology.org/2023.findings-emnlp.626/ | @inproceedings{luo-etal-2023-unifying,
title = "Unifying Text, Tables, and Images for Multimodal Question Answering",
author = "Luo, Haohao and
Shen, Ying and
Deng, Yang",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.626",
doi = "10.18653/v1/2023.findings-emnlp.626",
pages = "9355--9367",
abstract = "Multimodal question answering (MMQA), which aims to derive the answer from multiple knowledge modalities (e.g., text, tables, and images), has received increasing attention due to its board applications. Current approaches to MMQA often rely on single-modal or bi-modal QA models, which limits their ability to effectively integrate information across all modalities and leverage the power of pre-trained language models. To address these limitations, we propose a novel framework called UniMMQA, which unifies three different input modalities into a text-to-text format by employing position-enhanced table linearization and diversified image captioning techniques. Additionally, we enhance cross-modal reasoning by incorporating a multimodal rationale generator, which produces textual descriptions of cross-modal relations for adaptation into the text-to-text generation process. Experimental results on three MMQA benchmark datasets show the superiority of UniMMQA in both supervised and unsupervised settings.",
}
| Multimodal question answering (MMQA), which aims to derive the answer from multiple knowledge modalities (e.g., text, tables, and images), has received increasing attention due to its board applications. Current approaches to MMQA often rely on single-modal or bi-modal QA models, which limits their ability to effectively integrate information across all modalities and leverage the power of pre-trained language models. To address these limitations, we propose a novel framework called UniMMQA, which unifies three different input modalities into a text-to-text format by employing position-enhanced table linearization and diversified image captioning techniques. Additionally, we enhance cross-modal reasoning by incorporating a multimodal rationale generator, which produces textual descriptions of cross-modal relations for adaptation into the text-to-text generation process. Experimental results on three MMQA benchmark datasets show the superiority of UniMMQA in both supervised and unsupervised settings. | [
"Luo, Haohao",
"Shen, Ying",
"Deng, Yang"
] | Unifying Text, Tables, and Images for Multimodal Question Answering | findings-emnlp.626 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.627.bib | https://aclanthology.org/2023.findings-emnlp.627/ | @inproceedings{wada-etal-2023-unsupervised-lexical,
title = "Unsupervised Lexical Simplification with Context Augmentation",
author = "Wada, Takashi and
Baldwin, Timothy and
Lau, Jey",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.627",
doi = "10.18653/v1/2023.findings-emnlp.627",
pages = "9368--9379",
abstract = "We propose a new unsupervised lexical simplification method that uses only monolingual data and pre-trained language models. Given a target word and its context, our method generates substitutes based on the target context and also additional contexts sampled from monolingual data. We conduct experiments in English, Portuguese, and Spanish on the TSAR-2022 shared task, and show that our model substantially outperforms other unsupervised systems across all languages. We also establish a new state-of-the-art by ensembling our model with GPT-3.5. Lastly, we evaluate our model on the SWORDS lexical substitution data set, achieving a state-of-the-art result.",
}
| We propose a new unsupervised lexical simplification method that uses only monolingual data and pre-trained language models. Given a target word and its context, our method generates substitutes based on the target context and also additional contexts sampled from monolingual data. We conduct experiments in English, Portuguese, and Spanish on the TSAR-2022 shared task, and show that our model substantially outperforms other unsupervised systems across all languages. We also establish a new state-of-the-art by ensembling our model with GPT-3.5. Lastly, we evaluate our model on the SWORDS lexical substitution data set, achieving a state-of-the-art result. | [
"Wada, Takashi",
"Baldwin, Timothy",
"Lau, Jey"
] | Unsupervised Lexical Simplification with Context Augmentation | findings-emnlp.627 | 2311.00310 | [
"https://github.com/twadada/lexsub_decontextualised"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.628.bib | https://aclanthology.org/2023.findings-emnlp.628/ | @inproceedings{uthus-etal-2023-mlongt5,
title = "m{L}ong{T}5: A Multilingual and Efficient Text-To-Text Transformer for Longer Sequences",
author = "Uthus, David and
Ontanon, Santiago and
Ainslie, Joshua and
Guo, Mandy",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.628",
doi = "10.18653/v1/2023.findings-emnlp.628",
pages = "9380--9386",
abstract = "We present our work on developing a multilingual, efficient text-to-text transformer that is suitable for handling long inputs. This model, called mLongT5, builds upon the architecture of LongT5, while leveraging the multilingual datasets used for pretraining mT5 and the pretraining tasks of UL2. We evaluate this model on a variety of multilingual summarization and question-answering tasks, and the results show stronger performance for mLongT5 when compared to existing multilingual models such as mBART or M-BERT.",
}
| We present our work on developing a multilingual, efficient text-to-text transformer that is suitable for handling long inputs. This model, called mLongT5, builds upon the architecture of LongT5, while leveraging the multilingual datasets used for pretraining mT5 and the pretraining tasks of UL2. We evaluate this model on a variety of multilingual summarization and question-answering tasks, and the results show stronger performance for mLongT5 when compared to existing multilingual models such as mBART or M-BERT. | [
"Uthus, David",
"Ontanon, Santiago",
"Ainslie, Joshua",
"Guo, M",
"y"
] | mLongT5: A Multilingual and Efficient Text-To-Text Transformer for Longer Sequences | findings-emnlp.628 | 2305.11129 | [
"https://github.com/google-research/longt5"
] | https://huggingface.co/papers/2305.11129 | 0 | 2 | 1 | 4 | [
"agemagician/mlong-t5-tglobal-xl",
"agemagician/mlong-t5-tglobal-base",
"agemagician/mlong-t5-tglobal-large"
] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.629.bib | https://aclanthology.org/2023.findings-emnlp.629/ | @inproceedings{lee-hwang-2023-multilingual,
title = "Multilingual Lottery Tickets to Pretrain Language Models",
author = "Lee, Jaeseong and
Hwang, Seung-won",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.629",
doi = "10.18653/v1/2023.findings-emnlp.629",
pages = "9387--9398",
abstract = "The curse of multilinguality in training multilingual pretrained language models (mPLMs) refers to the negative interference between languages, especially when the capacity is limited. While increasing the capacity may appear intuitive for overcoming this curse, it negatively affects both training and inference costs. Our distinction is pursuing the competing goals of reducing negative interference, while keeping capacity per each language more or less the same. Specifically, we first scale the model to reduce interference, then search for a per-language subnetwork, or a lottery ticket, with comparable performance to the full model. According to lottery ticket hypothesis, this scale-then-find-ticket approach alleviates interfering signals as in the scaled model, but redistributes parameters to keep the parameters reduced. Finally, to avoid the cost of multiple retraining for searching multilingual tickets, we explore zero-shot neural architecture search (NAS) methods. We investigate the most appropriate zero-shot NAS method to find multilingual tickets. Our proposed multilingual tickets reduce the inference cost of models for each languages, while boosting the performances. The ticket search cost is negligible and tickets found qualitatively preserve linguistic similarity. Our code is publicly available.",
}
| The curse of multilinguality in training multilingual pretrained language models (mPLMs) refers to the negative interference between languages, especially when the capacity is limited. While increasing the capacity may appear intuitive for overcoming this curse, it negatively affects both training and inference costs. Our distinction is pursuing the competing goals of reducing negative interference, while keeping capacity per each language more or less the same. Specifically, we first scale the model to reduce interference, then search for a per-language subnetwork, or a lottery ticket, with comparable performance to the full model. According to lottery ticket hypothesis, this scale-then-find-ticket approach alleviates interfering signals as in the scaled model, but redistributes parameters to keep the parameters reduced. Finally, to avoid the cost of multiple retraining for searching multilingual tickets, we explore zero-shot neural architecture search (NAS) methods. We investigate the most appropriate zero-shot NAS method to find multilingual tickets. Our proposed multilingual tickets reduce the inference cost of models for each languages, while boosting the performances. The ticket search cost is negligible and tickets found qualitatively preserve linguistic similarity. Our code is publicly available. | [
"Lee, Jaeseong",
"Hwang, Seung-won"
] | Multilingual Lottery Tickets to Pretrain Language Models | findings-emnlp.629 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.630.bib | https://aclanthology.org/2023.findings-emnlp.630/ | @inproceedings{jiang-yin-2023-target,
title = "Target-Aware Spatio-Temporal Reasoning via Answering Questions in Dynamic Audio-Visual Scenarios",
author = "Jiang, Yuanyuan and
Yin, Jianqin",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.630",
doi = "10.18653/v1/2023.findings-emnlp.630",
pages = "9399--9409",
abstract = "Audio-visual question answering (AVQA) is a challenging task that requires multistep spatio-temporal reasoning over multimodal contexts. Recent works rely on elaborate target-agnostic parsing of audio-visual scenes for spatial grounding while mistreating audio and video as separate entities for temporal grounding. This paper proposes a new target-aware joint spatio-temporal grounding network for AVQA. It consists of two key components: the target-aware spatial grounding module (TSG) and the single-stream joint audio-visual temporal grounding module (JTG). The TSG can focus on audio-visual cues relevant to the query subject by utilizing explicit semantics from the question. Unlike previous two-stream temporal grounding modules that required an additional audio-visual fusion module, JTG incorporates audio-visual fusion and question-aware temporal grounding into one module with a simpler single-stream architecture. The temporal synchronization between audio and video in the JTG is facilitated by our proposed cross-modal synchrony loss (CSL). Extensive experiments verified the effectiveness of our proposed method over existing state-of-the-art methods.",
}
| Audio-visual question answering (AVQA) is a challenging task that requires multistep spatio-temporal reasoning over multimodal contexts. Recent works rely on elaborate target-agnostic parsing of audio-visual scenes for spatial grounding while mistreating audio and video as separate entities for temporal grounding. This paper proposes a new target-aware joint spatio-temporal grounding network for AVQA. It consists of two key components: the target-aware spatial grounding module (TSG) and the single-stream joint audio-visual temporal grounding module (JTG). The TSG can focus on audio-visual cues relevant to the query subject by utilizing explicit semantics from the question. Unlike previous two-stream temporal grounding modules that required an additional audio-visual fusion module, JTG incorporates audio-visual fusion and question-aware temporal grounding into one module with a simpler single-stream architecture. The temporal synchronization between audio and video in the JTG is facilitated by our proposed cross-modal synchrony loss (CSL). Extensive experiments verified the effectiveness of our proposed method over existing state-of-the-art methods. | [
"Jiang, Yuanyuan",
"Yin, Jianqin"
] | Target-Aware Spatio-Temporal Reasoning via Answering Questions in Dynamic Audio-Visual Scenarios | findings-emnlp.630 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.631.bib | https://aclanthology.org/2023.findings-emnlp.631/ | @inproceedings{kim-etal-2023-kg,
title = "{KG}-{GPT}: A General Framework for Reasoning on Knowledge Graphs Using Large Language Models",
author = "Kim, Jiho and
Kwon, Yeonsu and
Jo, Yohan and
Choi, Edward",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.631",
doi = "10.18653/v1/2023.findings-emnlp.631",
pages = "9410--9421",
abstract = "While large language models (LLMs) have made considerable advancements in understanding and generating unstructured text, their application in structured data remains underexplored. Particularly, using LLMs for complex reasoning tasks on knowledge graphs (KGs) remains largely untouched. To address this, we propose KG-GPT, a multi-purpose framework leveraging LLMs for tasks employing KGs. KG-GPT comprises three steps: Sentence Segmentation, Graph Retrieval, and Inference, each aimed at partitioning sentences, retrieving relevant graph components, and deriving logical conclusions, respectively. We evaluate KG-GPT using KG-based fact verification and KGQA benchmarks, with the model showing competitive and robust performance, even outperforming several fully-supervised models. Our work, therefore, marks a significant step in unifying structured and unstructured data processing within the realm of LLMs.",
}
| While large language models (LLMs) have made considerable advancements in understanding and generating unstructured text, their application in structured data remains underexplored. Particularly, using LLMs for complex reasoning tasks on knowledge graphs (KGs) remains largely untouched. To address this, we propose KG-GPT, a multi-purpose framework leveraging LLMs for tasks employing KGs. KG-GPT comprises three steps: Sentence Segmentation, Graph Retrieval, and Inference, each aimed at partitioning sentences, retrieving relevant graph components, and deriving logical conclusions, respectively. We evaluate KG-GPT using KG-based fact verification and KGQA benchmarks, with the model showing competitive and robust performance, even outperforming several fully-supervised models. Our work, therefore, marks a significant step in unifying structured and unstructured data processing within the realm of LLMs. | [
"Kim, Jiho",
"Kwon, Yeonsu",
"Jo, Yohan",
"Choi, Edward"
] | KG-GPT: A General Framework for Reasoning on Knowledge Graphs Using Large Language Models | findings-emnlp.631 | 2310.11220 | [
"https://github.com/jiho283/kg-gpt"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.632.bib | https://aclanthology.org/2023.findings-emnlp.632/ | @inproceedings{foroutan-etal-2023-breaking,
title = "Breaking the Language Barrier: Improving Cross-Lingual Reasoning with Structured Self-Attention",
author = "Foroutan, Negar and
Banaei, Mohammadreza and
Aberer, Karl and
Bosselut, Antoine",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.632",
doi = "10.18653/v1/2023.findings-emnlp.632",
pages = "9422--9442",
abstract = "In this work, we study whether multilingual language models (MultiLMs) can transfer logical reasoning abilities to other languages when they are fine-tuned for reasoning in a different language. We evaluate the cross-lingual reasoning abilities of MultiLMs in two schemes: (1) where the language of the context and the question remain the same in the new languages that are tested (i.e., the reasoning is still monolingual, but the model must transfer the learned reasoning ability across languages), and (2) where the language of the context and the question is different (which we term code-switched reasoning). On two logical reasoning datasets, RuleTaker and LeapOfThought, we demonstrate that although MultiLMs can transfer reasoning ability across languages in a monolingual setting, they struggle to transfer reasoning abilities in a code-switched setting. Following this observation, we propose a novel attention mechanism that uses a dedicated set of parameters to encourage cross-lingual attention in code-switched sequences, which improves the reasoning performance by up to 14{\%} and 4{\%} on the RuleTaker and LeapOfThought datasets, respectively.",
}
| In this work, we study whether multilingual language models (MultiLMs) can transfer logical reasoning abilities to other languages when they are fine-tuned for reasoning in a different language. We evaluate the cross-lingual reasoning abilities of MultiLMs in two schemes: (1) where the language of the context and the question remain the same in the new languages that are tested (i.e., the reasoning is still monolingual, but the model must transfer the learned reasoning ability across languages), and (2) where the language of the context and the question is different (which we term code-switched reasoning). On two logical reasoning datasets, RuleTaker and LeapOfThought, we demonstrate that although MultiLMs can transfer reasoning ability across languages in a monolingual setting, they struggle to transfer reasoning abilities in a code-switched setting. Following this observation, we propose a novel attention mechanism that uses a dedicated set of parameters to encourage cross-lingual attention in code-switched sequences, which improves the reasoning performance by up to 14{\%} and 4{\%} on the RuleTaker and LeapOfThought datasets, respectively. | [
"Foroutan, Negar",
"Banaei, Mohammadreza",
"Aberer, Karl",
"Bosselut, Antoine"
] | Breaking the Language Barrier: Improving Cross-Lingual Reasoning with Structured Self-Attention | findings-emnlp.632 | 2310.15258 | [
"https://github.com/negar-foroutan/multilingual-code-switched-reasoning"
] | https://huggingface.co/papers/2310.15258 | 0 | 2 | 0 | 4 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.633.bib | https://aclanthology.org/2023.findings-emnlp.633/ | @inproceedings{zhang-etal-2023-citb,
title = "{CITB}: A Benchmark for Continual Instruction Tuning",
author = "Zhang, Zihan and
Fang, Meng and
Chen, Ling and
Namazi-Rad, Mohammad-Reza",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.633",
doi = "10.18653/v1/2023.findings-emnlp.633",
pages = "9443--9455",
abstract = "Continual learning (CL) is a paradigm that aims to replicate the human ability to learn and accumulate knowledge continually without forgetting previous knowledge and transferring it to new tasks. Recent instruction tuning (IT) involves fine-tuning models to make them more adaptable to solving NLP tasks in general. However, it is still uncertain how instruction tuning works in the context of CL tasks. This challenging yet practical problem is formulated as Continual Instruction Tuning (CIT). In this work, we establish a CIT benchmark consisting of learning and evaluation protocols. We curate two long dialogue task streams of different types, InstrDialog and InstrDialog++, to study various CL methods systematically. Our experiments show that existing CL methods do not effectively leverage the rich natural language instructions, and fine-tuning an instruction-tuned model sequentially can yield similar or better results. We further explore different aspects that might affect the learning of CIT. We hope this benchmark will facilitate more research in this direction.",
}
| Continual learning (CL) is a paradigm that aims to replicate the human ability to learn and accumulate knowledge continually without forgetting previous knowledge and transferring it to new tasks. Recent instruction tuning (IT) involves fine-tuning models to make them more adaptable to solving NLP tasks in general. However, it is still uncertain how instruction tuning works in the context of CL tasks. This challenging yet practical problem is formulated as Continual Instruction Tuning (CIT). In this work, we establish a CIT benchmark consisting of learning and evaluation protocols. We curate two long dialogue task streams of different types, InstrDialog and InstrDialog++, to study various CL methods systematically. Our experiments show that existing CL methods do not effectively leverage the rich natural language instructions, and fine-tuning an instruction-tuned model sequentially can yield similar or better results. We further explore different aspects that might affect the learning of CIT. We hope this benchmark will facilitate more research in this direction. | [
"Zhang, Zihan",
"Fang, Meng",
"Chen, Ling",
"Namazi-Rad, Mohammad-Reza"
] | CITB: A Benchmark for Continual Instruction Tuning | findings-emnlp.633 | 2310.14510 | [
"https://github.com/hyintell/citb"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.634.bib | https://aclanthology.org/2023.findings-emnlp.634/ | @inproceedings{li-etal-2023-mixture,
title = "Mixture-of-Linguistic-Experts Adapters for Improving and Interpreting Pre-trained Language Models",
author = "Li, Raymond and
Murray, Gabriel and
Carenini, Giuseppe",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.634",
doi = "10.18653/v1/2023.findings-emnlp.634",
pages = "9456--9469",
abstract = "In this work, we propose a method that combines two popular research areas by injecting linguistic structures into pre-trained language models in the parameter-efficient fine-tuning (PEFT) setting. In our approach, parallel adapter modules encoding different linguistic structures are combined using a novel Mixture-of-Linguistic-Experts architecture, where Gumbel-Softmax gates are used to determine the importance of these modules at each layer of the model. To reduce the number of parameters, we first train the model for a fixed small number of steps before pruning the experts based on their important scores. Our experiment results with three different pre-trained models show that our approach can outperform state-of-the-art PEFT methods with a comparable number of parameters. In addition, we provide additional analysis to examine the experts selected by each model at each layer to provide insights for future studies.",
}
| In this work, we propose a method that combines two popular research areas by injecting linguistic structures into pre-trained language models in the parameter-efficient fine-tuning (PEFT) setting. In our approach, parallel adapter modules encoding different linguistic structures are combined using a novel Mixture-of-Linguistic-Experts architecture, where Gumbel-Softmax gates are used to determine the importance of these modules at each layer of the model. To reduce the number of parameters, we first train the model for a fixed small number of steps before pruning the experts based on their important scores. Our experiment results with three different pre-trained models show that our approach can outperform state-of-the-art PEFT methods with a comparable number of parameters. In addition, we provide additional analysis to examine the experts selected by each model at each layer to provide insights for future studies. | [
"Li, Raymond",
"Murray, Gabriel",
"Carenini, Giuseppe"
] | Mixture-of-Linguistic-Experts Adapters for Improving and Interpreting Pre-trained Language Models | findings-emnlp.634 | 2310.16240 | [
""
] | https://huggingface.co/papers/2310.16240 | 0 | 1 | 0 | 3 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.635.bib | https://aclanthology.org/2023.findings-emnlp.635/ | @inproceedings{li-etal-2023-towards-better,
title = "Towards Better Representations for Multi-Label Text Classification with Multi-granularity Information",
author = "Li, Fangfang and
Su, Puzhen and
Duan, Junwen and
Xiao, Weidong",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.635",
doi = "10.18653/v1/2023.findings-emnlp.635",
pages = "9470--9480",
abstract = "Multi-label text classification (MLTC) aims to assign multiple labels to a given text. Previous works have focused on text representation learning and label correlations modeling using pre-trained language models (PLMs). However, studies have shown that PLMs generate word frequency-oriented text representations, causing texts with different labels to be closely distributed in a narrow region, which is difficult to classify. To address this, we present a novel framework $\textbf{CL}$($\underline{C}$ontrastive $\underline{L}$earning)-$\textbf{MIL}$ ($\underline{M}$ulti-granularity $\underline{I}$nformation $\underline{L}$earning) to refine the text representation for MLTC task. We first use contrastive learning to generate uniform initial text representation and incorporate label frequency implicitly. Then, we design a multi-task learning module to integrate multi-granularity (diverse text-labels correlations, label-label relations and label frequency) information into text representations, enhancing their discriminative ability. Experimental results demonstrate the complementarity of the modules in CL-MIL, improving the quality of text representations and yielding stable and competitive improvements for MLTC.",
}
| Multi-label text classification (MLTC) aims to assign multiple labels to a given text. Previous works have focused on text representation learning and label correlations modeling using pre-trained language models (PLMs). However, studies have shown that PLMs generate word frequency-oriented text representations, causing texts with different labels to be closely distributed in a narrow region, which is difficult to classify. To address this, we present a novel framework $\textbf{CL}$($\underline{C}$ontrastive $\underline{L}$earning)-$\textbf{MIL}$ ($\underline{M}$ulti-granularity $\underline{I}$nformation $\underline{L}$earning) to refine the text representation for MLTC task. We first use contrastive learning to generate uniform initial text representation and incorporate label frequency implicitly. Then, we design a multi-task learning module to integrate multi-granularity (diverse text-labels correlations, label-label relations and label frequency) information into text representations, enhancing their discriminative ability. Experimental results demonstrate the complementarity of the modules in CL-MIL, improving the quality of text representations and yielding stable and competitive improvements for MLTC. | [
"Li, Fangfang",
"Su, Puzhen",
"Duan, Junwen",
"Xiao, Weidong"
] | Towards Better Representations for Multi-Label Text Classification with Multi-granularity Information | findings-emnlp.635 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.636.bib | https://aclanthology.org/2023.findings-emnlp.636/ | @inproceedings{song-etal-2023-pcmid,
title = "{PCMID}: Multi-Intent Detection through Supervised Prototypical Contrastive Learning",
author = "Song, Yurun and
Zhao, Junchen and
Koehler, Spencer and
Abdullah, Amir and
Harris, Ian",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.636",
doi = "10.18653/v1/2023.findings-emnlp.636",
pages = "9481--9495",
abstract = "Intent detection is a major task in Natural Language Understanding (NLU) and is the component of dialogue systems for interpreting users{'} intentions based on their utterances. Many works have explored detecting intents by assuming that each utterance represents only a single intent. Such systems have achieved very good results; however, intent detection is a far more challenging task in typical real-world scenarios, where each user utterance can be highly complex and express multiple intents. Therefore, in this paper, we propose PCMID, a novel Multi-Intent Detection framework enabled by Prototypical Contrastive Learning under a supervised setting. The PCMID model can learn multiple semantic representations of a given user utterance under the context of different intent labels in an optimized semantic space. Our experiments show that PCMID achieves the current state-of-the-art performance on both multiple public benchmark datasets and a private real-world dataset for the multi-intent detection task.",
}
| Intent detection is a major task in Natural Language Understanding (NLU) and is the component of dialogue systems for interpreting users{'} intentions based on their utterances. Many works have explored detecting intents by assuming that each utterance represents only a single intent. Such systems have achieved very good results; however, intent detection is a far more challenging task in typical real-world scenarios, where each user utterance can be highly complex and express multiple intents. Therefore, in this paper, we propose PCMID, a novel Multi-Intent Detection framework enabled by Prototypical Contrastive Learning under a supervised setting. The PCMID model can learn multiple semantic representations of a given user utterance under the context of different intent labels in an optimized semantic space. Our experiments show that PCMID achieves the current state-of-the-art performance on both multiple public benchmark datasets and a private real-world dataset for the multi-intent detection task. | [
"Song, Yurun",
"Zhao, Junchen",
"Koehler, Spencer",
"Abdullah, Amir",
"Harris, Ian"
] | PCMID: Multi-Intent Detection through Supervised Prototypical Contrastive Learning | findings-emnlp.636 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.637.bib | https://aclanthology.org/2023.findings-emnlp.637/ | @inproceedings{cheng-etal-2023-gpt,
title = "Is {GPT}-4 a Good Data Analyst?",
author = "Cheng, Liying and
Li, Xingxuan and
Bing, Lidong",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.637",
doi = "10.18653/v1/2023.findings-emnlp.637",
pages = "9496--9514",
abstract = "As large language models (LLMs) have demonstrated their powerful capabilities in plenty of domains and tasks, including context understanding, code generation, language generation, data storytelling, etc., many data analysts may raise concerns if their jobs will be replaced by artificial intelligence (AI). This controversial topic has drawn great attention in public. However, we are still at a stage of divergent opinions without any definitive conclusion. Motivated by this, we raise the research question of {``}is GPT-4 a good data analyst?{''} in this work and aim to answer it by conducting head-to-head comparative studies. In detail, we regard GPT-4 as a data analyst to perform end-to-end data analysis with databases from a wide range of domains. We propose a framework to tackle the problems by carefully designing the prompts for GPT-4 to conduct experiments. We also design several task-specific evaluation metrics to systematically compare the performance between several professional human data analysts and GPT-4. Experimental results show that GPT-4 can achieve comparable performance to humans. We also provide in-depth discussions about our results to shed light on further studies before reaching the conclusion that GPT-4 can replace data analysts.",
}
| As large language models (LLMs) have demonstrated their powerful capabilities in plenty of domains and tasks, including context understanding, code generation, language generation, data storytelling, etc., many data analysts may raise concerns if their jobs will be replaced by artificial intelligence (AI). This controversial topic has drawn great attention in public. However, we are still at a stage of divergent opinions without any definitive conclusion. Motivated by this, we raise the research question of {``}is GPT-4 a good data analyst?{''} in this work and aim to answer it by conducting head-to-head comparative studies. In detail, we regard GPT-4 as a data analyst to perform end-to-end data analysis with databases from a wide range of domains. We propose a framework to tackle the problems by carefully designing the prompts for GPT-4 to conduct experiments. We also design several task-specific evaluation metrics to systematically compare the performance between several professional human data analysts and GPT-4. Experimental results show that GPT-4 can achieve comparable performance to humans. We also provide in-depth discussions about our results to shed light on further studies before reaching the conclusion that GPT-4 can replace data analysts. | [
"Cheng, Liying",
"Li, Xingxuan",
"Bing, Lidong"
] | Is GPT-4 a Good Data Analyst? | findings-emnlp.637 | 2305.15038 | [
"https://github.com/damo-nlp-sg/gpt4-as-dataanalyst"
] | https://huggingface.co/papers/2305.15038 | 2 | 5 | 2 | 3 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.638.bib | https://aclanthology.org/2023.findings-emnlp.638/ | @inproceedings{qiao-etal-2023-diffusionret,
title = "{D}iffusion{R}et: Diffusion-Enhanced Generative Retriever using Constrained Decoding",
author = "Qiao, Shanbao and
Liu, Xuebing and
Na, Seung-Hoon",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.638",
doi = "10.18653/v1/2023.findings-emnlp.638",
pages = "9515--9529",
abstract = "Generative retrieval, which maps from a query to its relevant document identifiers (docids), has recently emerged as a new information retrieval (IR) paradigm, however, having suffered from 1) the $\textit{lack of the intermediate reasoning step}$, caused by the manner of merely using a query to perform the hierarchical classification, and 2) the $\textit{pretrain-finetune discrepancy}$, which comes from the use of the artificial symbols of docids. To address these limitations, we propose the novel approach of using the document generation from a query as an intermediate step before the retrieval, thus presenting $\underline{diffusion}$-enhanced generative $\underline{ret}$rieval ($\textbf{DiffusionRet}$), which consists of two processing steps: 1) the $\textit{diffusion-based document generation}$, which employs the sequence-to-sequence diffusion model to produce a pseudo document sample from a query, being expected to semantically close to a relevant document; 2) $\textit{N-gram-based generative retrieval}$, which use another sequence-to-sequence model to generate n-grams that appear in the collection index for linking a generated sample to an original document. Experiment results on MS MARCO and Natural Questions dataset show that the proposed DiffusionRet significantly outperforms all the existing generative retrieval methods and leads to the state-of-the-art performances, even with much smaller number of parameters.",
}
| Generative retrieval, which maps from a query to its relevant document identifiers (docids), has recently emerged as a new information retrieval (IR) paradigm, however, having suffered from 1) the $\textit{lack of the intermediate reasoning step}$, caused by the manner of merely using a query to perform the hierarchical classification, and 2) the $\textit{pretrain-finetune discrepancy}$, which comes from the use of the artificial symbols of docids. To address these limitations, we propose the novel approach of using the document generation from a query as an intermediate step before the retrieval, thus presenting $\underline{diffusion}$-enhanced generative $\underline{ret}$rieval ($\textbf{DiffusionRet}$), which consists of two processing steps: 1) the $\textit{diffusion-based document generation}$, which employs the sequence-to-sequence diffusion model to produce a pseudo document sample from a query, being expected to semantically close to a relevant document; 2) $\textit{N-gram-based generative retrieval}$, which use another sequence-to-sequence model to generate n-grams that appear in the collection index for linking a generated sample to an original document. Experiment results on MS MARCO and Natural Questions dataset show that the proposed DiffusionRet significantly outperforms all the existing generative retrieval methods and leads to the state-of-the-art performances, even with much smaller number of parameters. | [
"Qiao, Shanbao",
"Liu, Xuebing",
"Na, Seung-Hoon"
] | DiffusionRet: Diffusion-Enhanced Generative Retriever using Constrained Decoding | findings-emnlp.638 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.639.bib | https://aclanthology.org/2023.findings-emnlp.639/ | @inproceedings{fu-etal-2023-estimating,
title = "Estimating Large Language Model Capabilities without Labeled Test Data",
author = "Fu, Harvey and
Ye, Qinyuan and
Xu, Albert and
Ren, Xiang and
Jia, Robin",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.639",
doi = "10.18653/v1/2023.findings-emnlp.639",
pages = "9530--9546",
abstract = "Large Language Models (LLMs) have exhibited an impressive ability to perform in-context learning (ICL) from only a few examples, but the success of ICL varies widely from task to task. Thus, it is important to quickly determine whether ICL is applicable to a new task, but directly evaluating ICL accuracy can be expensive in situations where test data is expensive to annotate{---}the exact situations where ICL is most appealing. In this paper, we propose the task of ICL accuracy estimation, in which we predict the accuracy of an LLM when doing in-context learning on a new task given only unlabeled test data for that task. To perform ICL accuracy estimation, we propose a method that trains a meta-model using LLM confidence scores as features. We compare our method to several strong accuracy estimation baselines on a new benchmark that covers 4 LLMs and 3 task collections. The meta-model improves over all baselines across 7 out of 12 settings and achieves the same estimation performance as directly evaluating on 40 collected labeled test examples per task. At the same time, no existing approach provides an accurate and reliable ICL accuracy estimation in every setting, highlighting the need for better ways to measure the uncertainty of LLM predictions.",
}
| Large Language Models (LLMs) have exhibited an impressive ability to perform in-context learning (ICL) from only a few examples, but the success of ICL varies widely from task to task. Thus, it is important to quickly determine whether ICL is applicable to a new task, but directly evaluating ICL accuracy can be expensive in situations where test data is expensive to annotate{---}the exact situations where ICL is most appealing. In this paper, we propose the task of ICL accuracy estimation, in which we predict the accuracy of an LLM when doing in-context learning on a new task given only unlabeled test data for that task. To perform ICL accuracy estimation, we propose a method that trains a meta-model using LLM confidence scores as features. We compare our method to several strong accuracy estimation baselines on a new benchmark that covers 4 LLMs and 3 task collections. The meta-model improves over all baselines across 7 out of 12 settings and achieves the same estimation performance as directly evaluating on 40 collected labeled test examples per task. At the same time, no existing approach provides an accurate and reliable ICL accuracy estimation in every setting, highlighting the need for better ways to measure the uncertainty of LLM predictions. | [
"Fu, Harvey",
"Ye, Qinyuan",
"Xu, Albert",
"Ren, Xiang",
"Jia, Robin"
] | Estimating Large Language Model Capabilities without Labeled Test Data | findings-emnlp.639 | 2305.14802 | [
"https://github.com/harvey-fin/icl-estimate"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.640.bib | https://aclanthology.org/2023.findings-emnlp.640/ | @inproceedings{broscoteanu-ionescu-2023-novel,
title = "A Novel Contrastive Learning Method for Clickbait Detection on {R}o{C}li{C}o: A {R}omanian Clickbait Corpus of News Articles",
author = "Broscoteanu, Daria and
Ionescu, Radu",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.640",
doi = "10.18653/v1/2023.findings-emnlp.640",
pages = "9547--9555",
abstract = "To increase revenue, news websites often resort to using deceptive news titles, luring users into clicking on the title and reading the full news. Clickbait detection is the task that aims to automatically detect this form of false advertisement and avoid wasting the precious time of online users. Despite the importance of the task, to the best of our knowledge, there is no publicly available clickbait corpus for the Romanian language. To this end, we introduce a novel Romanian Clickbait Corpus (RoCliCo) comprising 8,313 news samples which are manually annotated with clickbait and non-clickbait labels. Furthermore, we conduct experiments with four machine learning methods, ranging from handcrafted models to recurrent and transformer-based neural networks, to establish a line-up of competitive baselines. We also carry out experiments with a weighted voting ensemble. Among the considered baselines, we propose a novel BERT-based contrastive learning model that learns to encode news titles and contents into a deep metric space such that titles and contents of non-clickbait news have high cosine similarity, while titles and contents of clickbait news have low cosine similarity. Our data set and code to reproduce the baselines are publicly available for download at https://github.com/dariabroscoteanu/RoCliCo.",
}
| To increase revenue, news websites often resort to using deceptive news titles, luring users into clicking on the title and reading the full news. Clickbait detection is the task that aims to automatically detect this form of false advertisement and avoid wasting the precious time of online users. Despite the importance of the task, to the best of our knowledge, there is no publicly available clickbait corpus for the Romanian language. To this end, we introduce a novel Romanian Clickbait Corpus (RoCliCo) comprising 8,313 news samples which are manually annotated with clickbait and non-clickbait labels. Furthermore, we conduct experiments with four machine learning methods, ranging from handcrafted models to recurrent and transformer-based neural networks, to establish a line-up of competitive baselines. We also carry out experiments with a weighted voting ensemble. Among the considered baselines, we propose a novel BERT-based contrastive learning model that learns to encode news titles and contents into a deep metric space such that titles and contents of non-clickbait news have high cosine similarity, while titles and contents of clickbait news have low cosine similarity. Our data set and code to reproduce the baselines are publicly available for download at https://github.com/dariabroscoteanu/RoCliCo. | [
"Broscoteanu, Daria",
"Ionescu, Radu"
] | A Novel Contrastive Learning Method for Clickbait Detection on RoCliCo: A Romanian Clickbait Corpus of News Articles | findings-emnlp.640 | 2310.06540 | [
"https://github.com/dariabroscoteanu/roclico"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.641.bib | https://aclanthology.org/2023.findings-emnlp.641/ | @inproceedings{wang-etal-2023-large,
title = "Large Language Models as Source Planner for Personalized Knowledge-grounded Dialogues",
author = "Wang, Hongru and
Hu, Minda and
Deng, Yang and
Wang, Rui and
Mi, Fei and
Wang, Weichao and
Wang, Yasheng and
Kwan, Wai-Chung and
King, Irwin and
Wong, Kam-Fai",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.641",
doi = "10.18653/v1/2023.findings-emnlp.641",
pages = "9556--9569",
abstract = "Open-domain dialogue system usually requires different sources of knowledge to generate more informative and evidential responses. However, existing knowledge-grounded dialogue systems either focus on a single knowledge source or overlook the dependency between multiple sources of knowledge, which may result in generating inconsistent or even paradoxical responses. To incorporate multiple knowledge sources and dependencies between them, we propose SAFARI, a novel framework that leverages the exceptional capabilities of large language models (LLMs) in planning, understanding, and incorporating under both supervised and unsupervised settings. Specifically, SAFARI decouples the knowledge grounding into multiple sources and response generation, which allows easy extension to various knowledge sources including the possibility of not using any sources. To study the problem, we construct a personalized knowledge-grounded dialogue dataset Knowledge Behind Persona (KBP), which is the first to consider the dependency between persona and implicit knowledge. Experimental results on the KBP dataset demonstrate that the SAFARI framework can effectively produce persona-consistent and knowledge-enhanced responses.",
}
| Open-domain dialogue system usually requires different sources of knowledge to generate more informative and evidential responses. However, existing knowledge-grounded dialogue systems either focus on a single knowledge source or overlook the dependency between multiple sources of knowledge, which may result in generating inconsistent or even paradoxical responses. To incorporate multiple knowledge sources and dependencies between them, we propose SAFARI, a novel framework that leverages the exceptional capabilities of large language models (LLMs) in planning, understanding, and incorporating under both supervised and unsupervised settings. Specifically, SAFARI decouples the knowledge grounding into multiple sources and response generation, which allows easy extension to various knowledge sources including the possibility of not using any sources. To study the problem, we construct a personalized knowledge-grounded dialogue dataset Knowledge Behind Persona (KBP), which is the first to consider the dependency between persona and implicit knowledge. Experimental results on the KBP dataset demonstrate that the SAFARI framework can effectively produce persona-consistent and knowledge-enhanced responses. | [
"Wang, Hongru",
"Hu, Minda",
"Deng, Yang",
"Wang, Rui",
"Mi, Fei",
"Wang, Weichao",
"Wang, Yasheng",
"Kwan, Wai-Chung",
"King, Irwin",
"Wong, Kam-Fai"
] | Large Language Models as Source Planner for Personalized Knowledge-grounded Dialogues | findings-emnlp.641 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.642.bib | https://aclanthology.org/2023.findings-emnlp.642/ | @inproceedings{costa-jussa-etal-2023-toxicity,
title = "Toxicity in Multilingual Machine Translation at Scale",
author = "Costa-juss{\`a}, Marta and
Smith, Eric and
Ropers, Christophe and
Licht, Daniel and
Maillard, Jean and
Ferrando, Javier and
Escolano, Carlos",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.642",
doi = "10.18653/v1/2023.findings-emnlp.642",
pages = "9570--9586",
abstract = "Machine Translation systems can produce different types of errors, some of which are characterized as critical or catastrophic due to the specific negative impact that they can have on users. In this paper we focus on one type of critical error: added toxicity. We evaluate and analyze added toxicity when translating a large evaluation dataset (HOLISTICBIAS, over 472k sentences, covering 13 demographic axes) from English into 164 languages. An automatic toxicity evaluation shows that added toxicity across languages varies from 0{\%} to 5{\%}. The output languages with the most added toxicity tend to be low-resource ones, and the demographic axes with the most added toxicity include sexual orientation, gender and sex, and ability. We also perform human evaluation on a subset of 8 translation directions, confirming the prevalence of true added toxicity. We use a measurement of the amount of source contribution to the translation, where a low source contribution implies hallucination, to interpret what causes toxicity. Making use of the input attributions allows us to explain toxicity, because the source contributions significantly correlate with toxicity for 84{\%} of languages studied. Given our findings, our recommendations to reduce added toxicity are to curate training data to avoid mistranslations, mitigate hallucination and check unstable translations.",
}
| Machine Translation systems can produce different types of errors, some of which are characterized as critical or catastrophic due to the specific negative impact that they can have on users. In this paper we focus on one type of critical error: added toxicity. We evaluate and analyze added toxicity when translating a large evaluation dataset (HOLISTICBIAS, over 472k sentences, covering 13 demographic axes) from English into 164 languages. An automatic toxicity evaluation shows that added toxicity across languages varies from 0{\%} to 5{\%}. The output languages with the most added toxicity tend to be low-resource ones, and the demographic axes with the most added toxicity include sexual orientation, gender and sex, and ability. We also perform human evaluation on a subset of 8 translation directions, confirming the prevalence of true added toxicity. We use a measurement of the amount of source contribution to the translation, where a low source contribution implies hallucination, to interpret what causes toxicity. Making use of the input attributions allows us to explain toxicity, because the source contributions significantly correlate with toxicity for 84{\%} of languages studied. Given our findings, our recommendations to reduce added toxicity are to curate training data to avoid mistranslations, mitigate hallucination and check unstable translations. | [
"Costa-juss{\\`a}, Marta",
"Smith, Eric",
"Ropers, Christophe",
"Licht, Daniel",
"Maillard, Jean",
"Ferr",
"o, Javier",
"Escolano, Carlos"
] | Toxicity in Multilingual Machine Translation at Scale | findings-emnlp.642 | 2210.03070 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.643.bib | https://aclanthology.org/2023.findings-emnlp.643/ | @inproceedings{liu-etal-2023-conversational,
title = "Conversational Recommender System and Large Language Model Are Made for Each Other in {E}-commerce Pre-sales Dialogue",
author = "Liu, Yuanxing and
Zhang, Weinan and
Chen, Yifan and
Zhang, Yuchi and
Bai, Haopeng and
Feng, Fan and
Cui, Hengbin and
Li, Yongbin and
Che, Wanxiang",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.643",
doi = "10.18653/v1/2023.findings-emnlp.643",
pages = "9587--9605",
abstract = "E-commerce pre-sales dialogue aims to understand and elicit user needs and preferences for the items they are seeking so as to provide appropriate recommendations. Conversational recommender systems (CRSs) learn user representation and provide accurate recommendations based on dialogue context, but rely on external knowledge. Large language models (LLMs) generate responses that mimic pre-sales dialogues after fine-tuning, but lack domain-specific knowledge for accurate recommendations. Intuitively, the strengths of LLM and CRS in E-commerce pre-sales dialogues are complementary, yet no previous work has explored this. This paper investigates the effectiveness of combining LLM and CRS in E-commerce pre-sales dialogues, proposing two collaboration methods: CRS assisting LLM and LLM assisting CRS. We conduct extensive experiments on a real-world dataset of E-commerce pre-sales dialogues. We analyze the impact of two collaborative approaches with two CRSs and two LLMs on four tasks of E-commerce pre-sales dialogue. We find that collaborations between CRS and LLM can be very effective in some cases.",
}
| E-commerce pre-sales dialogue aims to understand and elicit user needs and preferences for the items they are seeking so as to provide appropriate recommendations. Conversational recommender systems (CRSs) learn user representation and provide accurate recommendations based on dialogue context, but rely on external knowledge. Large language models (LLMs) generate responses that mimic pre-sales dialogues after fine-tuning, but lack domain-specific knowledge for accurate recommendations. Intuitively, the strengths of LLM and CRS in E-commerce pre-sales dialogues are complementary, yet no previous work has explored this. This paper investigates the effectiveness of combining LLM and CRS in E-commerce pre-sales dialogues, proposing two collaboration methods: CRS assisting LLM and LLM assisting CRS. We conduct extensive experiments on a real-world dataset of E-commerce pre-sales dialogues. We analyze the impact of two collaborative approaches with two CRSs and two LLMs on four tasks of E-commerce pre-sales dialogue. We find that collaborations between CRS and LLM can be very effective in some cases. | [
"Liu, Yuanxing",
"Zhang, Weinan",
"Chen, Yifan",
"Zhang, Yuchi",
"Bai, Haopeng",
"Feng, Fan",
"Cui, Hengbin",
"Li, Yongbin",
"Che, Wanxiang"
] | Conversational Recommender System and Large Language Model Are Made for Each Other in E-commerce Pre-sales Dialogue | findings-emnlp.643 | 2310.14626 | [
"https://github.com/leeeeoliu/llm-crs"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.644.bib | https://aclanthology.org/2023.findings-emnlp.644/ | @inproceedings{geng-etal-2023-vip5,
title = "{VIP}5: Towards Multimodal Foundation Models for Recommendation",
author = "Geng, Shijie and
Tan, Juntao and
Liu, Shuchang and
Fu, Zuohui and
Zhang, Yongfeng",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.644",
doi = "10.18653/v1/2023.findings-emnlp.644",
pages = "9606--9620",
abstract = "Computer Vision (CV), Natural Language Processing (NLP), and Recommender Systems (RecSys) are three prominent AI applications that have traditionally developed independently, resulting in disparate modeling and engineering methodologies. This has impeded the ability for these fields to directly benefit from each other{'}s advancements. With the recent development of foundation models, large language models have emerged as a potential general-purpose interface for unifying different modalities and problem formulations. In light of this, we propose the development of a multimodal foundation model (MFM) considering visual, textual, and personalization modalities under the P5 recommendation paradigm, thus named VIP5 (Visual P5), to unify various modalities and recommendation tasks. This will enable the processing of multiple modalities in a shared architecture for improved recommendations. To achieve this, we introduce multimodal personalized prompts to accommodate multiple modalities under a shared format. Additionally, we propose a parameter-efficient training method for foundation models, which involves freezing the P5 backbone and fine-tuning lightweight adapters, resulting in improved recommendation performance and increased efficiency in terms of training time and memory usage. Code and data of VIP5 are available at https://github.com/jeykigung/VIP5.",
}
| Computer Vision (CV), Natural Language Processing (NLP), and Recommender Systems (RecSys) are three prominent AI applications that have traditionally developed independently, resulting in disparate modeling and engineering methodologies. This has impeded the ability for these fields to directly benefit from each other{'}s advancements. With the recent development of foundation models, large language models have emerged as a potential general-purpose interface for unifying different modalities and problem formulations. In light of this, we propose the development of a multimodal foundation model (MFM) considering visual, textual, and personalization modalities under the P5 recommendation paradigm, thus named VIP5 (Visual P5), to unify various modalities and recommendation tasks. This will enable the processing of multiple modalities in a shared architecture for improved recommendations. To achieve this, we introduce multimodal personalized prompts to accommodate multiple modalities under a shared format. Additionally, we propose a parameter-efficient training method for foundation models, which involves freezing the P5 backbone and fine-tuning lightweight adapters, resulting in improved recommendation performance and increased efficiency in terms of training time and memory usage. Code and data of VIP5 are available at https://github.com/jeykigung/VIP5. | [
"Geng, Shijie",
"Tan, Juntao",
"Liu, Shuchang",
"Fu, Zuohui",
"Zhang, Yongfeng"
] | VIP5: Towards Multimodal Foundation Models for Recommendation | findings-emnlp.644 | 2305.14302 | [
"https://github.com/jeykigung/vip5"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.645.bib | https://aclanthology.org/2023.findings-emnlp.645/ | @inproceedings{nguyen-etal-2023-spectral,
title = "A Spectral Viewpoint on Continual Relation Extraction",
author = "Nguyen, Huy and
Nguyen, Chien and
Ngo, Linh and
Luu, Anh and
Nguyen, Thien",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.645",
doi = "10.18653/v1/2023.findings-emnlp.645",
pages = "9621--9629",
abstract = "Continual Relation Extraction (CRE) aims to continuously train a model to learn new relations while preserving its ability on previously learned relations. Similar to other continual learning problems, in CRE, models experience representation shift, where learned deep space changes in the continual learning process, which leads to the downgrade in the performance of the old tasks. In this work, we will provide an insight into this phenomenon under the spectral viewpoint. Our key argument is that, for each class shape, if its eigenvectors (or spectral components) do not change much, the shape is well-preserved. We then conduct a spectral experiment and show that, for the shape of each class, the eigenvectors with larger eigenvalue are more preserved after learning new tasks which means these vectors are good at keeping class shapes. Based on this analysis, we propose a simple yet effective class-wise regularization that improve the eigenvalues in the representation learning. We observe that our proposed regularization leads to an increase in the eigenvalues. Extensive experiments on two benchmark datasets, FewRel and TACRED, show the effectiveness of our proposed method with significant improvement in performance compared to the state-of-the-art models. Further analyses also verify our hypothesis that larger eigenvalues lead to better performance and vice versa.",
}
| Continual Relation Extraction (CRE) aims to continuously train a model to learn new relations while preserving its ability on previously learned relations. Similar to other continual learning problems, in CRE, models experience representation shift, where learned deep space changes in the continual learning process, which leads to the downgrade in the performance of the old tasks. In this work, we will provide an insight into this phenomenon under the spectral viewpoint. Our key argument is that, for each class shape, if its eigenvectors (or spectral components) do not change much, the shape is well-preserved. We then conduct a spectral experiment and show that, for the shape of each class, the eigenvectors with larger eigenvalue are more preserved after learning new tasks which means these vectors are good at keeping class shapes. Based on this analysis, we propose a simple yet effective class-wise regularization that improve the eigenvalues in the representation learning. We observe that our proposed regularization leads to an increase in the eigenvalues. Extensive experiments on two benchmark datasets, FewRel and TACRED, show the effectiveness of our proposed method with significant improvement in performance compared to the state-of-the-art models. Further analyses also verify our hypothesis that larger eigenvalues lead to better performance and vice versa. | [
"Nguyen, Huy",
"Nguyen, Chien",
"Ngo, Linh",
"Luu, Anh",
"Nguyen, Thien"
] | A Spectral Viewpoint on Continual Relation Extraction | findings-emnlp.645 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.646.bib | https://aclanthology.org/2023.findings-emnlp.646/ | @inproceedings{chakrabarty-etal-2023-learning,
title = "Learning to Follow Object-Centric Image Editing Instructions Faithfully",
author = "Chakrabarty, Tuhin and
Singh, Kanishk and
Saakyan, Arkadiy and
Muresan, Smaranda",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.646",
doi = "10.18653/v1/2023.findings-emnlp.646",
pages = "9630--9646",
abstract = "Natural language instructions are a powerful interface for editing the outputs of text-to-image diffusion models. However, several challenges need to be addressed: 1) underspecification (the need to model the implicit meaning of instructions) 2) grounding (the need to localize where the edit has to be performed), 3) faithfulness (the need to preserve the elements of the image not affected by the edit instruction). Current approaches focusing on image editing with natural language instructions rely on automatically generated paired data, which, as shown in our investigation, is noisy and sometimes nonsensical, exacerbating the above issues. Building on recent advances in segmentation, Chain-of-Thought prompting, and visual question answering, we significantly improve the quality of the paired data. In addition, we enhance the supervision signal by highlighting parts of the image that need to be changed by the instruction. The model fine-tuned on the improved data is capable of performing fine-grained object-centric edits better than state-of-the-art baselines, mitigating the problems outlined above, as shown by automatic and human evaluations. Moreover, our model is capable of generalizing to domains unseen during training, such as visual metaphors.",
}
| Natural language instructions are a powerful interface for editing the outputs of text-to-image diffusion models. However, several challenges need to be addressed: 1) underspecification (the need to model the implicit meaning of instructions) 2) grounding (the need to localize where the edit has to be performed), 3) faithfulness (the need to preserve the elements of the image not affected by the edit instruction). Current approaches focusing on image editing with natural language instructions rely on automatically generated paired data, which, as shown in our investigation, is noisy and sometimes nonsensical, exacerbating the above issues. Building on recent advances in segmentation, Chain-of-Thought prompting, and visual question answering, we significantly improve the quality of the paired data. In addition, we enhance the supervision signal by highlighting parts of the image that need to be changed by the instruction. The model fine-tuned on the improved data is capable of performing fine-grained object-centric edits better than state-of-the-art baselines, mitigating the problems outlined above, as shown by automatic and human evaluations. Moreover, our model is capable of generalizing to domains unseen during training, such as visual metaphors. | [
"Chakrabarty, Tuhin",
"Singh, Kanishk",
"Saakyan, Arkadiy",
"Muresan, Smar",
"a"
] | Learning to Follow Object-Centric Image Editing Instructions Faithfully | findings-emnlp.646 | 2310.19145 | [
"https://github.com/tuhinjubcse/faithfuledits_emnlp2023"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.647.bib | https://aclanthology.org/2023.findings-emnlp.647/ | @inproceedings{gretz-etal-2023-zero,
title = "Zero-shot Topical Text Classification with {LLM}s - an Experimental Study",
author = "Gretz, Shai and
Halfon, Alon and
Shnayderman, Ilya and
Toledo-Ronen, Orith and
Spector, Artem and
Dankin, Lena and
Katsis, Yannis and
Arviv, Ofir and
Katz, Yoav and
Slonim, Noam and
Ein-Dor, Liat",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.647",
doi = "10.18653/v1/2023.findings-emnlp.647",
pages = "9647--9676",
abstract = "Topical Text Classification (TTC) is an ancient, yet timely research area in natural language processing, with many practical applications. The recent dramatic advancements in large LMs raise the question of how well these models can perform in this task in a zero-shot scenario. Here, we share a first comprehensive study, comparing the zero-shot performance of a variety of LMs over TTC23, a large benchmark collection of 23 publicly available TTC datasets, covering a wide range of domains and styles. In addition, we leverage this new TTC benchmark to create LMs that are specialized in TTC, by fine-tuning these LMs over a subset of the datasets and evaluating their performance over the remaining, held-out datasets. We show that the TTC-specialized LMs obtain the top performance on our benchmark, by a significant margin. Our code and model are made available for the community. We hope that the results presented in this work will serve as a useful guide for practitioners interested in topical text classification.",
}
| Topical Text Classification (TTC) is an ancient, yet timely research area in natural language processing, with many practical applications. The recent dramatic advancements in large LMs raise the question of how well these models can perform in this task in a zero-shot scenario. Here, we share a first comprehensive study, comparing the zero-shot performance of a variety of LMs over TTC23, a large benchmark collection of 23 publicly available TTC datasets, covering a wide range of domains and styles. In addition, we leverage this new TTC benchmark to create LMs that are specialized in TTC, by fine-tuning these LMs over a subset of the datasets and evaluating their performance over the remaining, held-out datasets. We show that the TTC-specialized LMs obtain the top performance on our benchmark, by a significant margin. Our code and model are made available for the community. We hope that the results presented in this work will serve as a useful guide for practitioners interested in topical text classification. | [
"Gretz, Shai",
"Halfon, Alon",
"Shnayderman, Ilya",
"Toledo-Ronen, Orith",
"Spector, Artem",
"Dankin, Lena",
"Katsis, Yannis",
"Arviv, Ofir",
"Katz, Yoav",
"Slonim, Noam",
"Ein-Dor, Liat"
] | Zero-shot Topical Text Classification with LLMs - an Experimental Study | findings-emnlp.647 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.648.bib | https://aclanthology.org/2023.findings-emnlp.648/ | @inproceedings{wan-etal-2023-personalized,
title = "Are Personalized Stochastic Parrots More Dangerous? Evaluating Persona Biases in Dialogue Systems",
author = "Wan, Yixin and
Zhao, Jieyu and
Chadha, Aman and
Peng, Nanyun and
Chang, Kai-Wei",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.648",
doi = "10.18653/v1/2023.findings-emnlp.648",
pages = "9677--9705",
abstract = "Recent advancements in Large Language Models empower them to follow freeform instructions, including imitating generic or specific demographic personas in conversations. We define generic personas to represent demographic groups, such as {``}an Asian person{''}, whereas specific personas may take the form of specific popular Asian names like {``}Yumi{''}. While the adoption of personas enriches user experiences by making dialogue systems more engaging and approachable, it also casts a shadow of potential risk by exacerbating social biases within model responses, thereby causing societal harm through interactions with users. In this paper, we systematically study {``}persona biases{''}, which we define to be the sensitivity of dialogue models{'} harmful behaviors contingent upon the personas they adopt. We categorize persona biases into biases in harmful expression and harmful agreement, and establish a comprehensive evaluation framework to measure persona biases in five aspects: Offensiveness, Toxic Continuation, Regard, Stereotype Agreement, and Toxic Agreement. Additionally, we propose to investigate persona biases by experimenting with UNIVERSALPERSONA, a systematically constructed persona dataset encompassing various types of both generic and specific model personas. Through benchmarking on four different models- including Blender, ChatGPT, Alpaca, and Vicuna- our study uncovers significant persona biases in dialogue systems. Our findings also underscore the pressing need to revisit the use of personas in dialogue agents to ensure safe application.",
}
| Recent advancements in Large Language Models empower them to follow freeform instructions, including imitating generic or specific demographic personas in conversations. We define generic personas to represent demographic groups, such as {``}an Asian person{''}, whereas specific personas may take the form of specific popular Asian names like {``}Yumi{''}. While the adoption of personas enriches user experiences by making dialogue systems more engaging and approachable, it also casts a shadow of potential risk by exacerbating social biases within model responses, thereby causing societal harm through interactions with users. In this paper, we systematically study {``}persona biases{''}, which we define to be the sensitivity of dialogue models{'} harmful behaviors contingent upon the personas they adopt. We categorize persona biases into biases in harmful expression and harmful agreement, and establish a comprehensive evaluation framework to measure persona biases in five aspects: Offensiveness, Toxic Continuation, Regard, Stereotype Agreement, and Toxic Agreement. Additionally, we propose to investigate persona biases by experimenting with UNIVERSALPERSONA, a systematically constructed persona dataset encompassing various types of both generic and specific model personas. Through benchmarking on four different models- including Blender, ChatGPT, Alpaca, and Vicuna- our study uncovers significant persona biases in dialogue systems. Our findings also underscore the pressing need to revisit the use of personas in dialogue agents to ensure safe application. | [
"Wan, Yixin",
"Zhao, Jieyu",
"Chadha, Aman",
"Peng, Nanyun",
"Chang, Kai-Wei"
] | Are Personalized Stochastic Parrots More Dangerous? Evaluating Persona Biases in Dialogue Systems | findings-emnlp.648 | 2310.05280 | [
"https://github.com/uclanlp/persona-biases"
] | https://huggingface.co/papers/2310.05280 | 2 | 1 | 0 | 5 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.649.bib | https://aclanthology.org/2023.findings-emnlp.649/ | @inproceedings{zhang-etal-2023-black,
title = "A Black-Box Attack on Code Models via Representation Nearest Neighbor Search",
author = "Zhang, Jie and
Ma, Wei and
Hu, Qiang and
Liu, Shangqing and
Xie, Xiaofei and
Le Traon, Yves and
Liu, Yang",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.649",
doi = "10.18653/v1/2023.findings-emnlp.649",
pages = "9706--9716",
abstract = "Existing methods for generating adversarial code examples face several challenges: limted availability of substitute variables, high verification costs for these substitutes, and the creation of adversarial samples with noticeable perturbations. To address these concerns, our proposed approach, RNNS, uses a search seed based on historical attacks to find potential adversarial substitutes. Rather than directly using the discrete substitutes, they are mapped to a continuous vector space using a pre-trained variable name encoder. Based on the vector representation, RNNS predicts and selects better substitutes for attacks. We evaluated the performance of RNNS across six coding tasks encompassing three programming languages: Java, Python, and C. We employed three pre-trained code models (CodeBERT, GraphCodeBERT, and CodeT5) that resulted in a cumulative of 18 victim models. The results demonstrate that RNNS outperforms baselines in terms of ASR and QT. Furthermore, the perturbation of adversarial examples introduced by RNNS is smaller compared to the baselines in terms of the number of replaced variables and the change in variable length. Lastly, our experiments indicate that RNNS is efficient in attacking defended models and can be employed for adversarial training.",
}
| Existing methods for generating adversarial code examples face several challenges: limted availability of substitute variables, high verification costs for these substitutes, and the creation of adversarial samples with noticeable perturbations. To address these concerns, our proposed approach, RNNS, uses a search seed based on historical attacks to find potential adversarial substitutes. Rather than directly using the discrete substitutes, they are mapped to a continuous vector space using a pre-trained variable name encoder. Based on the vector representation, RNNS predicts and selects better substitutes for attacks. We evaluated the performance of RNNS across six coding tasks encompassing three programming languages: Java, Python, and C. We employed three pre-trained code models (CodeBERT, GraphCodeBERT, and CodeT5) that resulted in a cumulative of 18 victim models. The results demonstrate that RNNS outperforms baselines in terms of ASR and QT. Furthermore, the perturbation of adversarial examples introduced by RNNS is smaller compared to the baselines in terms of the number of replaced variables and the change in variable length. Lastly, our experiments indicate that RNNS is efficient in attacking defended models and can be employed for adversarial training. | [
"Zhang, Jie",
"Ma, Wei",
"Hu, Qiang",
"Liu, Shangqing",
"Xie, Xiaofei",
"Le Traon, Yves",
"Liu, Yang"
] | A Black-Box Attack on Code Models via Representation Nearest Neighbor Search | findings-emnlp.649 | 2305.05896 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.650.bib | https://aclanthology.org/2023.findings-emnlp.650/ | @inproceedings{zhang-etal-2023-well,
title = "How Well Do Text Embedding Models Understand Syntax?",
author = "Zhang, Yan and
Feng, Zhaopeng and
Teng, Zhiyang and
Liu, Zuozhu and
Li, Haizhou",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.650",
doi = "10.18653/v1/2023.findings-emnlp.650",
pages = "9717--9728",
abstract = "Text embedding models have significantly contributed to advancements in natural language processing by adeptly capturing semantic properties of textual data. However, the ability of these models to generalize across a wide range of syntactic contexts remains under-explored. In this paper, we first develop an evaluation set, named SR, to scrutinize the capability for syntax understanding of text embedding models from two crucial syntactic aspects: Structural heuristics, and Relational understanding among concepts, as revealed by the performance gaps in previous studies. Our findings reveal that existing text embedding models have not sufficiently addressed these syntactic understanding challenges, and such ineffectiveness becomes even more apparent when evaluated against existing benchmark datasets. Furthermore, we conduct rigorous analysis to unearth factors that lead to such limitations and examine why previous evaluations fail to detect such ineffectiveness. Lastly, we propose strategies to augment the generalization ability of text embedding models in diverse syntactic scenarios. This study serves to highlight the hurdles associated with syntactic generalization and provides pragmatic guidance for boosting model performance across varied syntactic contexts.",
}
| Text embedding models have significantly contributed to advancements in natural language processing by adeptly capturing semantic properties of textual data. However, the ability of these models to generalize across a wide range of syntactic contexts remains under-explored. In this paper, we first develop an evaluation set, named SR, to scrutinize the capability for syntax understanding of text embedding models from two crucial syntactic aspects: Structural heuristics, and Relational understanding among concepts, as revealed by the performance gaps in previous studies. Our findings reveal that existing text embedding models have not sufficiently addressed these syntactic understanding challenges, and such ineffectiveness becomes even more apparent when evaluated against existing benchmark datasets. Furthermore, we conduct rigorous analysis to unearth factors that lead to such limitations and examine why previous evaluations fail to detect such ineffectiveness. Lastly, we propose strategies to augment the generalization ability of text embedding models in diverse syntactic scenarios. This study serves to highlight the hurdles associated with syntactic generalization and provides pragmatic guidance for boosting model performance across varied syntactic contexts. | [
"Zhang, Yan",
"Feng, Zhaopeng",
"Teng, Zhiyang",
"Liu, Zuozhu",
"Li, Haizhou"
] | How Well Do Text Embedding Models Understand Syntax? | findings-emnlp.650 | 2311.07996 | [
"https://github.com/fzp0424/sr"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.651.bib | https://aclanthology.org/2023.findings-emnlp.651/ | @inproceedings{surana-etal-2023-cassi,
title = "{CASSI}: Contextual and Semantic Structure-based Interpolation Augmentation for Low-Resource {NER}",
author = "Surana, Tanmay and
Ho, Thi-Nga and
Tun, Kyaw and
Chng, Eng Siong",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.651",
doi = "10.18653/v1/2023.findings-emnlp.651",
pages = "9729--9742",
abstract = "While text augmentation methods have been successful in improving performance in the low-resource setting, they suffer from annotation corruption for a token-level task like NER. Moreover, existing methods cannot reliably add context diversity to the dataset, which has been shown to be crucial for low-resource NER. In this work, we propose Contextual and Semantic Structure-based Interpolation (CASSI), a novel augmentation scheme that generates high-quality contextually diverse augmentations while avoiding annotation corruption by structurally combining a pair of semantically similar sentences to generate a new sentence while maintaining semantic correctness and fluency. To accomplish this, we generate candidate augmentations by performing multiple dependency parsing-based exchanges in a pair of semantically similar sentences that are filtered via scoring with a pretrained Masked Language Model and a metric to promote specificity. Experiments show that CASSI consistently outperforms existing methods at multiple low resource levels, in multiple languages, and for noisy and clean text.",
}
| While text augmentation methods have been successful in improving performance in the low-resource setting, they suffer from annotation corruption for a token-level task like NER. Moreover, existing methods cannot reliably add context diversity to the dataset, which has been shown to be crucial for low-resource NER. In this work, we propose Contextual and Semantic Structure-based Interpolation (CASSI), a novel augmentation scheme that generates high-quality contextually diverse augmentations while avoiding annotation corruption by structurally combining a pair of semantically similar sentences to generate a new sentence while maintaining semantic correctness and fluency. To accomplish this, we generate candidate augmentations by performing multiple dependency parsing-based exchanges in a pair of semantically similar sentences that are filtered via scoring with a pretrained Masked Language Model and a metric to promote specificity. Experiments show that CASSI consistently outperforms existing methods at multiple low resource levels, in multiple languages, and for noisy and clean text. | [
"Surana, Tanmay",
"Ho, Thi-Nga",
"Tun, Kyaw",
"Chng, Eng Siong"
] | CASSI: Contextual and Semantic Structure-based Interpolation Augmentation for Low-Resource NER | findings-emnlp.651 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.652.bib | https://aclanthology.org/2023.findings-emnlp.652/ | @inproceedings{wang-etal-2023-newton,
title = "{NEWTON}: Are Large Language Models Capable of Physical Reasoning?",
author = "Wang, Yi and
Duan, Jiafei and
Fox, Dieter and
Srinivasa, Siddhartha",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.652",
doi = "10.18653/v1/2023.findings-emnlp.652",
pages = "9743--9758",
abstract = "Large Language Models (LLMs), through their contextualized representations, have been empirically proven to encapsulate syntactic, semantic, word sense, and common-sense knowledge. However, there has been limited exploration of their physical reasoning abilities, specifically concerning the crucial attributes for comprehending everyday objects. To address this gap, we introduce NEWTON, a repository and benchmark for evaluating the physics reasoning skills of LLMs. Further, to enable domain-specific adaptation of this benchmark, we present a pipeline to enable researchers to generate a variant of this benchmark that has been customized to the objects and attributes relevant for their application. The NEWTON repository comprises a collection of 2800 object-attribute pairs, providing the foundation for generating infinite-scale assessment templates. The NEWTON benchmark consists of 160K QA questions, curated using the NEWTON repository to investigate the physical reasoning capabilities of several mainstream language models across foundational, explicit, and implicit reasoning tasks. Through extensive empirical analysis, our results highlight the capabilities of LLMs for physical reasoning. We find that LLMs like GPT-4 demonstrate strong reasoning capabilities in scenario-based tasks but exhibit less consistency in object-attribute reasoning compared to humans (50{\%} vs. 84{\%}). Furthermore, the NEWTON platform demonstrates its potential for evaluating and enhancing language models, paving the way for their integration into physically grounded settings, such as robotic manipulation. Project site: https://newtonreasoning.github.io",
}
| Large Language Models (LLMs), through their contextualized representations, have been empirically proven to encapsulate syntactic, semantic, word sense, and common-sense knowledge. However, there has been limited exploration of their physical reasoning abilities, specifically concerning the crucial attributes for comprehending everyday objects. To address this gap, we introduce NEWTON, a repository and benchmark for evaluating the physics reasoning skills of LLMs. Further, to enable domain-specific adaptation of this benchmark, we present a pipeline to enable researchers to generate a variant of this benchmark that has been customized to the objects and attributes relevant for their application. The NEWTON repository comprises a collection of 2800 object-attribute pairs, providing the foundation for generating infinite-scale assessment templates. The NEWTON benchmark consists of 160K QA questions, curated using the NEWTON repository to investigate the physical reasoning capabilities of several mainstream language models across foundational, explicit, and implicit reasoning tasks. Through extensive empirical analysis, our results highlight the capabilities of LLMs for physical reasoning. We find that LLMs like GPT-4 demonstrate strong reasoning capabilities in scenario-based tasks but exhibit less consistency in object-attribute reasoning compared to humans (50{\%} vs. 84{\%}). Furthermore, the NEWTON platform demonstrates its potential for evaluating and enhancing language models, paving the way for their integration into physically grounded settings, such as robotic manipulation. Project site: https://newtonreasoning.github.io | [
"Wang, Yi",
"Duan, Jiafei",
"Fox, Dieter",
"Srinivasa, Siddhartha"
] | NEWTON: Are Large Language Models Capable of Physical Reasoning? | findings-emnlp.652 | 2310.07018 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.653.bib | https://aclanthology.org/2023.findings-emnlp.653/ | @inproceedings{mun-etal-2023-beyond,
title = "Beyond Denouncing Hate: Strategies for Countering Implied Biases and Stereotypes in Language",
author = "Mun, Jimin and
Allaway, Emily and
Yerukola, Akhila and
Vianna, Laura and
Leslie, Sarah-Jane and
Sap, Maarten",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.653",
doi = "10.18653/v1/2023.findings-emnlp.653",
pages = "9759--9777",
abstract = "Counterspeech, i.e., responses to counteract potential harms of hateful speech, has become an increasingly popular solution to address online hate speech without censorship. However, properly countering hateful language requires countering and dispelling the underlying inaccurate stereotypes implied by such language. In this work, we draw from psychology and philosophy literature to craft six psychologically inspired strategies to challenge the underlying stereotypical implications of hateful language. We first examine the convincingness of each of these strategies through a user study, and then compare their usages in both human- and machine-generated counterspeech datasets. Our results show that human-written counterspeech uses countering strategies that are more specific to the implied stereotype (e.g., counter examples to the stereotype, external factors about the stereotype{'}s origins), whereas machine-generated counterspeech uses less specific strategies (e.g., generally denouncing the hatefulness of speech). Furthermore, machine generated counterspeech often employs strategies that humans deem less convincing compared to human-produced counterspeech. Our findings point to the importance of accounting for the underlying stereotypical implications of speech when generating counterspeech and for better machine reasoning about anti-stereotypical examples.",
}
| Counterspeech, i.e., responses to counteract potential harms of hateful speech, has become an increasingly popular solution to address online hate speech without censorship. However, properly countering hateful language requires countering and dispelling the underlying inaccurate stereotypes implied by such language. In this work, we draw from psychology and philosophy literature to craft six psychologically inspired strategies to challenge the underlying stereotypical implications of hateful language. We first examine the convincingness of each of these strategies through a user study, and then compare their usages in both human- and machine-generated counterspeech datasets. Our results show that human-written counterspeech uses countering strategies that are more specific to the implied stereotype (e.g., counter examples to the stereotype, external factors about the stereotype{'}s origins), whereas machine-generated counterspeech uses less specific strategies (e.g., generally denouncing the hatefulness of speech). Furthermore, machine generated counterspeech often employs strategies that humans deem less convincing compared to human-produced counterspeech. Our findings point to the importance of accounting for the underlying stereotypical implications of speech when generating counterspeech and for better machine reasoning about anti-stereotypical examples. | [
"Mun, Jimin",
"Allaway, Emily",
"Yerukola, Akhila",
"Vianna, Laura",
"Leslie, Sarah-Jane",
"Sap, Maarten"
] | Beyond Denouncing Hate: Strategies for Countering Implied Biases and Stereotypes in Language | findings-emnlp.653 | 2311.00161 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.654.bib | https://aclanthology.org/2023.findings-emnlp.654/ | @inproceedings{zhu-etal-2023-calibration,
title = "On the Calibration of Large Language Models and Alignment",
author = "Zhu, Chiwei and
Xu, Benfeng and
Wang, Quan and
Zhang, Yongdong and
Mao, Zhendong",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.654",
doi = "10.18653/v1/2023.findings-emnlp.654",
pages = "9778--9795",
abstract = "As large language models attract increasing attention and find widespread application, concurrent challenges of reliability also arise at the same time. Confidence calibration, an effective analysis method for gauging the reliability of deep models, serves as a crucial tool for assessing and improving their reliability. However, such investigation has been comparatively underexplored. In this work, we conduct a systematic examination of the calibration of aligned language models throughout the entire construction process, including pretraining and alignment training. At each stage, we investigate how different training settings, such as parameter scales and training data, affect model calibration. To thoroughly assess model calibration, we evaluate models on three most concerned aspects: generation, factuality and understanding. Our work sheds light on whether popular LLMs are well-calibrated and how the training process influences model calibration.",
}
| As large language models attract increasing attention and find widespread application, concurrent challenges of reliability also arise at the same time. Confidence calibration, an effective analysis method for gauging the reliability of deep models, serves as a crucial tool for assessing and improving their reliability. However, such investigation has been comparatively underexplored. In this work, we conduct a systematic examination of the calibration of aligned language models throughout the entire construction process, including pretraining and alignment training. At each stage, we investigate how different training settings, such as parameter scales and training data, affect model calibration. To thoroughly assess model calibration, we evaluate models on three most concerned aspects: generation, factuality and understanding. Our work sheds light on whether popular LLMs are well-calibrated and how the training process influences model calibration. | [
"Zhu, Chiwei",
"Xu, Benfeng",
"Wang, Quan",
"Zhang, Yongdong",
"Mao, Zhendong"
] | On the Calibration of Large Language Models and Alignment | findings-emnlp.654 | 2311.13240 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.655.bib | https://aclanthology.org/2023.findings-emnlp.655/ | @inproceedings{liu-etal-2023-tcra,
title = "{TCRA}-{LLM}: Token Compression Retrieval Augmented Large Language Model for Inference Cost Reduction",
author = "Liu, Junyi and
Li, Liangzhi and
Xiang, Tong and
Wang, Bowen and
Qian, Yiming",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.655",
doi = "10.18653/v1/2023.findings-emnlp.655",
pages = "9796--9810",
abstract = "Since ChatGPT released its API for public use, the number of applications built on top of commercial large language models (LLMs) increase exponentially. One popular usage of such models is leveraging its in-context learning ability and generating responses given user queries leveraging knowledge obtained by retrieval augmentation. One problem of deploying commercial retrieval-augmented LLMs is the cost due to the additionally retrieved context that largely increases the input token size of the LLMs. To mitigate this, we propose a token compression scheme that includes two methods: summarization compression and semantic compression. The first method applies a T5-based model that is fine-tuned by datasets generated using self-instruct containing samples with varying lengths and reduce token size by doing summarization. The second method further compresses the token size by removing words with lower impact on the semantic. In order to adequately evaluate the effectiveness of the proposed methods, we propose and utilize a dataset called Food-Recommendation DB (FRDB) focusing on food recommendation for women around pregnancy period or infants. Our summarization compression can reduce 65{\%} of the retrieval token size with further 0.3{\%} improvement on the accuracy; semantic compression provides a more flexible way to trade-off the token size with performance, for which we can reduce the token size by 20{\%} with only 1.6{\%} of accuracy drop.",
}
| Since ChatGPT released its API for public use, the number of applications built on top of commercial large language models (LLMs) increase exponentially. One popular usage of such models is leveraging its in-context learning ability and generating responses given user queries leveraging knowledge obtained by retrieval augmentation. One problem of deploying commercial retrieval-augmented LLMs is the cost due to the additionally retrieved context that largely increases the input token size of the LLMs. To mitigate this, we propose a token compression scheme that includes two methods: summarization compression and semantic compression. The first method applies a T5-based model that is fine-tuned by datasets generated using self-instruct containing samples with varying lengths and reduce token size by doing summarization. The second method further compresses the token size by removing words with lower impact on the semantic. In order to adequately evaluate the effectiveness of the proposed methods, we propose and utilize a dataset called Food-Recommendation DB (FRDB) focusing on food recommendation for women around pregnancy period or infants. Our summarization compression can reduce 65{\%} of the retrieval token size with further 0.3{\%} improvement on the accuracy; semantic compression provides a more flexible way to trade-off the token size with performance, for which we can reduce the token size by 20{\%} with only 1.6{\%} of accuracy drop. | [
"Liu, Junyi",
"Li, Liangzhi",
"Xiang, Tong",
"Wang, Bowen",
"Qian, Yiming"
] | TCRA-LLM: Token Compression Retrieval Augmented Large Language Model for Inference Cost Reduction | findings-emnlp.655 | 2310.15556 | [
""
] | https://huggingface.co/papers/2310.15556 | 0 | 1 | 0 | 5 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.656.bib | https://aclanthology.org/2023.findings-emnlp.656/ | @inproceedings{lei-huang-2023-identifying,
title = "Identifying Conspiracy Theories News based on Event Relation Graph",
author = "Lei, Yuanyuan and
Huang, Ruihong",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.656",
doi = "10.18653/v1/2023.findings-emnlp.656",
pages = "9811--9822",
abstract = "Conspiracy theories, as a type of misinformation, are narratives that explains an event or situation in an irrational or malicious manner. While most previous work examined conspiracy theory in social media short texts, limited attention was put on such misinformation in long news documents. In this paper, we aim to identify whether a news article contains conspiracy theories. We observe that a conspiracy story can be made up by mixing uncorrelated events together, or by presenting an unusual distribution of relations between events. Achieving a contextualized understanding of events in a story is essential for detecting conspiracy theories. Thus, we propose to incorporate an event relation graph for each article, in which events are nodes, and four common types of event relations, coreference, temporal, causal, and subevent relations, are considered as edges. Then, we integrate the event relation graph into conspiracy theory identification in two ways: an event-aware language model is developed to augment the basic language model with the knowledge of events and event relations via soft labels; further, a heterogeneous graph attention network is designed to derive a graph embedding based on hard labels. Experiments on a large benchmark dataset show that our approach based on event relation graph improves both precision and recall of conspiracy theory identification, and generalizes well for new unseen media sources.",
}
| Conspiracy theories, as a type of misinformation, are narratives that explains an event or situation in an irrational or malicious manner. While most previous work examined conspiracy theory in social media short texts, limited attention was put on such misinformation in long news documents. In this paper, we aim to identify whether a news article contains conspiracy theories. We observe that a conspiracy story can be made up by mixing uncorrelated events together, or by presenting an unusual distribution of relations between events. Achieving a contextualized understanding of events in a story is essential for detecting conspiracy theories. Thus, we propose to incorporate an event relation graph for each article, in which events are nodes, and four common types of event relations, coreference, temporal, causal, and subevent relations, are considered as edges. Then, we integrate the event relation graph into conspiracy theory identification in two ways: an event-aware language model is developed to augment the basic language model with the knowledge of events and event relations via soft labels; further, a heterogeneous graph attention network is designed to derive a graph embedding based on hard labels. Experiments on a large benchmark dataset show that our approach based on event relation graph improves both precision and recall of conspiracy theory identification, and generalizes well for new unseen media sources. | [
"Lei, Yuanyuan",
"Huang, Ruihong"
] | Identifying Conspiracy Theories News based on Event Relation Graph | findings-emnlp.656 | 2310.18545 | [
"https://github.com/yuanyuanlei-nlp/conspiracy_theories_emnlp_2023"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.657.bib | https://aclanthology.org/2023.findings-emnlp.657/ | @inproceedings{murakhovska-etal-2023-salespeople,
title = "Salespeople vs {S}ales{B}ot: Exploring the Role of Educational Value in Conversational Recommender Systems",
author = "Murakhovs{'}ka, Lidiya and
Laban, Philippe and
Xie, Tian and
Xiong, Caiming and
Wu, Chien-Sheng",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.657",
doi = "10.18653/v1/2023.findings-emnlp.657",
pages = "9823--9838",
abstract = "Making big purchases requires consumers to research or consult a salesperson to gain domain expertise. However, existing conversational recommender systems (CRS) often overlook users{'} lack of background knowledge, focusing solely on gathering preferences. In this work, we define a new problem space for conversational agents that aim to provide both product recommendations and educational value through mixed-type mixed-initiative dialog. We introduce SalesOps, a framework that facilitates the simulation and evaluation of such systems by leveraging recent advancements in large language models (LLMs). We build SalesBot and ShopperBot, a pair of LLM-powered agents that can simulate either side of the framework. A comprehensive human study compares SalesBot against professional salespeople, revealing that although SalesBot approaches professional performance in terms of fluency and informativeness, it lags behind in recommendation quality. We emphasize the distinct limitations both face in providing truthful information, highlighting the challenges of ensuring faithfulness in the CRS context. We release our code and make all data available.",
}
| Making big purchases requires consumers to research or consult a salesperson to gain domain expertise. However, existing conversational recommender systems (CRS) often overlook users{'} lack of background knowledge, focusing solely on gathering preferences. In this work, we define a new problem space for conversational agents that aim to provide both product recommendations and educational value through mixed-type mixed-initiative dialog. We introduce SalesOps, a framework that facilitates the simulation and evaluation of such systems by leveraging recent advancements in large language models (LLMs). We build SalesBot and ShopperBot, a pair of LLM-powered agents that can simulate either side of the framework. A comprehensive human study compares SalesBot against professional salespeople, revealing that although SalesBot approaches professional performance in terms of fluency and informativeness, it lags behind in recommendation quality. We emphasize the distinct limitations both face in providing truthful information, highlighting the challenges of ensuring faithfulness in the CRS context. We release our code and make all data available. | [
"Murakhovs{'}ka, Lidiya",
"Laban, Philippe",
"Xie, Tian",
"Xiong, Caiming",
"Wu, Chien-Sheng"
] | Salespeople vs SalesBot: Exploring the Role of Educational Value in Conversational Recommender Systems | findings-emnlp.657 | 2310.17749 | [
"https://github.com/salesforce/salesbot"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.658.bib | https://aclanthology.org/2023.findings-emnlp.658/ | @inproceedings{ma-etal-2023-dynamic,
title = "Dynamic Open-book Prompt for Conversational Recommender System",
author = "Ma, Xuan and
Qian, Tieyun and
Sun, Ke",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.658",
doi = "10.18653/v1/2023.findings-emnlp.658",
pages = "9839--9849",
abstract = "Conversational Recommender System (CRS) aims to deliver personalized recommendations through interactive dialogues. Recent advances in prompt learning have shed light on this task. However, the performance of existing methods is confined by the limited context within ongoing conversations. Moreover, these methods utilize training samples only for prompt parameter training. The constructed prompt lacks the ability to refer to the training data during inference, which exacerbates the problem of limited context. To solve this problem, we propose a novel Dynamic Open-book Prompt approach, where the open book stores users{'} experiences in historical data, and we dynamically construct the prompt to memorize the user{'}s current utterance and selectively retrieve relevant contexts from the open book. Specifically, we first build an item-recommendation graph from the open book and convolute on the graph to form a base prompt which contains more information besides the finite dialogue. Then, we enhance the representation learning process of the prompt by tailoring similar contexts in the graph into the prompt to meet the user{'}s current need. This ensures the prompt provides targeted suggestions that are both informed and contextually relevant. Extensive experimental results on the ReDial dataset demonstrate the significant improvements achieved by our proposed model over the state-of-the-art methods. Our code and data are available at https://github.com/NLPWM-WHU/DOP.",
}
| Conversational Recommender System (CRS) aims to deliver personalized recommendations through interactive dialogues. Recent advances in prompt learning have shed light on this task. However, the performance of existing methods is confined by the limited context within ongoing conversations. Moreover, these methods utilize training samples only for prompt parameter training. The constructed prompt lacks the ability to refer to the training data during inference, which exacerbates the problem of limited context. To solve this problem, we propose a novel Dynamic Open-book Prompt approach, where the open book stores users{'} experiences in historical data, and we dynamically construct the prompt to memorize the user{'}s current utterance and selectively retrieve relevant contexts from the open book. Specifically, we first build an item-recommendation graph from the open book and convolute on the graph to form a base prompt which contains more information besides the finite dialogue. Then, we enhance the representation learning process of the prompt by tailoring similar contexts in the graph into the prompt to meet the user{'}s current need. This ensures the prompt provides targeted suggestions that are both informed and contextually relevant. Extensive experimental results on the ReDial dataset demonstrate the significant improvements achieved by our proposed model over the state-of-the-art methods. Our code and data are available at https://github.com/NLPWM-WHU/DOP. | [
"Ma, Xuan",
"Qian, Tieyun",
"Sun, Ke"
] | Dynamic Open-book Prompt for Conversational Recommender System | findings-emnlp.658 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.659.bib | https://aclanthology.org/2023.findings-emnlp.659/ | @inproceedings{zhang-etal-2023-auto,
title = "Auto-Instruct: Automatic Instruction Generation and Ranking for Black-Box Language Models",
author = "Zhang, Zhihan and
Wang, Shuohang and
Yu, Wenhao and
Xu, Yichong and
Iter, Dan and
Zeng, Qingkai and
Liu, Yang and
Zhu, Chenguang and
Jiang, Meng",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.659",
doi = "10.18653/v1/2023.findings-emnlp.659",
pages = "9850--9867",
abstract = "Large language models (LLMs) can perform a wide range of tasks by following natural language instructions, without the necessity of task-specific fine-tuning. Unfortunately, the performance of LLMs is greatly influenced by the quality of these instructions, and manually writing effective instructions for each task is a laborious and subjective process. In this paper, we introduce Auto-Instruct, a novel method to automatically improve the quality of instructions provided to LLMs. Our method leverages the inherent generative ability of LLMs to produce diverse candidate instructions for a given task, and then ranks them using a scoring model trained on a variety of 575 existing NLP tasks. In experiments on 118 out-of-domain tasks, Auto-Instruct surpasses both human-written instructions and existing baselines of LLM-generated instructions. Furthermore, our method exhibits notable generalizability even with other LLMs that are not incorporated into its training process.",
}
| Large language models (LLMs) can perform a wide range of tasks by following natural language instructions, without the necessity of task-specific fine-tuning. Unfortunately, the performance of LLMs is greatly influenced by the quality of these instructions, and manually writing effective instructions for each task is a laborious and subjective process. In this paper, we introduce Auto-Instruct, a novel method to automatically improve the quality of instructions provided to LLMs. Our method leverages the inherent generative ability of LLMs to produce diverse candidate instructions for a given task, and then ranks them using a scoring model trained on a variety of 575 existing NLP tasks. In experiments on 118 out-of-domain tasks, Auto-Instruct surpasses both human-written instructions and existing baselines of LLM-generated instructions. Furthermore, our method exhibits notable generalizability even with other LLMs that are not incorporated into its training process. | [
"Zhang, Zhihan",
"Wang, Shuohang",
"Yu, Wenhao",
"Xu, Yichong",
"Iter, Dan",
"Zeng, Qingkai",
"Liu, Yang",
"Zhu, Chenguang",
"Jiang, Meng"
] | Auto-Instruct: Automatic Instruction Generation and Ranking for Black-Box Language Models | findings-emnlp.659 | 2310.13127 | [
""
] | https://huggingface.co/papers/2310.13127 | 4 | 11 | 1 | 9 | [
"zhihz0535/Auto-Instruct-Flan-T5-davinci003-zeroshot",
"zhihz0535/Auto-Instruct-Flan-T5-davinci003-fewshot"
] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.660.bib | https://aclanthology.org/2023.findings-emnlp.660/ | @inproceedings{gong-etal-2023-diffuseq,
title = "{D}iffu{S}eq-v2: Bridging Discrete and Continuous Text Spaces for Accelerated {S}eq2{S}eq Diffusion Models",
author = "Gong, Shansan and
Li, Mukai and
Feng, Jiangtao and
Wu, Zhiyong and
Kong, Lingpeng",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.660",
doi = "10.18653/v1/2023.findings-emnlp.660",
pages = "9868--9875",
abstract = "Diffusion models have gained prominence in generating high-quality sequences of text. Nevertheless, current approaches predominantly represent discrete text within a continuous diffusion space, which incurs substantial computational overhead during training and results in slower sampling speeds. In this paper, we introduce a soft absorbing state that facilitates the diffusion model in learning to reconstruct discrete mutations based on the underlying Gaussian space, thereby enhancing its capacity to recover conditional signals. During the sampling phase, we employ state-of-the-art ODE solvers within the continuous space to expedite the sampling process. Comprehensive experimental evaluations reveal that our proposed method effectively accelerates the training convergence by 4x and generates samples of similar quality 800x faster, rendering it significantly closer to practical application.",
}
| Diffusion models have gained prominence in generating high-quality sequences of text. Nevertheless, current approaches predominantly represent discrete text within a continuous diffusion space, which incurs substantial computational overhead during training and results in slower sampling speeds. In this paper, we introduce a soft absorbing state that facilitates the diffusion model in learning to reconstruct discrete mutations based on the underlying Gaussian space, thereby enhancing its capacity to recover conditional signals. During the sampling phase, we employ state-of-the-art ODE solvers within the continuous space to expedite the sampling process. Comprehensive experimental evaluations reveal that our proposed method effectively accelerates the training convergence by 4x and generates samples of similar quality 800x faster, rendering it significantly closer to practical application. | [
"Gong, Shansan",
"Li, Mukai",
"Feng, Jiangtao",
"Wu, Zhiyong",
"Kong, Lingpeng"
] | DiffuSeq-v2: Bridging Discrete and Continuous Text Spaces for Accelerated Seq2Seq Diffusion Models | findings-emnlp.660 | 2310.05793 | [
"https://github.com/Shark-NLP/DiffuSeq"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.661.bib | https://aclanthology.org/2023.findings-emnlp.661/ | @inproceedings{guo-etal-2023-m2c,
title = "{M}2{C}: Towards Automatic Multimodal Manga Complement",
author = "Guo, Hongcheng and
Wang, Boyang and
Bai, Jiaqi and
Liu, Jiaheng and
Yang, Jian and
Li, Zhoujun",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.661",
doi = "10.18653/v1/2023.findings-emnlp.661",
pages = "9876--9882",
abstract = "Multimodal manga analysis focuses on enhancing manga understanding with visual and textual features, which has attracted considerable attention from both natural language processing and computer vision communities. Currently, most comics are hand-drawn and prone to problems such as missing pages, text contamination, and text aging, resulting in missing comic text content and seriously hindering human comprehension. In other words, the Multimodal Manga Complement (\textbf{M2C}) task has not been investigated, which aims to handle the aforementioned issues by providing a shared semantic space for vision and language understanding. To this end, we first propose the Multimodal Manga Complement task by establishing a new M2C benchmark dataset covering two languages. First, we design a manga argumentation method called MCoT to mine event knowledge in comics with large language models. Then, an effective baseline FVP-M$^{2}$ using fine-grained visual prompts is proposed to support manga complement. Extensive experimental results show the effectiveness of FVP-M$^{2}$ method for Multimodal Mange Complement.",
}
| Multimodal manga analysis focuses on enhancing manga understanding with visual and textual features, which has attracted considerable attention from both natural language processing and computer vision communities. Currently, most comics are hand-drawn and prone to problems such as missing pages, text contamination, and text aging, resulting in missing comic text content and seriously hindering human comprehension. In other words, the Multimodal Manga Complement (\textbf{M2C}) task has not been investigated, which aims to handle the aforementioned issues by providing a shared semantic space for vision and language understanding. To this end, we first propose the Multimodal Manga Complement task by establishing a new M2C benchmark dataset covering two languages. First, we design a manga argumentation method called MCoT to mine event knowledge in comics with large language models. Then, an effective baseline FVP-M$^{2}$ using fine-grained visual prompts is proposed to support manga complement. Extensive experimental results show the effectiveness of FVP-M$^{2}$ method for Multimodal Mange Complement. | [
"Guo, Hongcheng",
"Wang, Boyang",
"Bai, Jiaqi",
"Liu, Jiaheng",
"Yang, Jian",
"Li, Zhoujun"
] | M2C: Towards Automatic Multimodal Manga Complement | findings-emnlp.661 | 2310.17130 | [
"https://github.com/hc-guo/m2c"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.662.bib | https://aclanthology.org/2023.findings-emnlp.662/ | @inproceedings{thawani-etal-2023-learn,
title = "Learn Your Tokens: Word-Pooled Tokenization for Language Modeling",
author = "Thawani, Avijit and
Ghanekar, Saurabh and
Zhu, Xiaoyuan and
Pujara, Jay",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.662",
doi = "10.18653/v1/2023.findings-emnlp.662",
pages = "9883--9893",
abstract = "Language models typically tokenize text into subwords, using a deterministic, hand-engineered heuristic of combining characters into longer surface-level strings such as {`}ing{'} or whole words. Recent literature has repeatedly shown the limitations of such a tokenization strategy, particularly for documents not written in English and for representing numbers. On the other extreme, byte/character-level language models are much less restricted but suffer from increased sequence description lengths and a subsequent quadratic expansion in self-attention computation. Recent attempts to compress and limit these context lengths with fixed size convolutions is helpful but completely ignores the word boundary. This paper considers an alternative {`}learn your tokens{'} scheme which utilizes the word boundary to pool bytes/characters into word representations, which are fed to the primary language model, before again decoding individual characters/bytes per word in parallel. We find that our moderately expressive and moderately fast end-to-end tokenizer outperform by over {`}300{\%}{`} both subwords and byte/character models over the intrinsic language modeling metric of next-word prediction across datasets. It particularly outshines on rare words, outperforming by a factor of 30! We extensively study the language modeling setup for all three categories of tokenizers and theoretically analyze how our end-to-end models can also be a strong trade-off in efficiency and robustness.",
}
| Language models typically tokenize text into subwords, using a deterministic, hand-engineered heuristic of combining characters into longer surface-level strings such as {`}ing{'} or whole words. Recent literature has repeatedly shown the limitations of such a tokenization strategy, particularly for documents not written in English and for representing numbers. On the other extreme, byte/character-level language models are much less restricted but suffer from increased sequence description lengths and a subsequent quadratic expansion in self-attention computation. Recent attempts to compress and limit these context lengths with fixed size convolutions is helpful but completely ignores the word boundary. This paper considers an alternative {`}learn your tokens{'} scheme which utilizes the word boundary to pool bytes/characters into word representations, which are fed to the primary language model, before again decoding individual characters/bytes per word in parallel. We find that our moderately expressive and moderately fast end-to-end tokenizer outperform by over {`}300{\%}{`} both subwords and byte/character models over the intrinsic language modeling metric of next-word prediction across datasets. It particularly outshines on rare words, outperforming by a factor of 30! We extensively study the language modeling setup for all three categories of tokenizers and theoretically analyze how our end-to-end models can also be a strong trade-off in efficiency and robustness. | [
"Thawani, Avijit",
"Ghanekar, Saurabh",
"Zhu, Xiaoyuan",
"Pujara, Jay"
] | Learn Your Tokens: Word-Pooled Tokenization for Language Modeling | findings-emnlp.662 | 2310.11628 | [
"https://github.com/avi-jit/etok"
] | https://huggingface.co/papers/2310.11628 | 1 | 0 | 0 | 4 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.663.bib | https://aclanthology.org/2023.findings-emnlp.663/ | @inproceedings{yang-etal-2023-towards-detecting,
title = "Towards Detecting Contextual Real-Time Toxicity for In-Game Chat",
author = "Yang, Zachary and
Grenon-Godbout, Nicolas and
Rabbany, Reihaneh",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.663",
doi = "10.18653/v1/2023.findings-emnlp.663",
pages = "9894--9906",
abstract = "Real-time toxicity detection in online environments poses a significant challenge, due to the increasing prevalence of social media and gaming platforms. We introduce ToxBuster, a simple and scalable model that reliably detects toxic content in real-time for a line of chat by including chat history and metadata. ToxBuster consistently outperforms conventional toxicity models across popular multiplayer games, including Rainbow Six Siege, For Honor, and DOTA 2. We conduct an ablation study to assess the importance of each model component and explore ToxBuster{'}s transferability across the datasets. Furthermore, we showcase ToxBuster{'}s efficacy in post-game moderation, successfully flagging 82.1{\%} of chat-reported players at a precision level of 90.0{\%}. Additionally, we show how an additional 6{\%} of unreported toxic players can be proactively moderated.",
}
| Real-time toxicity detection in online environments poses a significant challenge, due to the increasing prevalence of social media and gaming platforms. We introduce ToxBuster, a simple and scalable model that reliably detects toxic content in real-time for a line of chat by including chat history and metadata. ToxBuster consistently outperforms conventional toxicity models across popular multiplayer games, including Rainbow Six Siege, For Honor, and DOTA 2. We conduct an ablation study to assess the importance of each model component and explore ToxBuster{'}s transferability across the datasets. Furthermore, we showcase ToxBuster{'}s efficacy in post-game moderation, successfully flagging 82.1{\%} of chat-reported players at a precision level of 90.0{\%}. Additionally, we show how an additional 6{\%} of unreported toxic players can be proactively moderated. | [
"Yang, Zachary",
"Grenon-Godbout, Nicolas",
"Rabbany, Reihaneh"
] | Towards Detecting Contextual Real-Time Toxicity for In-Game Chat | findings-emnlp.663 | 2310.18330 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.664.bib | https://aclanthology.org/2023.findings-emnlp.664/ | @inproceedings{gueuwou-etal-2023-jwsign,
title = "{JWS}ign: A Highly Multilingual Corpus of {B}ible Translations for more Diversity in Sign Language Processing",
author = {Gueuwou, Shester and
Siake, Sophie and
Leong, Colin and
M{\"u}ller, Mathias},
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.664",
doi = "10.18653/v1/2023.findings-emnlp.664",
pages = "9907--9927",
abstract = "Advancements in sign language processing have been hindered by a lack of sufficient data, impeding progress in recognition, translation, and production tasks. The absence of comprehensive sign language datasets across the world{'}s sign languages has widened the gap in this field, resulting in a few sign languages being studied more than others, making this research area extremely skewed mostly towards sign languages from high-income countries. In this work we introduce a new large and highly multilingual dataset for sign language translation: JWSign. The dataset consists of 2,530 hours of Bible translations in 98 sign languages, featuring more than 1,500 individual signers. On this dataset, we report neural machine translation experiments. Apart from bilingual baseline systems, we also train multilingual systems, including some that take into account the typological relatedness of signed or spoken languages. Our experiments highlight that multilingual systems are superior to bilingual baselines, and that in higher-resource scenarios, clustering language pairs that are related improves translation quality.",
}
| Advancements in sign language processing have been hindered by a lack of sufficient data, impeding progress in recognition, translation, and production tasks. The absence of comprehensive sign language datasets across the world{'}s sign languages has widened the gap in this field, resulting in a few sign languages being studied more than others, making this research area extremely skewed mostly towards sign languages from high-income countries. In this work we introduce a new large and highly multilingual dataset for sign language translation: JWSign. The dataset consists of 2,530 hours of Bible translations in 98 sign languages, featuring more than 1,500 individual signers. On this dataset, we report neural machine translation experiments. Apart from bilingual baseline systems, we also train multilingual systems, including some that take into account the typological relatedness of signed or spoken languages. Our experiments highlight that multilingual systems are superior to bilingual baselines, and that in higher-resource scenarios, clustering language pairs that are related improves translation quality. | [
"Gueuwou, Shester",
"Siake, Sophie",
"Leong, Colin",
"M{\\\"u}ller, Mathias"
] | JWSign: A Highly Multilingual Corpus of Bible Translations for more Diversity in Sign Language Processing | findings-emnlp.664 | 2311.10174 | [
"https://github.com/shesterg/jwsign-machine-translation"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.665.bib | https://aclanthology.org/2023.findings-emnlp.665/ | @inproceedings{cowap-etal-2023-stochastic,
title = "Do Stochastic Parrots have Feelings Too? Improving Neural Detection of Synthetic Text via Emotion Recognition",
author = "Cowap, Alan and
Graham, Yvette and
Foster, Jennifer",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.665",
doi = "10.18653/v1/2023.findings-emnlp.665",
pages = "9928--9946",
abstract = "Recent developments in generative AI have shone a spotlight on high-performance synthetic text generation technologies. The now wide availability and ease of use of such models highlights the urgent need to provide equally powerful technologies capable of identifying synthetic text. With this in mind, we draw inspiration from psychological studies which suggest that people can be driven by emotion and encode emotion in the text they compose. We hypothesize that pretrained language models (PLMs) have an affective deficit because they lack such an emotional driver when generating text and consequently may generate synthetic text which has affective incoherence i.e. lacking the kind of emotional coherence present in human-authored text. We subsequently develop an emotionally aware detector by fine-tuning a PLM on emotion. Experiment results indicate that our emotionally-aware detector achieves improvements across a range of synthetic text generators, various sized models, datasets, and domains. Finally, we compare our emotionally-aware synthetic text detector to ChatGPT in the task of identification of its own output and show substantial gains, reinforcing the potential of emotion as a signal to identify synthetic text. Code, models, and datasets are available at https: //github.com/alanagiasi/emoPLMsynth",
}
| Recent developments in generative AI have shone a spotlight on high-performance synthetic text generation technologies. The now wide availability and ease of use of such models highlights the urgent need to provide equally powerful technologies capable of identifying synthetic text. With this in mind, we draw inspiration from psychological studies which suggest that people can be driven by emotion and encode emotion in the text they compose. We hypothesize that pretrained language models (PLMs) have an affective deficit because they lack such an emotional driver when generating text and consequently may generate synthetic text which has affective incoherence i.e. lacking the kind of emotional coherence present in human-authored text. We subsequently develop an emotionally aware detector by fine-tuning a PLM on emotion. Experiment results indicate that our emotionally-aware detector achieves improvements across a range of synthetic text generators, various sized models, datasets, and domains. Finally, we compare our emotionally-aware synthetic text detector to ChatGPT in the task of identification of its own output and show substantial gains, reinforcing the potential of emotion as a signal to identify synthetic text. Code, models, and datasets are available at https: //github.com/alanagiasi/emoPLMsynth | [
"Cowap, Alan",
"Graham, Yvette",
"Foster, Jennifer"
] | Do Stochastic Parrots have Feelings Too? Improving Neural Detection of Synthetic Text via Emotion Recognition | findings-emnlp.665 | 2310.15904 | [
"https://github.com/alanagiasi/emoplmsynth"
] | https://huggingface.co/papers/2310.15904 | 1 | 0 | 0 | 3 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.666.bib | https://aclanthology.org/2023.findings-emnlp.666/ | @inproceedings{xiao-etal-2023-variator,
title = "Variator: Accelerating Pre-trained Models with Plug-and-Play Compression Modules",
author = "Xiao, Chaojun and
Luo, Yuqi and
Zhang, Wenbin and
Zhang, Pengle and
Han, Xu and
Lin, Yankai and
Zhang, Zhengyan and
Xie, Ruobing and
Liu, Zhiyuan and
Sun, Maosong and
Zhou, Jie",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.666",
doi = "10.18653/v1/2023.findings-emnlp.666",
pages = "9947--9959",
abstract = "Large language models (LLMs) have achieved remarkable results on NLP tasks but at the expense of huge parameter sizes and the consequent computational costs. In this paper, we propose Variator, a parameter-efficient acceleration method that enhances computational efficiency through plug-and-play compression plugins. Compression plugins are designed to reduce the sequence length via compressing multiple hidden vectors into one and trained with original LLMs frozen. Different from traditional model acceleration methods, which compress LLMs to smaller sizes, Variator offers two distinct advantages: (1) In real-world applications, the plug-and-play nature of our compression plugins enables dynamic selection of different compression plugins with varying acceleration ratios based on the current workload. (2) The compression plugin comprises a few compact neural network layers with minimal parameters, significantly saving storage and memory overhead, particularly in scenarios with a growing number of tasks. We validate the effectiveness of Variator on seven datasets. Experimental results show that Variator can save 53{\%} computational costs using only 0.9{\%} additional parameters with a performance drop of less than 2{\%}. Moreover, when the model scales to billions of parameters, Variator matches the strong performance of uncompressed LLMs. Our code and checkpoints will be released to facilitate future work.",
}
| Large language models (LLMs) have achieved remarkable results on NLP tasks but at the expense of huge parameter sizes and the consequent computational costs. In this paper, we propose Variator, a parameter-efficient acceleration method that enhances computational efficiency through plug-and-play compression plugins. Compression plugins are designed to reduce the sequence length via compressing multiple hidden vectors into one and trained with original LLMs frozen. Different from traditional model acceleration methods, which compress LLMs to smaller sizes, Variator offers two distinct advantages: (1) In real-world applications, the plug-and-play nature of our compression plugins enables dynamic selection of different compression plugins with varying acceleration ratios based on the current workload. (2) The compression plugin comprises a few compact neural network layers with minimal parameters, significantly saving storage and memory overhead, particularly in scenarios with a growing number of tasks. We validate the effectiveness of Variator on seven datasets. Experimental results show that Variator can save 53{\%} computational costs using only 0.9{\%} additional parameters with a performance drop of less than 2{\%}. Moreover, when the model scales to billions of parameters, Variator matches the strong performance of uncompressed LLMs. Our code and checkpoints will be released to facilitate future work. | [
"Xiao, Chaojun",
"Luo, Yuqi",
"Zhang, Wenbin",
"Zhang, Pengle",
"Han, Xu",
"Lin, Yankai",
"Zhang, Zhengyan",
"Xie, Ruobing",
"Liu, Zhiyuan",
"Sun, Maosong",
"Zhou, Jie"
] | Variator: Accelerating Pre-trained Models with Plug-and-Play Compression Modules | findings-emnlp.666 | 2310.15724 | [
"https://github.com/thunlp/compression-plugin"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.667.bib | https://aclanthology.org/2023.findings-emnlp.667/ | @inproceedings{he-etal-2023-pivotfec,
title = "{P}ivot{FEC}: Enhancing Few-shot Factual Error Correction with a Pivot Task Approach using Large Language Models",
author = "He, Xingwei and
Jin, A-Long and
Ma, Jun and
Yuan, Yuan and
Yiu, Siu",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.667",
doi = "10.18653/v1/2023.findings-emnlp.667",
pages = "9960--9976",
abstract = "Factual Error Correction (FEC) aims to rectify false claims by making minimal revisions to align them more accurately with supporting evidence. However, the lack of datasets containing false claims and their corresponding corrections has impeded progress in this field. Existing distantly supervised models typically employ the mask-then-correct paradigm, where a masker identifies problematic spans in false claims, followed by a corrector to predict the masked portions. Unfortunately, accurately identifying errors in claims is challenging, leading to issues like over-erasure and incorrect masking. To overcome these challenges, we present PivotFEC, a method that enhances few-shot FEC with a pivot task approach using large language models (LLMs). Specifically, we introduce a pivot task called factual error injection, which leverages LLMs (e.g., ChatGPT) to intentionally generate text containing factual errors under few-shot settings; then, the generated text with factual errors can be used to train the FEC corrector. Our experiments on a public dataset demonstrate the effectiveness of PivotFEC in two significant ways: Firstly, it improves the widely-adopted SARI metrics by 11.3 compared to the best-performing distantly supervised methods. Secondly, it outperforms its few-shot counterpart (i.e., LLMs are directly used to solve FEC) by 7.9 points in SARI, validating the efficacy of our proposed pivot task.",
}
| Factual Error Correction (FEC) aims to rectify false claims by making minimal revisions to align them more accurately with supporting evidence. However, the lack of datasets containing false claims and their corresponding corrections has impeded progress in this field. Existing distantly supervised models typically employ the mask-then-correct paradigm, where a masker identifies problematic spans in false claims, followed by a corrector to predict the masked portions. Unfortunately, accurately identifying errors in claims is challenging, leading to issues like over-erasure and incorrect masking. To overcome these challenges, we present PivotFEC, a method that enhances few-shot FEC with a pivot task approach using large language models (LLMs). Specifically, we introduce a pivot task called factual error injection, which leverages LLMs (e.g., ChatGPT) to intentionally generate text containing factual errors under few-shot settings; then, the generated text with factual errors can be used to train the FEC corrector. Our experiments on a public dataset demonstrate the effectiveness of PivotFEC in two significant ways: Firstly, it improves the widely-adopted SARI metrics by 11.3 compared to the best-performing distantly supervised methods. Secondly, it outperforms its few-shot counterpart (i.e., LLMs are directly used to solve FEC) by 7.9 points in SARI, validating the efficacy of our proposed pivot task. | [
"He, Xingwei",
"Jin, A-Long",
"Ma, Jun",
"Yuan, Yuan",
"Yiu, Siu"
] | PivotFEC: Enhancing Few-shot Factual Error Correction with a Pivot Task Approach using Large Language Models | findings-emnlp.667 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.668.bib | https://aclanthology.org/2023.findings-emnlp.668/ | @inproceedings{becquin-esmeir-2023-semantic,
title = "Semantic Similarity Covariance Matrix Shrinkage",
author = "Becquin, Guillaume and
Esmeir, Saher",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.668",
doi = "10.18653/v1/2023.findings-emnlp.668",
pages = "9977--9992",
abstract = "An accurate estimation of the covariance matrix is a critical component of many applications in finance, including portfolio optimization. The sample covariance suffers from the curse of dimensionality when the number of observations is in the same order or lower than the number of variables. This tends to be the case in portfolio optimization, where a portfolio manager can choose between thousands of stocks using historical daily returns to guide their investment decisions. To address this issue, past works proposed linear covariance shrinkage to regularize the estimated matrix. While effective, the proposed methods relied solely on historical price data and thus ignored company fundamental data. In this work, we propose to utilise semantic similarity derived from textual descriptions or knowledge graphs to improve the covariance estimation. Rather than using the semantic similarity directly as a biased estimator to the covariance, we employ it as a shrinkage target. The resulting covariance estimators leverage both semantic similarity and recent price history, and can be readily adapted to a broad range of financial securities. The effectiveness of the approach is demonstrated for a period including diverse market conditions and compared with the covariance shrinkage prior art.",
}
| An accurate estimation of the covariance matrix is a critical component of many applications in finance, including portfolio optimization. The sample covariance suffers from the curse of dimensionality when the number of observations is in the same order or lower than the number of variables. This tends to be the case in portfolio optimization, where a portfolio manager can choose between thousands of stocks using historical daily returns to guide their investment decisions. To address this issue, past works proposed linear covariance shrinkage to regularize the estimated matrix. While effective, the proposed methods relied solely on historical price data and thus ignored company fundamental data. In this work, we propose to utilise semantic similarity derived from textual descriptions or knowledge graphs to improve the covariance estimation. Rather than using the semantic similarity directly as a biased estimator to the covariance, we employ it as a shrinkage target. The resulting covariance estimators leverage both semantic similarity and recent price history, and can be readily adapted to a broad range of financial securities. The effectiveness of the approach is demonstrated for a period including diverse market conditions and compared with the covariance shrinkage prior art. | [
"Becquin, Guillaume",
"Esmeir, Saher"
] | Semantic Similarity Covariance Matrix Shrinkage | findings-emnlp.668 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.669.bib | https://aclanthology.org/2023.findings-emnlp.669/ | @inproceedings{dai-etal-2023-llm,
title = "{LLM}-in-the-loop: Leveraging Large Language Model for Thematic Analysis",
author = "Dai, Shih-Chieh and
Xiong, Aiping and
Ku, Lun-Wei",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.669",
doi = "10.18653/v1/2023.findings-emnlp.669",
pages = "9993--10001",
abstract = "Thematic analysis (TA) has been widely used for analyzing qualitative data in many disciplines and fields. To ensure reliable analysis, the same piece of data is typically assigned to at least two human coders. Moreover, to produce meaningful and useful analysis, human coders develop and deepen their data interpretation and coding over multiple iterations, making TA labor-intensive and time-consuming. Recently the emerging field of large language models (LLMs) research has shown that LLMs have the potential replicate human-like behavior in various tasks: in particular, LLMs outperform crowd workers on text-annotation tasks, suggesting an opportunity to leverage LLMs on TA. We propose a human{--}LLM collaboration framework (i.e., LLM-in-the-loop) to conduct TA with in-context learning (ICL). This framework provides the prompt to frame discussions with a LLM (e.g., GPT-3.5) to generate the final codebook for TA. We demonstrate the utility of this framework using survey datasets on the aspects of the music listening experience and the usage of a password manager. Results of the two case studies show that the proposed framework yields similar coding quality to that of human coders but reduces TA{'}s labor and time demands.",
}
| Thematic analysis (TA) has been widely used for analyzing qualitative data in many disciplines and fields. To ensure reliable analysis, the same piece of data is typically assigned to at least two human coders. Moreover, to produce meaningful and useful analysis, human coders develop and deepen their data interpretation and coding over multiple iterations, making TA labor-intensive and time-consuming. Recently the emerging field of large language models (LLMs) research has shown that LLMs have the potential replicate human-like behavior in various tasks: in particular, LLMs outperform crowd workers on text-annotation tasks, suggesting an opportunity to leverage LLMs on TA. We propose a human{--}LLM collaboration framework (i.e., LLM-in-the-loop) to conduct TA with in-context learning (ICL). This framework provides the prompt to frame discussions with a LLM (e.g., GPT-3.5) to generate the final codebook for TA. We demonstrate the utility of this framework using survey datasets on the aspects of the music listening experience and the usage of a password manager. Results of the two case studies show that the proposed framework yields similar coding quality to that of human coders but reduces TA{'}s labor and time demands. | [
"Dai, Shih-Chieh",
"Xiong, Aiping",
"Ku, Lun-Wei"
] | LLM-in-the-loop: Leveraging Large Language Model for Thematic Analysis | findings-emnlp.669 | 2310.15100 | [
"https://github.com/sjdai/llm-thematic-analysis"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.670.bib | https://aclanthology.org/2023.findings-emnlp.670/ | @inproceedings{mishra-etal-2023-llm,
title = "{LLM} aided semi-supervision for efficient Extractive Dialog Summarization",
author = "Mishra, Nishant and
Sahu, Gaurav and
Calixto, Iacer and
Abu-Hanna, Ameen and
Laradji, Issam",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.670",
doi = "10.18653/v1/2023.findings-emnlp.670",
pages = "10002--10009",
abstract = "Generating high-quality summaries for chat dialogs often requires large labeled datasets. We propose a method to efficiently use unlabeled data for extractive summarization of customer-agent dialogs. In our method, we frame summarization as a question-answering problem and use state-of-the-art large language models (LLMs) to generate pseudo-labels for a dialog. We then use these pseudo-labels to fine-tune a chat summarization model, effectively transferring knowledge from the large LLM into a smaller specialized model. We demonstrate our method on the TWEETSUMM dataset, and show that using 10{\%} of the original labelled data set we can achieve 65.9/57.0/61.0 ROUGE-1/-2/-L, whereas the current state-of-the-art trained on the entire training data set obtains 65.16/55.81/64.37 ROUGE-1/-2/-L. In other words, in the worst case (i.e., ROUGE-L) we still effectively retain 94.7{\%} of the performance while using only 10{\%} of the data.",
}
| Generating high-quality summaries for chat dialogs often requires large labeled datasets. We propose a method to efficiently use unlabeled data for extractive summarization of customer-agent dialogs. In our method, we frame summarization as a question-answering problem and use state-of-the-art large language models (LLMs) to generate pseudo-labels for a dialog. We then use these pseudo-labels to fine-tune a chat summarization model, effectively transferring knowledge from the large LLM into a smaller specialized model. We demonstrate our method on the TWEETSUMM dataset, and show that using 10{\%} of the original labelled data set we can achieve 65.9/57.0/61.0 ROUGE-1/-2/-L, whereas the current state-of-the-art trained on the entire training data set obtains 65.16/55.81/64.37 ROUGE-1/-2/-L. In other words, in the worst case (i.e., ROUGE-L) we still effectively retain 94.7{\%} of the performance while using only 10{\%} of the data. | [
"Mishra, Nishant",
"Sahu, Gaurav",
"Calixto, Iacer",
"Abu-Hanna, Ameen",
"Laradji, Issam"
] | LLM aided semi-supervision for efficient Extractive Dialog Summarization | findings-emnlp.670 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.671.bib | https://aclanthology.org/2023.findings-emnlp.671/ | @inproceedings{chai-strube-2023-investigating,
title = "Investigating Multilingual Coreference Resolution by Universal Annotations",
author = "Chai, Haixia and
Strube, Michael",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.671",
doi = "10.18653/v1/2023.findings-emnlp.671",
pages = "10010--10024",
abstract = "Multilingual coreference resolution (MCR) has been a long-standing and challenging task. With the newly proposed multilingual coreference dataset, CorefUD (Nedoluzhko et al., 2022), we conduct an investigation into the task by using its harmonized universal morphosyntactic and coreference annotations. First, we study coreference by examining the ground truth data at different linguistic levels, namely mention, entity and document levels, and across different genres, to gain insights into the characteristics of coreference across multiple languages. Second, we perform an error analysis of the most challenging cases that the SotA system fails to resolve in the CRAC 2022 shared task using the universal annotations. Last, based on this analysis, we extract features from universal morphosyntactic annotations and integrate these features into a baseline system to assess their potential benefits for the MCR task. Our results show that our best configuration of features improves the baseline by 0.9{\%} F1 score.",
}
| Multilingual coreference resolution (MCR) has been a long-standing and challenging task. With the newly proposed multilingual coreference dataset, CorefUD (Nedoluzhko et al., 2022), we conduct an investigation into the task by using its harmonized universal morphosyntactic and coreference annotations. First, we study coreference by examining the ground truth data at different linguistic levels, namely mention, entity and document levels, and across different genres, to gain insights into the characteristics of coreference across multiple languages. Second, we perform an error analysis of the most challenging cases that the SotA system fails to resolve in the CRAC 2022 shared task using the universal annotations. Last, based on this analysis, we extract features from universal morphosyntactic annotations and integrate these features into a baseline system to assess their potential benefits for the MCR task. Our results show that our best configuration of features improves the baseline by 0.9{\%} F1 score. | [
"Chai, Haixia",
"Strube, Michael"
] | Investigating Multilingual Coreference Resolution by Universal Annotations | findings-emnlp.671 | 2310.17734 | [
"https://github.com/haixiachai/multi-coref"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.672.bib | https://aclanthology.org/2023.findings-emnlp.672/ | @inproceedings{zhang-etal-2023-factspotter,
title = "{F}act{S}potter: Evaluating the Factual Faithfulness of Graph-to-Text Generation",
author = "Zhang, Kun and
Balalau, Oana and
Manolescu, Ioana",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.672",
doi = "10.18653/v1/2023.findings-emnlp.672",
pages = "10025--10042",
abstract = "Graph-to-text (G2T) generation takes a graph as input and aims to generate a fluent and faith- ful textual representation of the information in the graph. The task has many applications, such as dialogue generation and question an- swering. In this work, we investigate to what extent the G2T generation problem is solved for previously studied datasets, and how pro- posed metrics perform when comparing generated texts. To help address their limitations, we propose a new metric that correctly identifies factual faithfulness, i.e., given a triple (subject, predicate, object), it decides if the triple is present in a generated text. We show that our metric FactSpotter achieves the highest correlation with human annotations on data correct- ness, data coverage, and relevance. In addition, FactSpotter can be used as a plug-in feature to improve the factual faithfulness of existing models. Finally, we investigate if existing G2T datasets are still challenging for state-of-the-art models. Our code is available online: https://github.com/guihuzhang/FactSpotter.",
}
| Graph-to-text (G2T) generation takes a graph as input and aims to generate a fluent and faith- ful textual representation of the information in the graph. The task has many applications, such as dialogue generation and question an- swering. In this work, we investigate to what extent the G2T generation problem is solved for previously studied datasets, and how pro- posed metrics perform when comparing generated texts. To help address their limitations, we propose a new metric that correctly identifies factual faithfulness, i.e., given a triple (subject, predicate, object), it decides if the triple is present in a generated text. We show that our metric FactSpotter achieves the highest correlation with human annotations on data correct- ness, data coverage, and relevance. In addition, FactSpotter can be used as a plug-in feature to improve the factual faithfulness of existing models. Finally, we investigate if existing G2T datasets are still challenging for state-of-the-art models. Our code is available online: https://github.com/guihuzhang/FactSpotter. | [
"Zhang, Kun",
"Balalau, Oana",
"Manolescu, Ioana"
] | FactSpotter: Evaluating the Factual Faithfulness of Graph-to-Text Generation | findings-emnlp.672 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.673.bib | https://aclanthology.org/2023.findings-emnlp.673/ | @inproceedings{zhang-etal-2023-layoutdit,
title = "{L}ayout{DIT}: Layout-Aware End-to-End Document Image Translation with Multi-Step Conductive Decoder",
author = "Zhang, Zhiyang and
Zhang, Yaping and
Liang, Yupu and
Xiang, Lu and
Zhao, Yang and
Zhou, Yu and
Zong, Chengqing",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.673",
doi = "10.18653/v1/2023.findings-emnlp.673",
pages = "10043--10053",
abstract = "Document image translation (DIT) aims to translate text embedded in images from one language to another. It is a challenging task that needs to understand visual layout with text semantics simultaneously. However, existing methods struggle to capture the crucial visual layout in real-world complex document images. In this work, we make the first attempt to incorporate layout knowledge into DIT in an end-to-end way. Specifically, we propose a novel Layout-aware end-to-end Document Image Translation (LayoutDIT) with multi-step conductive decoder. A layout-aware encoder is first introduced to model visual layout relations with raw OCR results. Then a novel multi-step conductive decoder is unified with hidden states conduction across three step-decoders to achieve the document translation step by step. Benefiting from the layout-aware end-to-end joint training, our LayoutDIT outperforms state-of-the-art methods with better parameter efficiency. Besides, we create a new multi-domain document image translation dataset to validate the model{'}s generalization. Extensive experiments show that LayoutDIT has a good generalization in diverse and complex layout scenes.",
}
| Document image translation (DIT) aims to translate text embedded in images from one language to another. It is a challenging task that needs to understand visual layout with text semantics simultaneously. However, existing methods struggle to capture the crucial visual layout in real-world complex document images. In this work, we make the first attempt to incorporate layout knowledge into DIT in an end-to-end way. Specifically, we propose a novel Layout-aware end-to-end Document Image Translation (LayoutDIT) with multi-step conductive decoder. A layout-aware encoder is first introduced to model visual layout relations with raw OCR results. Then a novel multi-step conductive decoder is unified with hidden states conduction across three step-decoders to achieve the document translation step by step. Benefiting from the layout-aware end-to-end joint training, our LayoutDIT outperforms state-of-the-art methods with better parameter efficiency. Besides, we create a new multi-domain document image translation dataset to validate the model{'}s generalization. Extensive experiments show that LayoutDIT has a good generalization in diverse and complex layout scenes. | [
"Zhang, Zhiyang",
"Zhang, Yaping",
"Liang, Yupu",
"Xiang, Lu",
"Zhao, Yang",
"Zhou, Yu",
"Zong, Chengqing"
] | LayoutDIT: Layout-Aware End-to-End Document Image Translation with Multi-Step Conductive Decoder | findings-emnlp.673 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.674.bib | https://aclanthology.org/2023.findings-emnlp.674/ | @inproceedings{mircea-cheung-2023-balaur,
title = "Balaur: Language Model Pretraining with Lexical Semantic Relations",
author = "Mircea, Andrei and
Cheung, Jackie",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.674",
doi = "10.18653/v1/2023.findings-emnlp.674",
pages = "10054--10070",
abstract = "Lexical semantic relations (LSRs) characterize meaning relationships between words and play an important role in systematic generalization on lexical inference tasks. Notably, several tasks that require knowledge of hypernymy still pose a challenge for pretrained language models (LMs) such as BERT, underscoring the need to better align their linguistic behavior with our knowledge of LSRs. In this paper, we propose Balaur, a model that addresses this challenge by modeling LSRs directly in the LM{'}s hidden states throughout pretraining. Motivating our approach is the hypothesis that the internal representations of LMs can provide an interface to their observable linguistic behavior, and that by controlling one we can influence the other. We validate our hypothesis and demonstrate that Balaur generally improves the performance of large transformer-based LMs on a comprehensive set of hypernymy-informed tasks, as well as on the original LM objective. Code and data are made available at https://github.com/mirandrom/balaur",
}
| Lexical semantic relations (LSRs) characterize meaning relationships between words and play an important role in systematic generalization on lexical inference tasks. Notably, several tasks that require knowledge of hypernymy still pose a challenge for pretrained language models (LMs) such as BERT, underscoring the need to better align their linguistic behavior with our knowledge of LSRs. In this paper, we propose Balaur, a model that addresses this challenge by modeling LSRs directly in the LM{'}s hidden states throughout pretraining. Motivating our approach is the hypothesis that the internal representations of LMs can provide an interface to their observable linguistic behavior, and that by controlling one we can influence the other. We validate our hypothesis and demonstrate that Balaur generally improves the performance of large transformer-based LMs on a comprehensive set of hypernymy-informed tasks, as well as on the original LM objective. Code and data are made available at https://github.com/mirandrom/balaur | [
"Mircea, Andrei",
"Cheung, Jackie"
] | Balaur: Language Model Pretraining with Lexical Semantic Relations | findings-emnlp.674 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.675.bib | https://aclanthology.org/2023.findings-emnlp.675/ | @inproceedings{chen-etal-2023-exploring-context,
title = "Exploring In-Context Learning for Knowledge Grounded Dialog Generation",
author = "Chen, Qinyu and
Wu, Wenhao and
Li, Sujian",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.675",
doi = "10.18653/v1/2023.findings-emnlp.675",
pages = "10071--10081",
abstract = "Large neural-based dialog generation models have been applied in many real-life scenarios, yet they are prone to hallucination and tend to produce factually inaccurate outputs which raise great concerns. To alleviate this problem, we propose a plug-and-play retrieval-based framework IKA, which leverages in-context learning and retrieval techniques to enhance LLMs on knowledge grounded dialog generation. We design thorough experiments on a large-scale knowledge graph with 1M+ facts to investigate the effectiveness and generalization of our framework. Experiments show that our method surpasses previous training-based SOTA by a large margin, specifically 46.67{\%} in BLEU4, 26.01{\%} in ROUGE-L, 122.90{\%} in BARTScore and 30.50{\%} in Entity Coverage F1. Further analysis show promising abilities of LLMs to perform knowledge-intensive tasks, which is previously considered weak and understudied.",
}
| Large neural-based dialog generation models have been applied in many real-life scenarios, yet they are prone to hallucination and tend to produce factually inaccurate outputs which raise great concerns. To alleviate this problem, we propose a plug-and-play retrieval-based framework IKA, which leverages in-context learning and retrieval techniques to enhance LLMs on knowledge grounded dialog generation. We design thorough experiments on a large-scale knowledge graph with 1M+ facts to investigate the effectiveness and generalization of our framework. Experiments show that our method surpasses previous training-based SOTA by a large margin, specifically 46.67{\%} in BLEU4, 26.01{\%} in ROUGE-L, 122.90{\%} in BARTScore and 30.50{\%} in Entity Coverage F1. Further analysis show promising abilities of LLMs to perform knowledge-intensive tasks, which is previously considered weak and understudied. | [
"Chen, Qinyu",
"Wu, Wenhao",
"Li, Sujian"
] | Exploring In-Context Learning for Knowledge Grounded Dialog Generation | findings-emnlp.675 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.676.bib | https://aclanthology.org/2023.findings-emnlp.676/ | @inproceedings{wu-etal-2023-towards-enhancing,
title = "Towards Enhancing Relational Rules for Knowledge Graph Link Prediction",
author = "Wu, Shuhan and
Wan, Huaiyu and
Chen, Wei and
Wu, Yuting and
Shen, Junfeng and
Lin, Youfang",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.676",
doi = "10.18653/v1/2023.findings-emnlp.676",
pages = "10082--10097",
abstract = "Graph neural networks (GNNs) have shown promising performance for knowledge graph reasoning. A recent variant of GNN called progressive relational graph neural network (PRGNN), utilizes relational rules to infer missing knowledge in relational digraphs and achieves notable results. However, during reasoning with PRGNN, two important properties are often overlooked: (1) the sequentiality of relation composition, where the order of combining different relations affects the semantics of the relational rules, and (2) the lagged entity information propagation, where the transmission speed of required information lags behind the appearance speed of new entities. Ignoring these properties leads to incorrect relational rule learning and decreased reasoning accuracy. To address these issues, we propose a novel knowledge graph reasoning approach, the Relational rUle eNhanced Graph Neural Network (RUN-GNN). Specifically, RUN-GNN employs a query related fusion gate unit to model the sequentiality of relation composition and utilizes a buffering update mechanism to alleviate the negative effect of lagged entity information propagation, resulting in higher-quality relational rule learning. Experimental results on multiple datasets demonstrate the superiority of RUN-GNN is superior on both transductive and inductive link prediction tasks.",
}
| Graph neural networks (GNNs) have shown promising performance for knowledge graph reasoning. A recent variant of GNN called progressive relational graph neural network (PRGNN), utilizes relational rules to infer missing knowledge in relational digraphs and achieves notable results. However, during reasoning with PRGNN, two important properties are often overlooked: (1) the sequentiality of relation composition, where the order of combining different relations affects the semantics of the relational rules, and (2) the lagged entity information propagation, where the transmission speed of required information lags behind the appearance speed of new entities. Ignoring these properties leads to incorrect relational rule learning and decreased reasoning accuracy. To address these issues, we propose a novel knowledge graph reasoning approach, the Relational rUle eNhanced Graph Neural Network (RUN-GNN). Specifically, RUN-GNN employs a query related fusion gate unit to model the sequentiality of relation composition and utilizes a buffering update mechanism to alleviate the negative effect of lagged entity information propagation, resulting in higher-quality relational rule learning. Experimental results on multiple datasets demonstrate the superiority of RUN-GNN is superior on both transductive and inductive link prediction tasks. | [
"Wu, Shuhan",
"Wan, Huaiyu",
"Chen, Wei",
"Wu, Yuting",
"Shen, Junfeng",
"Lin, Youfang"
] | Towards Enhancing Relational Rules for Knowledge Graph Link Prediction | findings-emnlp.676 | 2310.13411 | [
"https://github.com/ninggirsu/run-gnn"
] | https://huggingface.co/papers/2310.13411 | 0 | 0 | 0 | 6 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.677.bib | https://aclanthology.org/2023.findings-emnlp.677/ | @inproceedings{zhu-etal-2023-nlp,
title = "Are {NLP} Models Good at Tracing Thoughts: An Overview of Narrative Understanding",
author = "Zhu, Lixing and
Zhao, Runcong and
Gui, Lin and
He, Yulan",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.677",
doi = "10.18653/v1/2023.findings-emnlp.677",
pages = "10098--10121",
abstract = "Narrative understanding involves capturing the author{'}s cognitive processes, providing insights into their knowledge, intentions, beliefs, and desires. Although large language models (LLMs) excel in generating grammatically coherent text, their ability to comprehend the author{'}s thoughts remains uncertain. This limitation hinders the practical applications of narrative understanding. In this paper, we conduct a comprehensive survey of narrative understanding tasks, thoroughly examining their key features, definitions, taxonomy, associated datasets, training objectives, evaluation metrics, and limitations. Furthermore, we explore the potential of expanding the capabilities of modularized LLMs to address novel narrative understanding tasks. By framing narrative understanding as the retrieval of the author{'}s imaginative cues that outline the narrative structure, our study introduces a fresh perspective on enhancing narrative comprehension.",
}
| Narrative understanding involves capturing the author{'}s cognitive processes, providing insights into their knowledge, intentions, beliefs, and desires. Although large language models (LLMs) excel in generating grammatically coherent text, their ability to comprehend the author{'}s thoughts remains uncertain. This limitation hinders the practical applications of narrative understanding. In this paper, we conduct a comprehensive survey of narrative understanding tasks, thoroughly examining their key features, definitions, taxonomy, associated datasets, training objectives, evaluation metrics, and limitations. Furthermore, we explore the potential of expanding the capabilities of modularized LLMs to address novel narrative understanding tasks. By framing narrative understanding as the retrieval of the author{'}s imaginative cues that outline the narrative structure, our study introduces a fresh perspective on enhancing narrative comprehension. | [
"Zhu, Lixing",
"Zhao, Runcong",
"Gui, Lin",
"He, Yulan"
] | Are NLP Models Good at Tracing Thoughts: An Overview of Narrative Understanding | findings-emnlp.677 | 2310.18783 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.678.bib | https://aclanthology.org/2023.findings-emnlp.678/ | @inproceedings{qamar-etal-2023-speaking,
title = "Who is Speaking? Speaker-Aware Multiparty Dialogue Act Classification",
author = "Qamar, Ayesha and
Pyarelal, Adarsh and
Huang, Ruihong",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.678",
doi = "10.18653/v1/2023.findings-emnlp.678",
pages = "10122--10135",
abstract = "Utterances do not occur in isolation in dialogues; it is essential to have the information of who the speaker of an utterance is to be able to recover the speaker{'}s intention with respect to the surrounding context. Beyond simply capturing speaker switches, identifying how speakers interact with each other in a dialogue is crucial to understanding conversational flow. This becomes increasingly important and simultaneously difficult to model when more than two interlocutors take part in a conversation. To overcome this challenge, we propose to explicitly add speaker awareness to each utterance representation. To that end, we use a graph neural network to model how each speaker is behaving within the local context of a conversation. The speaker representations learned this way are then used to update their respective utterance representations. We experiment with both multiparticipant and dyadic conversations on the MRDA and SwDA datasets and show the effectiveness of our approach.",
}
| Utterances do not occur in isolation in dialogues; it is essential to have the information of who the speaker of an utterance is to be able to recover the speaker{'}s intention with respect to the surrounding context. Beyond simply capturing speaker switches, identifying how speakers interact with each other in a dialogue is crucial to understanding conversational flow. This becomes increasingly important and simultaneously difficult to model when more than two interlocutors take part in a conversation. To overcome this challenge, we propose to explicitly add speaker awareness to each utterance representation. To that end, we use a graph neural network to model how each speaker is behaving within the local context of a conversation. The speaker representations learned this way are then used to update their respective utterance representations. We experiment with both multiparticipant and dyadic conversations on the MRDA and SwDA datasets and show the effectiveness of our approach. | [
"Qamar, Ayesha",
"Pyarelal, Adarsh",
"Huang, Ruihong"
] | Who is Speaking? Speaker-Aware Multiparty Dialogue Act Classification | findings-emnlp.678 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.679.bib | https://aclanthology.org/2023.findings-emnlp.679/ | @inproceedings{gonen-etal-2023-demystifying,
title = "Demystifying Prompts in Language Models via Perplexity Estimation",
author = "Gonen, Hila and
Iyer, Srini and
Blevins, Terra and
Smith, Noah and
Zettlemoyer, Luke",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.679",
doi = "10.18653/v1/2023.findings-emnlp.679",
pages = "10136--10148",
abstract = "Language models can be prompted to perform a wide variety of tasks with zero- and few-shot in-context learning. However, performance varies significantly with the choice of prompt, and we do not yet understand why this happens. In this paper, we analyze the factors that contribute to this variance and establish a new empirical hypothesis: the performance of a prompt is predicted by the extent to which the model is familiar with the language it contains. Over a wide range of tasks, we show that the lower the perplexity of the prompt, the better it is able to perform the task, when considering reasonable prompts that are related to it. As part of our analysis, we also devise a method to automatically extend a small seed set of manually written prompts by paraphrasing with GPT3 and backtranslation. This larger set allows us to verify that perplexity is a strong predictor of the success of a prompt and we show that the lowest perplexity prompts are consistently effective.",
}
| Language models can be prompted to perform a wide variety of tasks with zero- and few-shot in-context learning. However, performance varies significantly with the choice of prompt, and we do not yet understand why this happens. In this paper, we analyze the factors that contribute to this variance and establish a new empirical hypothesis: the performance of a prompt is predicted by the extent to which the model is familiar with the language it contains. Over a wide range of tasks, we show that the lower the perplexity of the prompt, the better it is able to perform the task, when considering reasonable prompts that are related to it. As part of our analysis, we also devise a method to automatically extend a small seed set of manually written prompts by paraphrasing with GPT3 and backtranslation. This larger set allows us to verify that perplexity is a strong predictor of the success of a prompt and we show that the lowest perplexity prompts are consistently effective. | [
"Gonen, Hila",
"Iyer, Srini",
"Blevins, Terra",
"Smith, Noah",
"Zettlemoyer, Luke"
] | Demystifying Prompts in Language Models via Perplexity Estimation | findings-emnlp.679 | 2212.04037 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.680.bib | https://aclanthology.org/2023.findings-emnlp.680/ | @inproceedings{wang-etal-2023-c2d2,
title = "{C}2{D}2 Dataset: A Resource for the Cognitive Distortion Analysis and Its Impact on Mental Health",
author = "Wang, Bichen and
Deng, Pengfei and
Zhao, Yanyan and
Qin, Bing",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.680",
doi = "10.18653/v1/2023.findings-emnlp.680",
pages = "10149--10160",
abstract = "Cognitive distortions refer to patterns of irrational thinking that can lead to distorted perceptions of reality and mental health problems in individuals. Despite previous attempts to detect cognitive distortion through language, progress has been slow due to the lack of appropriate data. In this paper, we present the C2D2 dataset, the first expert-supervised \textbf{C}hinese \textbf{C}ognitive \textbf{D}istortion \textbf{D}ataset, which contains 7,500 cognitive distortion thoughts in everyday life scenes. Additionally, we examine the presence of cognitive distortions in social media texts shared by individuals diagnosed with mental disorders, providing insights into the association between cognitive distortions and mental health conditions. We propose that incorporating information about users{'} cognitive distortions can enhance the performance of existing models mental disorder detection. We contribute to a better understanding of how cognitive distortions appear in individuals{'} language and their impact on mental health.",
}
| Cognitive distortions refer to patterns of irrational thinking that can lead to distorted perceptions of reality and mental health problems in individuals. Despite previous attempts to detect cognitive distortion through language, progress has been slow due to the lack of appropriate data. In this paper, we present the C2D2 dataset, the first expert-supervised \textbf{C}hinese \textbf{C}ognitive \textbf{D}istortion \textbf{D}ataset, which contains 7,500 cognitive distortion thoughts in everyday life scenes. Additionally, we examine the presence of cognitive distortions in social media texts shared by individuals diagnosed with mental disorders, providing insights into the association between cognitive distortions and mental health conditions. We propose that incorporating information about users{'} cognitive distortions can enhance the performance of existing models mental disorder detection. We contribute to a better understanding of how cognitive distortions appear in individuals{'} language and their impact on mental health. | [
"Wang, Bichen",
"Deng, Pengfei",
"Zhao, Yanyan",
"Qin, Bing"
] | C2D2 Dataset: A Resource for the Cognitive Distortion Analysis and Its Impact on Mental Health | findings-emnlp.680 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.681.bib | https://aclanthology.org/2023.findings-emnlp.681/ | @inproceedings{ye-etal-2023-mixedit,
title = "{M}ix{E}dit: Revisiting Data Augmentation and Beyond for Grammatical Error Correction",
author = "Ye, Jingheng and
Li, Yinghui and
Li, Yangning and
Zheng, Hai-Tao",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.681",
doi = "10.18653/v1/2023.findings-emnlp.681",
pages = "10161--10175",
abstract = "Data Augmentation through generating pseudo data has been proven effective in mitigating the challenge of data scarcity in the field of Grammatical Error Correction (GEC). Various augmentation strategies have been widely explored, most of which are motivated by two heuristics, i.e., increasing the distribution similarity and diversity of pseudo data. However, the underlying mechanism responsible for the effectiveness of these strategies remains poorly understood. In this paper, we aim to clarify how data augmentation improves GEC models. To this end, we introduce two interpretable and computationally efficient measures: Affinity and Diversity. Our findings indicate that an excellent GEC data augmentation strategy characterized by high Affinity and appropriate Diversity can better improve the performance of GEC models. Based on this observation, we propose MixEdit, a data augmentation approach that strategically and dynamically augments realistic data, without requiring extra monolingual corpora. To verify the correctness of our findings and the effectiveness of the proposed MixEdit, we conduct experiments on mainstream English and Chinese GEC datasets. The results show that MixEdit substantially improves GEC models and is complementary to traditional data augmentation methods. All the source codes of MixEdit are released at https://github.com/THUKElab/MixEdit.",
}
| Data Augmentation through generating pseudo data has been proven effective in mitigating the challenge of data scarcity in the field of Grammatical Error Correction (GEC). Various augmentation strategies have been widely explored, most of which are motivated by two heuristics, i.e., increasing the distribution similarity and diversity of pseudo data. However, the underlying mechanism responsible for the effectiveness of these strategies remains poorly understood. In this paper, we aim to clarify how data augmentation improves GEC models. To this end, we introduce two interpretable and computationally efficient measures: Affinity and Diversity. Our findings indicate that an excellent GEC data augmentation strategy characterized by high Affinity and appropriate Diversity can better improve the performance of GEC models. Based on this observation, we propose MixEdit, a data augmentation approach that strategically and dynamically augments realistic data, without requiring extra monolingual corpora. To verify the correctness of our findings and the effectiveness of the proposed MixEdit, we conduct experiments on mainstream English and Chinese GEC datasets. The results show that MixEdit substantially improves GEC models and is complementary to traditional data augmentation methods. All the source codes of MixEdit are released at https://github.com/THUKElab/MixEdit. | [
"Ye, Jingheng",
"Li, Yinghui",
"Li, Yangning",
"Zheng, Hai-Tao"
] | MixEdit: Revisiting Data Augmentation and Beyond for Grammatical Error Correction | findings-emnlp.681 | 2310.11671 | [
"https://github.com/thukelab/mixedit"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.682.bib | https://aclanthology.org/2023.findings-emnlp.682/ | @inproceedings{lou-etal-2023-cceval,
title = "{CCE}val: A Representative Evaluation Benchmark for the {C}hinese-centric Multilingual Machine Translation",
author = "Lou, Lianzhang and
Yin, Xi and
Xie, Yutao and
Xiang, Yang",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.682",
doi = "10.18653/v1/2023.findings-emnlp.682",
pages = "10176--10184",
abstract = "The Chinese-centric Multilingual Machine Translation (MMT) has gained more importance recently due to increasing demands from international business development and cross-cultural exchanges. However, an important factor that limits the progress of this area is the lack of highly representative and high-quality evaluation benchmarks. To fill this gap, we propose CCEval, an impartial and representative Chinese-centric MMT evaluation dataset. This benchmark dataset consists of 2500 Chinese sentences we meticulously selected and processed, and covers more diverse linguistic features as compared to other MMT evaluation benchmarks. These sentences have been translated into 11 languages of various resource levels by professional translators via a rigorously controlled process pipeline to ensure their high quality. We conduct experiments to demonstrate our sampling methodology{'}s effectiveness in constructing evaluation datasets strongly correlated with human evaluations. The resulting dataset enables better assessments of the Chinese-centric MMT quality. Our CCEval benchmark dataset is available at https://bright.pcl.ac.cn/en/offlineTasks.",
}
| The Chinese-centric Multilingual Machine Translation (MMT) has gained more importance recently due to increasing demands from international business development and cross-cultural exchanges. However, an important factor that limits the progress of this area is the lack of highly representative and high-quality evaluation benchmarks. To fill this gap, we propose CCEval, an impartial and representative Chinese-centric MMT evaluation dataset. This benchmark dataset consists of 2500 Chinese sentences we meticulously selected and processed, and covers more diverse linguistic features as compared to other MMT evaluation benchmarks. These sentences have been translated into 11 languages of various resource levels by professional translators via a rigorously controlled process pipeline to ensure their high quality. We conduct experiments to demonstrate our sampling methodology{'}s effectiveness in constructing evaluation datasets strongly correlated with human evaluations. The resulting dataset enables better assessments of the Chinese-centric MMT quality. Our CCEval benchmark dataset is available at https://bright.pcl.ac.cn/en/offlineTasks. | [
"Lou, Lianzhang",
"Yin, Xi",
"Xie, Yutao",
"Xiang, Yang"
] | CCEval: A Representative Evaluation Benchmark for the Chinese-centric Multilingual Machine Translation | findings-emnlp.682 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.683.bib | https://aclanthology.org/2023.findings-emnlp.683/ | @inproceedings{zhou-etal-2023-rome,
title = "{ROME}: Evaluating Pre-trained Vision-Language Models on Reasoning beyond Visual Common Sense",
author = "Zhou, Kankan and
Lai, Eason and
Yeong, Wei Bin Au and
Mouratidis, Kyriakos and
Jiang, Jing",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.683",
doi = "10.18653/v1/2023.findings-emnlp.683",
pages = "10185--10197",
abstract = "Humans possess a strong capability for reasoning beyond common sense. For example, given an unconventional image of a goldfish laying on the table next to an empty fishbowl, a human would effortlessly determine that the fish is not inside the fishbowl. The case, however, may be different for a vision-language model, whose reasoning could gravitate towards the common scenario that the fish is inside the bowl, despite the visual input. In this paper, we introduce a novel probing dataset named ROME (reasoning beyond commonsense knowledge) to evaluate whether the state-of-the-art pre-trained vision-language models have the reasoning capability to correctly interpret counter-intuitive content. ROME contains images that defy commonsense knowledge with regards to color, shape, material, size and positional relation. Experiments on the state-of-the-art pre-trained vision-language models reveal that most of these models are still largely incapable of interpreting counter-intuitive scenarios. We hope that ROME will spur further investigations on reasoning beyond commonsense knowledge in vision-language research.",
}
| Humans possess a strong capability for reasoning beyond common sense. For example, given an unconventional image of a goldfish laying on the table next to an empty fishbowl, a human would effortlessly determine that the fish is not inside the fishbowl. The case, however, may be different for a vision-language model, whose reasoning could gravitate towards the common scenario that the fish is inside the bowl, despite the visual input. In this paper, we introduce a novel probing dataset named ROME (reasoning beyond commonsense knowledge) to evaluate whether the state-of-the-art pre-trained vision-language models have the reasoning capability to correctly interpret counter-intuitive content. ROME contains images that defy commonsense knowledge with regards to color, shape, material, size and positional relation. Experiments on the state-of-the-art pre-trained vision-language models reveal that most of these models are still largely incapable of interpreting counter-intuitive scenarios. We hope that ROME will spur further investigations on reasoning beyond commonsense knowledge in vision-language research. | [
"Zhou, Kankan",
"Lai, Eason",
"Yeong, Wei Bin Au",
"Mouratidis, Kyriakos",
"Jiang, Jing"
] | ROME: Evaluating Pre-trained Vision-Language Models on Reasoning beyond Visual Common Sense | findings-emnlp.683 | 2310.19301 | [
"https://github.com/k-square-00/rome"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.684.bib | https://aclanthology.org/2023.findings-emnlp.684/ | @inproceedings{guo-etal-2023-automatic,
title = "Automatic Analysis of Substantiation in Scientific Peer Reviews",
author = "Guo, Yanzhu and
Shang, Guokan and
Rennard, Virgile and
Vazirgiannis, Michalis and
Clavel, Chlo{\'e}",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.684",
doi = "10.18653/v1/2023.findings-emnlp.684",
pages = "10198--10216",
abstract = "With the increasing amount of problematic peer reviews in top AI conferences, the community is urgently in need of automatic quality control measures. In this paper, we restrict our attention to substantiation {---} one popular quality aspect indicating whether the claims in a review are sufficiently supported by evidence {---} and provide a solution automatizing this evaluation process. To achieve this goal, we first formulate the problem as claim-evidence pair extraction in scientific peer reviews, and collect SubstanReview, the first annotated dataset for this task. SubstanReview consists of 550 reviews from NLP conferences annotated by domain experts. On the basis of this dataset, we train an argument mining system to automatically analyze the level of substantiation in peer reviews. We also perform data analysis on the SubstanReview dataset to obtain meaningful insights on peer reviewing quality in NLP conferences over recent years. The dataset is available at https://github.com/YanzhuGuo/SubstanReview.",
}
| With the increasing amount of problematic peer reviews in top AI conferences, the community is urgently in need of automatic quality control measures. In this paper, we restrict our attention to substantiation {---} one popular quality aspect indicating whether the claims in a review are sufficiently supported by evidence {---} and provide a solution automatizing this evaluation process. To achieve this goal, we first formulate the problem as claim-evidence pair extraction in scientific peer reviews, and collect SubstanReview, the first annotated dataset for this task. SubstanReview consists of 550 reviews from NLP conferences annotated by domain experts. On the basis of this dataset, we train an argument mining system to automatically analyze the level of substantiation in peer reviews. We also perform data analysis on the SubstanReview dataset to obtain meaningful insights on peer reviewing quality in NLP conferences over recent years. The dataset is available at https://github.com/YanzhuGuo/SubstanReview. | [
"Guo, Yanzhu",
"Shang, Guokan",
"Rennard, Virgile",
"Vazirgiannis, Michalis",
"Clavel, Chlo{\\'e}"
] | Automatic Analysis of Substantiation in Scientific Peer Reviews | findings-emnlp.684 | 2311.11967 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.685.bib | https://aclanthology.org/2023.findings-emnlp.685/ | @inproceedings{lo-etal-2023-hierarchical,
title = "Hierarchical Prompting Assists Large Language Model on Web Navigation",
author = "Lo, Robert and
Sridhar, Abishek and
Xu, Frank and
Zhu, Hao and
Zhou, Shuyan",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.685",
doi = "10.18653/v1/2023.findings-emnlp.685",
pages = "10217--10244",
abstract = "Large language models (LLMs) struggle on processing complicated observations in interactive decision making. To alleviate this issue, we propose a simple hierarchical prompting approach. Diverging from previous prompting approaches that always put the full observation (a web page) to the prompt, we propose to first construct an action-aware observation which is more condensed and relevant with a dedicated Summarizer prompt. The Actor prompt then predicts the next action based on the summarized history. While our method has broad applicability, we particularly demonstrate its efficacy in the complex domain of web navigation where a full observation often contains redundant and irrelevant information. Our approach outperforms the previous state-of-the-art prompting mechanism with the same LLM by 6.2{\%} on task success rate, demonstrating its potential on interactive decision making tasks with long observation traces.",
}
| Large language models (LLMs) struggle on processing complicated observations in interactive decision making. To alleviate this issue, we propose a simple hierarchical prompting approach. Diverging from previous prompting approaches that always put the full observation (a web page) to the prompt, we propose to first construct an action-aware observation which is more condensed and relevant with a dedicated Summarizer prompt. The Actor prompt then predicts the next action based on the summarized history. While our method has broad applicability, we particularly demonstrate its efficacy in the complex domain of web navigation where a full observation often contains redundant and irrelevant information. Our approach outperforms the previous state-of-the-art prompting mechanism with the same LLM by 6.2{\%} on task success rate, demonstrating its potential on interactive decision making tasks with long observation traces. | [
"Lo, Robert",
"Sridhar, Abishek",
"Xu, Frank",
"Zhu, Hao",
"Zhou, Shuyan"
] | Hierarchical Prompting Assists Large Language Model on Web Navigation | findings-emnlp.685 | 2305.14257 | [
"https://github.com/robert1003/ash-prompting"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.686.bib | https://aclanthology.org/2023.findings-emnlp.686/ | @inproceedings{laskar-etal-2023-large,
title = "Can Large Language Models Fix Data Annotation Errors? An Empirical Study Using Debatepedia for Query-Focused Text Summarization",
author = "Laskar, Md Tahmid Rahman and
Rahman, Mizanur and
Jahan, Israt and
Hoque, Enamul and
Huang, Jimmy",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.686",
doi = "10.18653/v1/2023.findings-emnlp.686",
pages = "10245--10255",
abstract = "Debatepedia is a publicly available dataset consisting of arguments and counter-arguments on controversial topics that has been widely used for the single-document query-focused abstractive summarization task in recent years. However, it has been recently found that this dataset is limited by noise and even most queries in this dataset do not have any relevance to the respective document. In this paper, we study whether large language models (LLMs) can be utilized to clean the Debatepedia dataset to make it suitable for query-focused abstractive summarization. More specifically, we harness the language generation capabilities of two LLMs, namely, ChatGPT and PaLM to regenerate its queries. Based on our experiments, we find that solely depending on large language models for query correction may not be very useful for data cleaning. However, we observe that leveraging a rule-based approach for data sampling followed by query regeneration using LLMs (especially ChatGPT) for the sampled instances may ensure a higher quality version of this dataset suitable for the development of more generalized query-focused text summarization models.",
}
| Debatepedia is a publicly available dataset consisting of arguments and counter-arguments on controversial topics that has been widely used for the single-document query-focused abstractive summarization task in recent years. However, it has been recently found that this dataset is limited by noise and even most queries in this dataset do not have any relevance to the respective document. In this paper, we study whether large language models (LLMs) can be utilized to clean the Debatepedia dataset to make it suitable for query-focused abstractive summarization. More specifically, we harness the language generation capabilities of two LLMs, namely, ChatGPT and PaLM to regenerate its queries. Based on our experiments, we find that solely depending on large language models for query correction may not be very useful for data cleaning. However, we observe that leveraging a rule-based approach for data sampling followed by query regeneration using LLMs (especially ChatGPT) for the sampled instances may ensure a higher quality version of this dataset suitable for the development of more generalized query-focused text summarization models. | [
"Laskar, Md Tahmid Rahman",
"Rahman, Mizanur",
"Jahan, Israt",
"Hoque, Enamul",
"Huang, Jimmy"
] | Can Large Language Models Fix Data Annotation Errors? An Empirical Study Using Debatepedia for Query-Focused Text Summarization | findings-emnlp.686 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.687.bib | https://aclanthology.org/2023.findings-emnlp.687/ | @inproceedings{khatry-etal-2023-tstr,
title = "{TSTR}: Target Similarity Tuning Meets the Real World",
author = "Khatry, Anirudh and
Gulwani, Sumit and
Gupta, Priyanshu and
Le, Vu and
Singh, Mukul and
Singha, Ananya and
Verbruggen, Gust",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.687",
doi = "10.18653/v1/2023.findings-emnlp.687",
pages = "10256--10261",
abstract = "Target similarity tuning (TST) is a method of selecting relevant examples in natural language (NL) to code generation through large language models (LLMs) to improve performance. Its goal is to adapt a sentence embedding model to have the similarity between two NL inputs match the similarity between their associated code outputs. In this paper, we propose different methods to apply and improve TST in the real world. First, we replace the sentence transformer with embeddings from a larger model, which reduces sensitivity to the language distribution and thus provides more flexibility in synthetic generation of examples, and we train a tiny model that transforms these embeddings to a space where embedding similarity matches code similarity, which allows the model to remain a black box and only requires a few matrix multiplications at inference time. Second, we how to efficiently select a smaller number of training examples to train the TST model. Third, we introduce a ranking-based evaluation for TST that does not require end-to-end code generation experiments, which can be expensive to perform.",
}
| Target similarity tuning (TST) is a method of selecting relevant examples in natural language (NL) to code generation through large language models (LLMs) to improve performance. Its goal is to adapt a sentence embedding model to have the similarity between two NL inputs match the similarity between their associated code outputs. In this paper, we propose different methods to apply and improve TST in the real world. First, we replace the sentence transformer with embeddings from a larger model, which reduces sensitivity to the language distribution and thus provides more flexibility in synthetic generation of examples, and we train a tiny model that transforms these embeddings to a space where embedding similarity matches code similarity, which allows the model to remain a black box and only requires a few matrix multiplications at inference time. Second, we how to efficiently select a smaller number of training examples to train the TST model. Third, we introduce a ranking-based evaluation for TST that does not require end-to-end code generation experiments, which can be expensive to perform. | [
"Khatry, Anirudh",
"Gulwani, Sumit",
"Gupta, Priyanshu",
"Le, Vu",
"Singh, Mukul",
"Singha, Ananya",
"Verbruggen, Gust"
] | TSTR: Target Similarity Tuning Meets the Real World | findings-emnlp.687 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.688.bib | https://aclanthology.org/2023.findings-emnlp.688/ | @inproceedings{zhou-etal-2023-realbehavior,
title = "{R}eal{B}ehavior: A Framework for Faithfully Characterizing Foundation Models{'} Human-like Behavior Mechanisms",
author = "Zhou, Enyu and
Zheng, Rui and
Xi, Zhiheng and
Gao, Songyang and
Fan, Xiaoran and
Fei, Zichu and
Ye, Jingting and
Gui, Tao and
Zhang, Qi and
Huang, Xuanjing",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.688",
doi = "10.18653/v1/2023.findings-emnlp.688",
pages = "10262--10274",
abstract = "Reports of human-like behaviors in foundation models are growing, with psychological theories providing enduring tools to investigate these behaviors. However, current research tends to directly apply these human-oriented tools without verifying the faithfulness of their outcomes. In this paper, we introduce a framework, RealBehavior, which is designed to characterize the humanoid behaviors of models faithfully. Beyond simply measuring behaviors, our framework assesses the faithfulness of results based on reproducibility, internal and external consistency, and generalizability. Our findings suggest that a simple application of psychological tools cannot faithfully characterize all human-like behaviors. Moreover, we discuss the impacts of aligning models with human and social values, arguing for the necessity of diversifying alignment objectives to prevent the creation of models with restricted characteristics.",
}
| Reports of human-like behaviors in foundation models are growing, with psychological theories providing enduring tools to investigate these behaviors. However, current research tends to directly apply these human-oriented tools without verifying the faithfulness of their outcomes. In this paper, we introduce a framework, RealBehavior, which is designed to characterize the humanoid behaviors of models faithfully. Beyond simply measuring behaviors, our framework assesses the faithfulness of results based on reproducibility, internal and external consistency, and generalizability. Our findings suggest that a simple application of psychological tools cannot faithfully characterize all human-like behaviors. Moreover, we discuss the impacts of aligning models with human and social values, arguing for the necessity of diversifying alignment objectives to prevent the creation of models with restricted characteristics. | [
"Zhou, Enyu",
"Zheng, Rui",
"Xi, Zhiheng",
"Gao, Songyang",
"Fan, Xiaoran",
"Fei, Zichu",
"Ye, Jingting",
"Gui, Tao",
"Zhang, Qi",
"Huang, Xuanjing"
] | RealBehavior: A Framework for Faithfully Characterizing Foundation Models' Human-like Behavior Mechanisms | findings-emnlp.688 | 2310.11227 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.689.bib | https://aclanthology.org/2023.findings-emnlp.689/ | @inproceedings{wambsganss-etal-2023-unraveling,
title = "Unraveling Downstream Gender Bias from Large Language Models: A Study on {AI} Educational Writing Assistance",
author = {Wambsganss, Thiemo and
Su, Xiaotian and
Swamy, Vinitra and
Neshaei, Seyed and
Rietsche, Roman and
K{\"a}ser, Tanja},
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.689",
doi = "10.18653/v1/2023.findings-emnlp.689",
pages = "10275--10288",
abstract = "Large Language Models (LLMs) are increasingly utilized in educational tasks such as providing writing suggestions to students. Despite their potential, LLMs are known to harbor inherent biases which may negatively impact learners. Previous studies have investigated bias in models and data representations separately, neglecting the potential impact of LLM bias on human writing. In this paper, we investigate how bias transfers through an AI writing support pipeline. We conduct a large-scale user study with 231 students writing business case peer reviews in German. Students are divided into five groups with different levels of writing support: one in-classroom group with recommender system feature-based suggestions and four groups recruited from Prolific {--} a control group with no assistance, two groups with suggestions from fine-tuned GPT-2 and GPT-3 models, and one group with suggestions from pre-trained GPT-3.5. Using GenBit gender bias analysis and Word Embedding Association Tests (WEAT), we evaluate the gender bias at various stages of the pipeline: in reviews written by students, in suggestions generated by the models, and in model embeddings directly. Our results demonstrate that there is no significant difference in gender bias between the resulting peer reviews of groups with and without LLM suggestions. Our research is therefore optimistic about the use of AI writing support in the classroom, showcasing a context where bias in LLMs does not transfer to students{'} responses.",
}
| Large Language Models (LLMs) are increasingly utilized in educational tasks such as providing writing suggestions to students. Despite their potential, LLMs are known to harbor inherent biases which may negatively impact learners. Previous studies have investigated bias in models and data representations separately, neglecting the potential impact of LLM bias on human writing. In this paper, we investigate how bias transfers through an AI writing support pipeline. We conduct a large-scale user study with 231 students writing business case peer reviews in German. Students are divided into five groups with different levels of writing support: one in-classroom group with recommender system feature-based suggestions and four groups recruited from Prolific {--} a control group with no assistance, two groups with suggestions from fine-tuned GPT-2 and GPT-3 models, and one group with suggestions from pre-trained GPT-3.5. Using GenBit gender bias analysis and Word Embedding Association Tests (WEAT), we evaluate the gender bias at various stages of the pipeline: in reviews written by students, in suggestions generated by the models, and in model embeddings directly. Our results demonstrate that there is no significant difference in gender bias between the resulting peer reviews of groups with and without LLM suggestions. Our research is therefore optimistic about the use of AI writing support in the classroom, showcasing a context where bias in LLMs does not transfer to students{'} responses. | [
"Wambsganss, Thiemo",
"Su, Xiaotian",
"Swamy, Vinitra",
"Neshaei, Seyed",
"Rietsche, Roman",
"K{\\\"a}ser, Tanja"
] | Unraveling Downstream Gender Bias from Large Language Models: A Study on AI Educational Writing Assistance | findings-emnlp.689 | 2311.03311 | [
"https://github.com/epfl-ml4ed/unraveling-llm-bias"
] | https://huggingface.co/papers/2311.03311 | 0 | 0 | 0 | 6 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.690.bib | https://aclanthology.org/2023.findings-emnlp.690/ | @inproceedings{min-etal-2023-verve,
title = "{VERVE}: Template-based {R}eflecti{VE} Rewriting for {M}oti{V}ational {I}nt{E}rviewing",
author = "Min, Do June and
Perez-Rosas, Veronica and
Resnicow, Ken and
Mihalcea, Rada",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.690",
doi = "10.18653/v1/2023.findings-emnlp.690",
pages = "10289--10302",
abstract = "Reflective listening is a fundamental skill that counselors must acquire to achieve proficiency in motivational interviewing (MI). It involves responding in a manner that acknowledges and explores the meaning of what the client has expressed in the conversation. In this work, we introduce the task of counseling response rewriting, which transforms non-reflective statements into reflective responses. We introduce VERVE, a template-based rewriting system with paraphrase-augmented training and adaptive template updating. VERVE first creates a template by identifying and filtering out tokens that are not relevant to reflections and constructs a reflective response using the template. Paraphrase-augmented training allows the model to learn less-strict fillings of masked spans, and adaptive template updating helps discover effective templates for rewriting without significantly removing the original content. Using both automatic and human evaluations, we compare our method against text rewriting baselines and show that our framework is effective in turning non-reflective statements into more reflective responses while achieving a good content preservation-reflection style trade-off.",
}
| Reflective listening is a fundamental skill that counselors must acquire to achieve proficiency in motivational interviewing (MI). It involves responding in a manner that acknowledges and explores the meaning of what the client has expressed in the conversation. In this work, we introduce the task of counseling response rewriting, which transforms non-reflective statements into reflective responses. We introduce VERVE, a template-based rewriting system with paraphrase-augmented training and adaptive template updating. VERVE first creates a template by identifying and filtering out tokens that are not relevant to reflections and constructs a reflective response using the template. Paraphrase-augmented training allows the model to learn less-strict fillings of masked spans, and adaptive template updating helps discover effective templates for rewriting without significantly removing the original content. Using both automatic and human evaluations, we compare our method against text rewriting baselines and show that our framework is effective in turning non-reflective statements into more reflective responses while achieving a good content preservation-reflection style trade-off. | [
"Min, Do June",
"Perez-Rosas, Veronica",
"Resnicow, Ken",
"Mihalcea, Rada"
] | VERVE: Template-based ReflectiVE Rewriting for MotiVational IntErviewing | findings-emnlp.690 | 2311.08299 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.691.bib | https://aclanthology.org/2023.findings-emnlp.691/ | @inproceedings{wang-etal-2023-self-knowledge,
title = "Self-Knowledge Guided Retrieval Augmentation for Large Language Models",
author = "Wang, Yile and
Li, Peng and
Sun, Maosong and
Liu, Yang",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.691",
doi = "10.18653/v1/2023.findings-emnlp.691",
pages = "10303--10315",
abstract = "Large language models (LLMs) have shown superior performance without task-specific fine-tuning. Despite the success, the knowledge stored in the parameters of LLMs could still be incomplete and difficult to update due to the computational costs. As complementary, retrieval-based methods can offer non-parametric world knowledge and improve the performance on tasks such as question answering. However, we find that the retrieved knowledge does not always help and even has a negative impact on original responses occasionally. To better make use of both internal knowledge and external world knowledge, we investigate eliciting the model{'}s ability to recognize what they know and do not know (which is also called {``}self-knowledge{''}) and propose Self-Knowledge guided Retrieval augmentation (SKR), a simple yet effective method which can let LLMs refer to the questions they have previously encountered and adaptively call for external resources when dealing with new questions. We evaluate SKR on multiple datasets and demonstrate that it outperforms chain-of-thought based and fully retrieval-based methods by using either InstructGPT or ChatGPT.",
}
| Large language models (LLMs) have shown superior performance without task-specific fine-tuning. Despite the success, the knowledge stored in the parameters of LLMs could still be incomplete and difficult to update due to the computational costs. As complementary, retrieval-based methods can offer non-parametric world knowledge and improve the performance on tasks such as question answering. However, we find that the retrieved knowledge does not always help and even has a negative impact on original responses occasionally. To better make use of both internal knowledge and external world knowledge, we investigate eliciting the model{'}s ability to recognize what they know and do not know (which is also called {``}self-knowledge{''}) and propose Self-Knowledge guided Retrieval augmentation (SKR), a simple yet effective method which can let LLMs refer to the questions they have previously encountered and adaptively call for external resources when dealing with new questions. We evaluate SKR on multiple datasets and demonstrate that it outperforms chain-of-thought based and fully retrieval-based methods by using either InstructGPT or ChatGPT. | [
"Wang, Yile",
"Li, Peng",
"Sun, Maosong",
"Liu, Yang"
] | Self-Knowledge Guided Retrieval Augmentation for Large Language Models | findings-emnlp.691 | 2310.05002 | [
""
] | https://huggingface.co/papers/2310.05002 | 1 | 0 | 0 | 4 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.692.bib | https://aclanthology.org/2023.findings-emnlp.692/ | @inproceedings{zou-etal-2023-pretraining,
title = "Pretraining Language Models with Text-Attributed Heterogeneous Graphs",
author = "Zou, Tao and
Yu, Le and
Huang, Yifei and
Sun, Leilei and
Du, Bowen",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.692",
doi = "10.18653/v1/2023.findings-emnlp.692",
pages = "10316--10333",
abstract = "In many real-world scenarios (e.g., academic networks, social platforms), different types of entities are not only associated with texts but also connected by various relationships, which can be abstracted as Text-Attributed Heterogeneous Graphs (TAHGs). Current pretraining tasks for Language Models (LMs) primarily focus on separately learning the textual information of each entity and overlook the crucial aspect of capturing topological connections among entities in TAHGs. In this paper, we present a new pretraining framework for LMs that explicitly considers the topological and heterogeneous information in TAHGs. Firstly, we define a context graph as neighborhoods of a target node within specific orders and propose a topology-aware pretraining task to predict nodes involved in the context graph by jointly optimizing an LM and an auxiliary heterogeneous graph neural network. Secondly, based on the observation that some nodes are text-rich while others have little text, we devise a text augmentation strategy to enrich textless nodes with their neighbors{'} texts for handling the imbalance issue. We conduct link prediction and node classification tasks on three datasets from various domains. Experimental results demonstrate the superiority of our approach over existing methods and the rationality of each design. Our code is available at https://github.com/Hope-Rita/THLM.",
}
| In many real-world scenarios (e.g., academic networks, social platforms), different types of entities are not only associated with texts but also connected by various relationships, which can be abstracted as Text-Attributed Heterogeneous Graphs (TAHGs). Current pretraining tasks for Language Models (LMs) primarily focus on separately learning the textual information of each entity and overlook the crucial aspect of capturing topological connections among entities in TAHGs. In this paper, we present a new pretraining framework for LMs that explicitly considers the topological and heterogeneous information in TAHGs. Firstly, we define a context graph as neighborhoods of a target node within specific orders and propose a topology-aware pretraining task to predict nodes involved in the context graph by jointly optimizing an LM and an auxiliary heterogeneous graph neural network. Secondly, based on the observation that some nodes are text-rich while others have little text, we devise a text augmentation strategy to enrich textless nodes with their neighbors{'} texts for handling the imbalance issue. We conduct link prediction and node classification tasks on three datasets from various domains. Experimental results demonstrate the superiority of our approach over existing methods and the rationality of each design. Our code is available at https://github.com/Hope-Rita/THLM. | [
"Zou, Tao",
"Yu, Le",
"Huang, Yifei",
"Sun, Leilei",
"Du, Bowen"
] | Pretraining Language Models with Text-Attributed Heterogeneous Graphs | findings-emnlp.692 | 2310.12580 | [
"https://github.com/hope-rita/thlm"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.693.bib | https://aclanthology.org/2023.findings-emnlp.693/ | @inproceedings{chun-etal-2023-cretihc,
title = "{CR}e{TIHC}: Designing Causal Reasoning Tasks about Temporal Interventions and Hallucinated Confoundings",
author = "Chun, Changwoo and
Lee, SongEun and
Seo, Jaehyung and
Lim, Heuiseok",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.693",
doi = "10.18653/v1/2023.findings-emnlp.693",
pages = "10334--10343",
abstract = "Large language models (LLMs) have demonstrated impressive capabilities in natural language processing. However, their ability to establish causal relationships, particularly in the context of temporal interventions and language hallucinations, remains challenging. This paper presents \textbf{CReTIHC}, a novel dataset designed to test and enhance the causal reasoning abilities of LLMs. The dataset is constructed using a unique approach that incorporates elements of verbal hallucinations and temporal interventions through the reengineering of existing causal inference datasets. This transformation creates complex scenarios that push LLMs to critically evaluate the information presented and identify cause-and-effect relationships. The CReTIHC dataset serves as a pioneering tool for improving LLM{'}s causal inference capabilities, paving the way for a more nuanced understanding of causal relationships in natural language processing (NLP) tasks. The whole dataset is publicly accessible at: (https://github.com/ChangwooChun/CReTIHC)",
}
| Large language models (LLMs) have demonstrated impressive capabilities in natural language processing. However, their ability to establish causal relationships, particularly in the context of temporal interventions and language hallucinations, remains challenging. This paper presents \textbf{CReTIHC}, a novel dataset designed to test and enhance the causal reasoning abilities of LLMs. The dataset is constructed using a unique approach that incorporates elements of verbal hallucinations and temporal interventions through the reengineering of existing causal inference datasets. This transformation creates complex scenarios that push LLMs to critically evaluate the information presented and identify cause-and-effect relationships. The CReTIHC dataset serves as a pioneering tool for improving LLM{'}s causal inference capabilities, paving the way for a more nuanced understanding of causal relationships in natural language processing (NLP) tasks. The whole dataset is publicly accessible at: (https://github.com/ChangwooChun/CReTIHC) | [
"Chun, Changwoo",
"Lee, SongEun",
"Seo, Jaehyung",
"Lim, Heuiseok"
] | CReTIHC: Designing Causal Reasoning Tasks about Temporal Interventions and Hallucinated Confoundings | findings-emnlp.693 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.694.bib | https://aclanthology.org/2023.findings-emnlp.694/ | @inproceedings{wang-etal-2023-dimensionality,
title = "On the Dimensionality of Sentence Embeddings",
author = "Wang, Hongwei and
Zhang, Hongming and
Yu, Dong",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.694",
doi = "10.18653/v1/2023.findings-emnlp.694",
pages = "10344--10354",
abstract = "Learning sentence embeddings is a fundamental problem in natural language processing. While existing research primarily focuses on enhancing the quality of sentence embeddings, the exploration of sentence embedding dimensions is limited. Here we present a comprehensive and empirical analysis of the dimensionality of sentence embeddings. First, we demonstrate that the optimal dimension of sentence embeddings is usually smaller than the default value. Subsequently, to compress the dimension of sentence embeddings with minimum performance degradation, we identify two components contributing to the overall performance loss: the encoder{'}s performance loss and the pooler{'}s performance loss. Therefore, we propose a two-step training method for sentence representation learning models, wherein the encoder and the pooler are optimized separately to mitigate the overall performance loss in low-dimension scenarios. Experimental results on seven STS tasks and seven sentence classification tasks demonstrate that our method significantly improves the performance of low-dimensional sentence embeddings.",
}
| Learning sentence embeddings is a fundamental problem in natural language processing. While existing research primarily focuses on enhancing the quality of sentence embeddings, the exploration of sentence embedding dimensions is limited. Here we present a comprehensive and empirical analysis of the dimensionality of sentence embeddings. First, we demonstrate that the optimal dimension of sentence embeddings is usually smaller than the default value. Subsequently, to compress the dimension of sentence embeddings with minimum performance degradation, we identify two components contributing to the overall performance loss: the encoder{'}s performance loss and the pooler{'}s performance loss. Therefore, we propose a two-step training method for sentence representation learning models, wherein the encoder and the pooler are optimized separately to mitigate the overall performance loss in low-dimension scenarios. Experimental results on seven STS tasks and seven sentence classification tasks demonstrate that our method significantly improves the performance of low-dimensional sentence embeddings. | [
"Wang, Hongwei",
"Zhang, Hongming",
"Yu, Dong"
] | On the Dimensionality of Sentence Embeddings | findings-emnlp.694 | 2310.15285 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.695.bib | https://aclanthology.org/2023.findings-emnlp.695/ | @inproceedings{xue-aletras-2023-pit,
title = "Pit One Against Many: Leveraging Attention-head Embeddings for Parameter-efficient Multi-head Attention",
author = "Xue, Huiyin and
Aletras, Nikolaos",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.695",
doi = "10.18653/v1/2023.findings-emnlp.695",
pages = "10355--10373",
abstract = "Scaling pre-trained language models has resulted in large performance gains in various natural language processing tasks but comes with a large cost in memory requirements. Inspired by the position embeddings in transformers, we aim to simplify and reduce the memory footprint of the multi-head attention (MHA) mechanism. We propose an alternative module that uses only a single shared projection matrix and multiple head embeddings (MHE), i.e. one per head. We empirically demonstrate that our MHE attention is substantially more memory efficient compared to alternative attention mechanisms while achieving high predictive performance retention ratio to vanilla MHA on several downstream tasks. MHE attention only requires a negligible fraction of additional parameters ($3nd$, where $n$ is the number of attention heads and $d$ the size of the head embeddings) compared to a single-head attention, while MHA requires $(3n^2-3n)d^2-3nd$ additional parameters.",
}
| Scaling pre-trained language models has resulted in large performance gains in various natural language processing tasks but comes with a large cost in memory requirements. Inspired by the position embeddings in transformers, we aim to simplify and reduce the memory footprint of the multi-head attention (MHA) mechanism. We propose an alternative module that uses only a single shared projection matrix and multiple head embeddings (MHE), i.e. one per head. We empirically demonstrate that our MHE attention is substantially more memory efficient compared to alternative attention mechanisms while achieving high predictive performance retention ratio to vanilla MHA on several downstream tasks. MHE attention only requires a negligible fraction of additional parameters ($3nd$, where $n$ is the number of attention heads and $d$ the size of the head embeddings) compared to a single-head attention, while MHA requires $(3n^2-3n)d^2-3nd$ additional parameters. | [
"Xue, Huiyin",
"Aletras, Nikolaos"
] | Pit One Against Many: Leveraging Attention-head Embeddings for Parameter-efficient Multi-head Attention | findings-emnlp.695 | 2310.07911 | [
""
] | https://huggingface.co/papers/2310.07911 | 0 | 1 | 0 | 2 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.696.bib | https://aclanthology.org/2023.findings-emnlp.696/ | @inproceedings{zhou-tan-2023-entity,
title = "Entity-Based Evaluation of Political Bias in Automatic Summarization",
author = "Zhou, Karen and
Tan, Chenhao",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.696",
doi = "10.18653/v1/2023.findings-emnlp.696",
pages = "10374--10386",
abstract = "Growing literature has shown that NLP systems may encode social biases; however, the *political* bias of summarization models remains relatively unknown. In this work, we use an entity replacement method to investigate the portrayal of politicians in automatically generated summaries of news articles. We develop an entity-based computational framework to assess the sensitivities of several extractive and abstractive summarizers to the politicians Donald Trump and Joe Biden. We find consistent differences in these summaries upon entity replacement, such as reduced emphasis of Trump{'}s presence in the context of the same article and a more individualistic representation of Trump with respect to the collective US government (i.e., administration). These summary dissimilarities are most prominent when the entity is heavily featured in the source article. Our characterization provides a foundation for future studies of bias in summarization and for normative discussions on the ideal qualities of automatic summaries.",
}
| Growing literature has shown that NLP systems may encode social biases; however, the *political* bias of summarization models remains relatively unknown. In this work, we use an entity replacement method to investigate the portrayal of politicians in automatically generated summaries of news articles. We develop an entity-based computational framework to assess the sensitivities of several extractive and abstractive summarizers to the politicians Donald Trump and Joe Biden. We find consistent differences in these summaries upon entity replacement, such as reduced emphasis of Trump{'}s presence in the context of the same article and a more individualistic representation of Trump with respect to the collective US government (i.e., administration). These summary dissimilarities are most prominent when the entity is heavily featured in the source article. Our characterization provides a foundation for future studies of bias in summarization and for normative discussions on the ideal qualities of automatic summaries. | [
"Zhou, Karen",
"Tan, Chenhao"
] | Entity-Based Evaluation of Political Bias in Automatic Summarization | findings-emnlp.696 | 2305.02321 | [
"https://github.com/chicagohai/entity-based-political-bias"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.697.bib | https://aclanthology.org/2023.findings-emnlp.697/ | @inproceedings{wang-etal-2023-stylebart,
title = "{S}tyle{BART}: Decorate Pretrained Model with Style Adapters for Unsupervised Stylistic Headline Generation",
author = "Wang, Hanqing and
Luo, Yajing and
Xiong, Boya and
Chen, Guanhua and
Chen, Yun",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.697",
doi = "10.18653/v1/2023.findings-emnlp.697",
pages = "10387--10399",
abstract = "Stylistic headline generation is the task to generate a headline that not only summarizes the content of an article, but also reflects a desired style that attracts users. As style-specific article-headline pairs are scarce, previous researches focus on unsupervised approaches with a standard headline generation dataset and mono-style corpora. In this work, we follow this line and propose StyleBART, an unsupervised approach for stylistic headline generation. Our method decorates the pretrained BART model with adapters that are responsible for different styles and allows the generation of headlines with diverse styles by simply switching the adapters. Different from previous works, StyleBART separates the task of style learning and headline generation, making it possible to freely combine the base model and the style adapters during inference. We further propose an inverse paraphrasing task to enhance the style adapters. Extensive automatic and human evaluations show that StyleBART achieves new state-of-the-art performance in the unsupervised stylistic headline generation task, producing high-quality headlines with the desired style.",
}
| Stylistic headline generation is the task to generate a headline that not only summarizes the content of an article, but also reflects a desired style that attracts users. As style-specific article-headline pairs are scarce, previous researches focus on unsupervised approaches with a standard headline generation dataset and mono-style corpora. In this work, we follow this line and propose StyleBART, an unsupervised approach for stylistic headline generation. Our method decorates the pretrained BART model with adapters that are responsible for different styles and allows the generation of headlines with diverse styles by simply switching the adapters. Different from previous works, StyleBART separates the task of style learning and headline generation, making it possible to freely combine the base model and the style adapters during inference. We further propose an inverse paraphrasing task to enhance the style adapters. Extensive automatic and human evaluations show that StyleBART achieves new state-of-the-art performance in the unsupervised stylistic headline generation task, producing high-quality headlines with the desired style. | [
"Wang, Hanqing",
"Luo, Yajing",
"Xiong, Boya",
"Chen, Guanhua",
"Chen, Yun"
] | StyleBART: Decorate Pretrained Model with Style Adapters for Unsupervised Stylistic Headline Generation | findings-emnlp.697 | 2310.17743 | [
"https://github.com/sufenlp/stylebart"
] | https://huggingface.co/papers/2310.17743 | 0 | 1 | 0 | 5 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.698.bib | https://aclanthology.org/2023.findings-emnlp.698/ | @inproceedings{tang-etal-2023-rsvp,
title = "{RSVP}: Customer Intent Detection via Agent Response Contrastive and Generative Pre-Training",
author = "Tang, Yu-Chien and
Wang, Wei-Yao and
Yen, An-Zi and
Peng, Wen-Chih",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.698",
doi = "10.18653/v1/2023.findings-emnlp.698",
pages = "10400--10412",
abstract = "The dialogue systems in customer services have been developed with neural models to provide users with precise answers and round-the-clock support in task-oriented conversations by detecting customer intents based on their utterances. Existing intent detection approaches have highly relied on adaptively pre-training language models with large-scale datasets, yet the predominant cost of data collection may hinder their superiority. In addition, they neglect the information within the conversational responses of the agents, which have a lower collection cost, but are significant to customer intent as agents must tailor their replies based on the customers{'} intent. In this paper, we propose RSVP, a self-supervised framework dedicated to task-oriented dialogues, which utilizes agent responses for pre-training in a two-stage manner. Specifically, we introduce two pre-training tasks to incorporate the relations of utterance-response pairs: 1) Response Retrieval by selecting a correct response from a batch of candidates, and 2) Response Generation by mimicking agents to generate the response to a given utterance. Our benchmark results for two real-world customer service datasets show that RSVP significantly outperforms the state-of-the-art baselines by 4.95{\%} for accuracy, 3.4{\%} for MRR@3, and 2.75{\%} for MRR@5 on average. Extensive case studies are investigated to show the validity of incorporating agent responses into the pre-training stage.",
}
| The dialogue systems in customer services have been developed with neural models to provide users with precise answers and round-the-clock support in task-oriented conversations by detecting customer intents based on their utterances. Existing intent detection approaches have highly relied on adaptively pre-training language models with large-scale datasets, yet the predominant cost of data collection may hinder their superiority. In addition, they neglect the information within the conversational responses of the agents, which have a lower collection cost, but are significant to customer intent as agents must tailor their replies based on the customers{'} intent. In this paper, we propose RSVP, a self-supervised framework dedicated to task-oriented dialogues, which utilizes agent responses for pre-training in a two-stage manner. Specifically, we introduce two pre-training tasks to incorporate the relations of utterance-response pairs: 1) Response Retrieval by selecting a correct response from a batch of candidates, and 2) Response Generation by mimicking agents to generate the response to a given utterance. Our benchmark results for two real-world customer service datasets show that RSVP significantly outperforms the state-of-the-art baselines by 4.95{\%} for accuracy, 3.4{\%} for MRR@3, and 2.75{\%} for MRR@5 on average. Extensive case studies are investigated to show the validity of incorporating agent responses into the pre-training stage. | [
"Tang, Yu-Chien",
"Wang, Wei-Yao",
"Yen, An-Zi",
"Peng, Wen-Chih"
] | RSVP: Customer Intent Detection via Agent Response Contrastive and Generative Pre-Training | findings-emnlp.698 | 2310.09773 | [
"https://github.com/tommytyc/rsvp"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.699.bib | https://aclanthology.org/2023.findings-emnlp.699/ | @inproceedings{chen-etal-2023-improving-low,
title = "Improving Low-resource Question Answering by Augmenting Question Information",
author = "Chen, Andong and
Sun, Yuan and
Zhao, Xiaobing and
Galindo Esparza, Rosella and
Chen, Kehai and
Xiang, Yang and
Zhao, Tiejun and
Zhang, Min",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.699",
doi = "10.18653/v1/2023.findings-emnlp.699",
pages = "10413--10420",
abstract = "In the era of large models, low-resource question-answering tasks lag, emphasizing the importance of data augmentation - a key research avenue in natural language processing. The main challenges include leveraging the large model{'}s internal knowledge for data augmentation, determining which QA data component - the question, passage, or answer - benefits most from augmentation, and retaining consistency in the augmented content without inducing excessive noise. To tackle these, we introduce PQQ, an innovative approach for question data augmentation consisting of Prompt Answer, Question Generation, and Question Filter. Our experiments reveal that ChatGPT underperforms on the experimental data, yet our PQQ method excels beyond existing augmentation strategies. Further, its universal applicability is validated through successful tests on high-resource QA tasks like SQUAD1.1 and TriviaQA.",
}
| In the era of large models, low-resource question-answering tasks lag, emphasizing the importance of data augmentation - a key research avenue in natural language processing. The main challenges include leveraging the large model{'}s internal knowledge for data augmentation, determining which QA data component - the question, passage, or answer - benefits most from augmentation, and retaining consistency in the augmented content without inducing excessive noise. To tackle these, we introduce PQQ, an innovative approach for question data augmentation consisting of Prompt Answer, Question Generation, and Question Filter. Our experiments reveal that ChatGPT underperforms on the experimental data, yet our PQQ method excels beyond existing augmentation strategies. Further, its universal applicability is validated through successful tests on high-resource QA tasks like SQUAD1.1 and TriviaQA. | [
"Chen, Andong",
"Sun, Yuan",
"Zhao, Xiaobing",
"Galindo Esparza, Rosella",
"Chen, Kehai",
"Xiang, Yang",
"Zhao, Tiejun",
"Zhang, Min"
] | Improving Low-resource Question Answering by Augmenting Question Information | findings-emnlp.699 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.700.bib | https://aclanthology.org/2023.findings-emnlp.700/ | @inproceedings{zhang-etal-2023-instructsafety,
title = "{I}nstruct{S}afety: A Unified Framework for Building Multidimensional and Explainable Safety Detector through Instruction Tuning",
author = "Zhang, Zhexin and
Cheng, Jiale and
Sun, Hao and
Deng, Jiawen and
Huang, Minlie",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.700",
doi = "10.18653/v1/2023.findings-emnlp.700",
pages = "10421--10436",
abstract = "Safety detection has been an increasingly important topic in recent years and it has become even more necessary to develop reliable safety detection systems with the rapid development of large language models. However, currently available safety detection systems have limitations in terms of their versatility and interpretability. In this paper, we first introduce InstructSafety, a safety detection framework that unifies 7 common sub-tasks for safety detection. These tasks are unified into a similar form through different instructions. We then conduct a comprehensive survey of existing safety detection datasets and process 39 human-annotated datasets for instruction tuning. We also construct adversarial samples to enhance the model{'}s robustness. After fine-tuning Flan-T5 on the collected data, we have developed Safety-Flan-T5, a multidimensional and explainable safety detector. We conduct comprehensive experiments on a variety of datasets and tasks, and demonstrate the strong performance of Safety-Flan-T5 in comparison to supervised baselines and served APIs (Perspective API, ChatGPT and InstructGPT). We will release the processed data, fine-tuned Safety-Flan-T5 and related code for public use.",
}
| Safety detection has been an increasingly important topic in recent years and it has become even more necessary to develop reliable safety detection systems with the rapid development of large language models. However, currently available safety detection systems have limitations in terms of their versatility and interpretability. In this paper, we first introduce InstructSafety, a safety detection framework that unifies 7 common sub-tasks for safety detection. These tasks are unified into a similar form through different instructions. We then conduct a comprehensive survey of existing safety detection datasets and process 39 human-annotated datasets for instruction tuning. We also construct adversarial samples to enhance the model{'}s robustness. After fine-tuning Flan-T5 on the collected data, we have developed Safety-Flan-T5, a multidimensional and explainable safety detector. We conduct comprehensive experiments on a variety of datasets and tasks, and demonstrate the strong performance of Safety-Flan-T5 in comparison to supervised baselines and served APIs (Perspective API, ChatGPT and InstructGPT). We will release the processed data, fine-tuned Safety-Flan-T5 and related code for public use. | [
"Zhang, Zhexin",
"Cheng, Jiale",
"Sun, Hao",
"Deng, Jiawen",
"Huang, Minlie"
] | InstructSafety: A Unified Framework for Building Multidimensional and Explainable Safety Detector through Instruction Tuning | findings-emnlp.700 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.701.bib | https://aclanthology.org/2023.findings-emnlp.701/ | @inproceedings{roy-goldwasser-2023-tale,
title = "{``}A Tale of Two Movements{'}: Identifying and Comparing Perspectives in {\#}{B}lack{L}ives{M}atter and {\#}{B}lue{L}ives{M}atter Movements-related Tweets using Weakly Supervised Graph-based Structured Prediction",
author = "Roy, Shamik and
Goldwasser, Dan",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.701",
doi = "10.18653/v1/2023.findings-emnlp.701",
pages = "10437--10467",
abstract = "Social media has become a major driver of social change, by facilitating the formation of online social movements. Automatically understanding the perspectives driving the movement and the voices opposing it, is a challenging task as annotated data is difficult to obtain. We propose a weakly supervised graph-based approach that explicitly models perspectives in {\#}BackLivesMatter-related tweets. Our proposed approach utilizes a social-linguistic representation of the data. We convert the text to a graph by breaking it into structured elements and connect it with the social network of authors, then structured prediction is done over the elements for identifying perspectives. Our approach uses a small seed set of labeled examples. We experiment with large language models for generating artificial training examples, compare them to manual annotation, and find that it achieves comparable performance. We perform quantitative and qualitative analyses using a human-annotated test set. Our model outperforms multitask baselines by a large margin, successfully characterizing the perspectives supporting and opposing {\#}BLM.",
}
| Social media has become a major driver of social change, by facilitating the formation of online social movements. Automatically understanding the perspectives driving the movement and the voices opposing it, is a challenging task as annotated data is difficult to obtain. We propose a weakly supervised graph-based approach that explicitly models perspectives in {\#}BackLivesMatter-related tweets. Our proposed approach utilizes a social-linguistic representation of the data. We convert the text to a graph by breaking it into structured elements and connect it with the social network of authors, then structured prediction is done over the elements for identifying perspectives. Our approach uses a small seed set of labeled examples. We experiment with large language models for generating artificial training examples, compare them to manual annotation, and find that it achieves comparable performance. We perform quantitative and qualitative analyses using a human-annotated test set. Our model outperforms multitask baselines by a large margin, successfully characterizing the perspectives supporting and opposing {\#}BLM. | [
"Roy, Shamik",
"Goldwasser, Dan"
] | “A Tale of Two Movements': Identifying and Comparing Perspectives in #BlackLivesMatter and #BlueLivesMatter Movements-related Tweets using Weakly Supervised Graph-based Structured Prediction | findings-emnlp.701 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.702.bib | https://aclanthology.org/2023.findings-emnlp.702/ | @inproceedings{liang-liao-2023-clusterprompt,
title = "{C}luster{P}rompt: Cluster Semantic Enhanced Prompt Learning for New Intent Discovery",
author = "Liang, Jinggui and
Liao, Lizi",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.702",
doi = "10.18653/v1/2023.findings-emnlp.702",
pages = "10468--10481",
abstract = "The discovery of new intent categories from user utterances is a crucial task in expanding agent skills. The key lies in how to efficiently solicit semantic evidence from utterances and properly transfer knowledge from existing intents to new intents. However, previous methods laid too much emphasis on relations among utterances or clusters for transfer learning, while paying less attention to the usage of semantics. As a result, these methods suffer from in-domain over-fitting and often generate meaningless new intent clusters due to data distortion. In this paper, we present a novel approach called Cluster Semantic Enhanced Prompt Learning (CsePL) for discovering new intents. Our method leverages two-level contrastive learning with label semantic alignment to learn meaningful representations of intent clusters. These learned intent representations are then utilized as soft prompt initializations for discriminating new intents, reducing the dominance of existing intents. Extensive experiments conducted on three public datasets demonstrate the superiority of our proposed method. It not only outperforms existing methods but also suggests meaningful intent labels and enables early detection of new intents.",
}
| The discovery of new intent categories from user utterances is a crucial task in expanding agent skills. The key lies in how to efficiently solicit semantic evidence from utterances and properly transfer knowledge from existing intents to new intents. However, previous methods laid too much emphasis on relations among utterances or clusters for transfer learning, while paying less attention to the usage of semantics. As a result, these methods suffer from in-domain over-fitting and often generate meaningless new intent clusters due to data distortion. In this paper, we present a novel approach called Cluster Semantic Enhanced Prompt Learning (CsePL) for discovering new intents. Our method leverages two-level contrastive learning with label semantic alignment to learn meaningful representations of intent clusters. These learned intent representations are then utilized as soft prompt initializations for discriminating new intents, reducing the dominance of existing intents. Extensive experiments conducted on three public datasets demonstrate the superiority of our proposed method. It not only outperforms existing methods but also suggests meaningful intent labels and enables early detection of new intents. | [
"Liang, Jinggui",
"Liao, Lizi"
] | ClusterPrompt: Cluster Semantic Enhanced Prompt Learning for New Intent Discovery | findings-emnlp.702 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.703.bib | https://aclanthology.org/2023.findings-emnlp.703/ | @inproceedings{kabbara-cheung-2023-investigating,
title = "Investigating the Effect of Pre-finetuning {BERT} Models on {NLI} Involving Presuppositions",
author = "Kabbara, Jad and
Cheung, Jackie",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.703",
doi = "10.18653/v1/2023.findings-emnlp.703",
pages = "10482--10494",
abstract = "We explore the connection between presupposition, discourse and sarcasm and propose to leverage that connection in a transfer learning scenario with the goal of improving the performance of NLI models on cases involving presupposition. We exploit advances in training transformer-based models that show that pre-finetuning{---}{--}i.e., finetuning the model on an additional task or dataset before the actual finetuning phase{---}{--}can help these models, in some cases, achieve a higher performance on a given downstream task. Building on those advances and that aforementioned connection, we propose pre-finetuning NLI models on carefully chosen tasks in an attempt to improve their performance on NLI cases involving presupposition. We notice that, indeed, pre-finetuning on those tasks leads to performance improvements. Furthermore, we run several diagnostic tests to understand whether these gains are merely a byproduct of additional training data. The results show that, while additional training data seems to be helping on its own in some cases, the choice of the tasks plays a role in the performance improvements.",
}
| We explore the connection between presupposition, discourse and sarcasm and propose to leverage that connection in a transfer learning scenario with the goal of improving the performance of NLI models on cases involving presupposition. We exploit advances in training transformer-based models that show that pre-finetuning{---}{--}i.e., finetuning the model on an additional task or dataset before the actual finetuning phase{---}{--}can help these models, in some cases, achieve a higher performance on a given downstream task. Building on those advances and that aforementioned connection, we propose pre-finetuning NLI models on carefully chosen tasks in an attempt to improve their performance on NLI cases involving presupposition. We notice that, indeed, pre-finetuning on those tasks leads to performance improvements. Furthermore, we run several diagnostic tests to understand whether these gains are merely a byproduct of additional training data. The results show that, while additional training data seems to be helping on its own in some cases, the choice of the tasks plays a role in the performance improvements. | [
"Kabbara, Jad",
"Cheung, Jackie"
] | Investigating the Effect of Pre-finetuning BERT Models on NLI Involving Presuppositions | findings-emnlp.703 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.704.bib | https://aclanthology.org/2023.findings-emnlp.704/ | @inproceedings{cheng-etal-2023-mrrl,
title = "{MRRL}: Modifying the Reference via Reinforcement Learning for Non-Autoregressive Joint Multiple Intent Detection and Slot Filling",
author = "Cheng, Xuxin and
Zhu, Zhihong and
Cao, Bowen and
Ye, Qichen and
Zou, Yuexian",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.704",
doi = "10.18653/v1/2023.findings-emnlp.704",
pages = "10495--10505",
abstract = "With the rise of non-autoregressive approach, some non-autoregressive models for joint multiple intent detection and slot filling have obtained the promising inference speed. However, most existing SLU models (1) suffer from the multi-modality problem that leads to reference intents and slots may not be suitable for training; (2) lack of alignment between the correct predictions of the two tasks, which extremely limits the overall accuracy. Therefore, in this paper, we propose $\textbf{M}$odifying the $\textbf{R}$eference via $\textbf{R}$einforcement $\textbf{L}$earning (MRRL), a novel method for multiple intent detection and slot filling, which introduces a modifier and employs reinforcement learning. Specifically, we try to provide the better training target for the non-autoregressive SLU model via modifying the reference based on the output of the non-autoregressive SLU model, and propose a suitability reward to ensure that the output of the modifier module could fit well with the output of the non-autoregressive SLU model and does not deviate too far from the reference. In addition, we also propose a compromise reward to realize a flexible trade-off between the two subtasks. Experiments on two multi-intent datasets and non-autoregressive baselines demonstrate that our MRRL could consistently improve the performance of baselines. More encouragingly, our best variant achieves new state-of-the-art results, outperforming the previous best approach by 3.6 overall accuracy on MixATIS dataset.",
}
| With the rise of non-autoregressive approach, some non-autoregressive models for joint multiple intent detection and slot filling have obtained the promising inference speed. However, most existing SLU models (1) suffer from the multi-modality problem that leads to reference intents and slots may not be suitable for training; (2) lack of alignment between the correct predictions of the two tasks, which extremely limits the overall accuracy. Therefore, in this paper, we propose $\textbf{M}$odifying the $\textbf{R}$eference via $\textbf{R}$einforcement $\textbf{L}$earning (MRRL), a novel method for multiple intent detection and slot filling, which introduces a modifier and employs reinforcement learning. Specifically, we try to provide the better training target for the non-autoregressive SLU model via modifying the reference based on the output of the non-autoregressive SLU model, and propose a suitability reward to ensure that the output of the modifier module could fit well with the output of the non-autoregressive SLU model and does not deviate too far from the reference. In addition, we also propose a compromise reward to realize a flexible trade-off between the two subtasks. Experiments on two multi-intent datasets and non-autoregressive baselines demonstrate that our MRRL could consistently improve the performance of baselines. More encouragingly, our best variant achieves new state-of-the-art results, outperforming the previous best approach by 3.6 overall accuracy on MixATIS dataset. | [
"Cheng, Xuxin",
"Zhu, Zhihong",
"Cao, Bowen",
"Ye, Qichen",
"Zou, Yuexian"
] | MRRL: Modifying the Reference via Reinforcement Learning for Non-Autoregressive Joint Multiple Intent Detection and Slot Filling | findings-emnlp.704 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.705.bib | https://aclanthology.org/2023.findings-emnlp.705/ | @inproceedings{dong-etal-2023-demonsf,
title = "{D}emo{NSF}: A Multi-task Demonstration-based Generative Framework for Noisy Slot Filling Task",
author = "Dong, Guanting and
Hui, Tingfeng and
GongQue, Zhuoma and
Zhao, Jinxu and
Guo, Daichi and
Zhao, Gang and
He, Keqing and
Xu, Weiran",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.705",
doi = "10.18653/v1/2023.findings-emnlp.705",
pages = "10506--10518",
abstract = "Recently, prompt-based generative frameworks have shown impressive capabilities in sequence labeling tasks. However, in practical dialogue scenarios, relying solely on simplistic templates and traditional corpora presents a challenge for these methods in generalizing to unknown input perturbations. To address this gap, we propose a multi-task demonstration-based generative framework for noisy slot filling, named DemoNSF. Specifically, we introduce three noisy auxiliary tasks, namely noisy recovery (NR), random mask (RM), and hybrid discrimination (HD), to implicitly capture semantic structural information of input perturbations at different granularities. In the downstream main task, we design a noisy demonstration construction strategy for the generative framework, which explicitly incorporates task-specific information and perturbed distribution during training and inference. Experiments on two benchmarks demonstrate that DemoNSF outperforms all baseline methods and achieves strong generalization. Further analysis provides empirical guidance for the practical application of generative frameworks. Our code is released at https://github.com/dongguanting/Demo-NSF.",
}
| Recently, prompt-based generative frameworks have shown impressive capabilities in sequence labeling tasks. However, in practical dialogue scenarios, relying solely on simplistic templates and traditional corpora presents a challenge for these methods in generalizing to unknown input perturbations. To address this gap, we propose a multi-task demonstration-based generative framework for noisy slot filling, named DemoNSF. Specifically, we introduce three noisy auxiliary tasks, namely noisy recovery (NR), random mask (RM), and hybrid discrimination (HD), to implicitly capture semantic structural information of input perturbations at different granularities. In the downstream main task, we design a noisy demonstration construction strategy for the generative framework, which explicitly incorporates task-specific information and perturbed distribution during training and inference. Experiments on two benchmarks demonstrate that DemoNSF outperforms all baseline methods and achieves strong generalization. Further analysis provides empirical guidance for the practical application of generative frameworks. Our code is released at https://github.com/dongguanting/Demo-NSF. | [
"Dong, Guanting",
"Hui, Tingfeng",
"GongQue, Zhuoma",
"Zhao, Jinxu",
"Guo, Daichi",
"Zhao, Gang",
"He, Keqing",
"Xu, Weiran"
] | DemoNSF: A Multi-task Demonstration-based Generative Framework for Noisy Slot Filling Task | findings-emnlp.705 | 2310.10169 | [
"https://github.com/dongguanting/Demo-NSF"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.706.bib | https://aclanthology.org/2023.findings-emnlp.706/ | @inproceedings{salehi-etal-2023-sharcs,
title = "{SHARCS}: Efficient Transformers Through Routing with Dynamic Width Sub-networks",
author = "Salehi, Mohammadreza and
Mehta, Sachin and
Kusupati, Aditya and
Farhadi, Ali and
Hajishirzi, Hannaneh",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.706",
doi = "10.18653/v1/2023.findings-emnlp.706",
pages = "10519--10532",
abstract = "We introduce SHARCS for adaptive inference that takes into account the hardness of input samples. SHARCS can train a router on any transformer network, enabling the model to direct different samples to sub-networks with varying widths. Our experiments demonstrate that: (1) SHARCS outperforms or complements existing per-sample adaptive inference methods across various classification tasks in terms of accuracy vs. FLOPs; (2) SHARCS generalizes across different architectures and can be even applied to compressed and efficient transformer encoders to further improve their efficiency; (3) SHARCS can provide a 2 times inference speed up at an insignificant drop in accuracy.",
}
| We introduce SHARCS for adaptive inference that takes into account the hardness of input samples. SHARCS can train a router on any transformer network, enabling the model to direct different samples to sub-networks with varying widths. Our experiments demonstrate that: (1) SHARCS outperforms or complements existing per-sample adaptive inference methods across various classification tasks in terms of accuracy vs. FLOPs; (2) SHARCS generalizes across different architectures and can be even applied to compressed and efficient transformer encoders to further improve their efficiency; (3) SHARCS can provide a 2 times inference speed up at an insignificant drop in accuracy. | [
"Salehi, Mohammadreza",
"Mehta, Sachin",
"Kusupati, Aditya",
"Farhadi, Ali",
"Hajishirzi, Hannaneh"
] | SHARCS: Efficient Transformers Through Routing with Dynamic Width Sub-networks | findings-emnlp.706 | 2310.12126 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.707.bib | https://aclanthology.org/2023.findings-emnlp.707/ | @inproceedings{bai-etal-2023-always,
title = "Always the Best Fit: Adaptive Domain Gap Filling from Causal Perspective for Few-Shot Relation Extraction",
author = "Bai, Ge and
Lu, Chenji and
Geng, Jiaxiang and
Li, Shilong and
Shi, Yidong and
Liu, Xiyan and
Liu, Ying and
Zhang, Zhang and
Liu, Ruifang",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.707",
doi = "10.18653/v1/2023.findings-emnlp.707",
pages = "10533--10542",
abstract = "Cross-domain Relation Extraction aims to transfer knowledge from a source domain to a different target domain to address low-resource challenges. However, the semantic gap caused by data bias between domains is a major challenge, especially in few-shot scenarios. Previous work has mainly focused on transferring knowledge between domains through shared feature representations without analyzing the impact of each factor that may produce data bias based on the characteristics of each domain. This work takes a causal perspective and proposes a new framework CausalGF. By constructing a unified structural causal model, we estimating the causal effects of factors such as syntactic structure, label distribution,and entities on the outcome. CausalGF calculates the causal effects among the factors and adjusts them dynamically based on domain characteristics, enabling adaptive gap filling. Our experiments show that our approach better fills the domain gap, yielding significantly better results on the cross-domain few-shot relation extraction task.",
}
| Cross-domain Relation Extraction aims to transfer knowledge from a source domain to a different target domain to address low-resource challenges. However, the semantic gap caused by data bias between domains is a major challenge, especially in few-shot scenarios. Previous work has mainly focused on transferring knowledge between domains through shared feature representations without analyzing the impact of each factor that may produce data bias based on the characteristics of each domain. This work takes a causal perspective and proposes a new framework CausalGF. By constructing a unified structural causal model, we estimating the causal effects of factors such as syntactic structure, label distribution,and entities on the outcome. CausalGF calculates the causal effects among the factors and adjusts them dynamically based on domain characteristics, enabling adaptive gap filling. Our experiments show that our approach better fills the domain gap, yielding significantly better results on the cross-domain few-shot relation extraction task. | [
"Bai, Ge",
"Lu, Chenji",
"Geng, Jiaxiang",
"Li, Shilong",
"Shi, Yidong",
"Liu, Xiyan",
"Liu, Ying",
"Zhang, Zhang",
"Liu, Ruifang"
] | Always the Best Fit: Adaptive Domain Gap Filling from Causal Perspective for Few-Shot Relation Extraction | findings-emnlp.707 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.708.bib | https://aclanthology.org/2023.findings-emnlp.708/ | @inproceedings{kargupta-etal-2023-megclass,
title = "{MEGC}lass: Extremely Weakly Supervised Text Classification via Mutually-Enhancing Text Granularities",
author = "Kargupta, Priyanka and
Komarlu, Tanay and
Yoon, Susik and
Wang, Xuan and
Han, Jiawei",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.708",
doi = "10.18653/v1/2023.findings-emnlp.708",
pages = "10543--10558",
abstract = "Text classification is essential for organizing unstructured text. Traditional methods rely on human annotations or, more recently, a set of class seed words for supervision, which can be costly, particularly for specialized or emerging domains. To address this, using class surface names alone as extremely weak supervision has been proposed. However, existing approaches treat different levels of text granularity (documents, sentences, or words) independently, disregarding inter-granularity class disagreements and the context identifiable exclusively through joint extraction. In order to tackle these issues, we introduce MEGClass, an extremely weakly-supervised text classification method that leverages Mutually-Enhancing Text Granularities. MEGClass utilizes coarse- and fine-grained context signals obtained by jointly considering a document{'}s most class-indicative words and sentences. This approach enables the learning of a contextualized document representation that captures the most discriminative class indicators. By preserving the heterogeneity of potential classes, MEGClass can select the most informative class-indicative documents as iterative feedback to enhance the initial word-based class representations and ultimately fine-tune a pre-trained text classifier. Extensive experiments on seven benchmark datasets demonstrate that MEGClass outperforms other weakly and extremely weakly supervised methods.",
}
| Text classification is essential for organizing unstructured text. Traditional methods rely on human annotations or, more recently, a set of class seed words for supervision, which can be costly, particularly for specialized or emerging domains. To address this, using class surface names alone as extremely weak supervision has been proposed. However, existing approaches treat different levels of text granularity (documents, sentences, or words) independently, disregarding inter-granularity class disagreements and the context identifiable exclusively through joint extraction. In order to tackle these issues, we introduce MEGClass, an extremely weakly-supervised text classification method that leverages Mutually-Enhancing Text Granularities. MEGClass utilizes coarse- and fine-grained context signals obtained by jointly considering a document{'}s most class-indicative words and sentences. This approach enables the learning of a contextualized document representation that captures the most discriminative class indicators. By preserving the heterogeneity of potential classes, MEGClass can select the most informative class-indicative documents as iterative feedback to enhance the initial word-based class representations and ultimately fine-tune a pre-trained text classifier. Extensive experiments on seven benchmark datasets demonstrate that MEGClass outperforms other weakly and extremely weakly supervised methods. | [
"Kargupta, Priyanka",
"Komarlu, Tanay",
"Yoon, Susik",
"Wang, Xuan",
"Han, Jiawei"
] | MEGClass: Extremely Weakly Supervised Text Classification via Mutually-Enhancing Text Granularities | findings-emnlp.708 | 2304.01969 | [
"https://github.com/pkargupta/MEGClass"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.709.bib | https://aclanthology.org/2023.findings-emnlp.709/ | @inproceedings{zhou-he-2023-causal,
title = "Causal Inference from Text: Unveiling Interactions between Variables",
author = "Zhou, Yuxiang and
He, Yulan",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.709",
doi = "10.18653/v1/2023.findings-emnlp.709",
pages = "10559--10571",
abstract = "Adjusting for latent covariates is crucial for estimating causal effects from observational textual data. Most existing methods only account for confounding covariates that affect both treatment and outcome, potentially leading to biased causal effects. This bias arises from insufficient consideration of non-confounding covariates, which are relevant only to either the treatment or the outcome. In this work, we aim to mitigate the bias by unveiling interactions between different variables to disentangle the non-confounding covariates when estimating causal effects from text. The disentangling process ensures covariates only contribute to their respective objectives, enabling independence between variables. Additionally, we impose a constraint to balance representations from the treated group and control group to alleviate selection bias. We conduct experiments on two different treatment factors under various scenarios, and the proposed model significantly outperforms recent strong baselines. Furthermore, our thorough analysis on earnings call transcripts demonstrates that our model can effectively disentangle the variables, and further investigations into real-world scenarios provide guidance for investors to make informed decisions.",
}
| Adjusting for latent covariates is crucial for estimating causal effects from observational textual data. Most existing methods only account for confounding covariates that affect both treatment and outcome, potentially leading to biased causal effects. This bias arises from insufficient consideration of non-confounding covariates, which are relevant only to either the treatment or the outcome. In this work, we aim to mitigate the bias by unveiling interactions between different variables to disentangle the non-confounding covariates when estimating causal effects from text. The disentangling process ensures covariates only contribute to their respective objectives, enabling independence between variables. Additionally, we impose a constraint to balance representations from the treated group and control group to alleviate selection bias. We conduct experiments on two different treatment factors under various scenarios, and the proposed model significantly outperforms recent strong baselines. Furthermore, our thorough analysis on earnings call transcripts demonstrates that our model can effectively disentangle the variables, and further investigations into real-world scenarios provide guidance for investors to make informed decisions. | [
"Zhou, Yuxiang",
"He, Yulan"
] | Causal Inference from Text: Unveiling Interactions between Variables | findings-emnlp.709 | 2311.05286 | [
"https://github.com/zyxnlp/diva"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.710.bib | https://aclanthology.org/2023.findings-emnlp.710/ | @inproceedings{ma-etal-2023-large,
title = "Large Language Model Is Not a Good Few-shot Information Extractor, but a Good Reranker for Hard Samples!",
author = "Ma, Yubo and
Cao, Yixin and
Hong, Yong and
Sun, Aixin",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.710",
doi = "10.18653/v1/2023.findings-emnlp.710",
pages = "10572--10601",
abstract = "Large Language Models (LLMs) have made remarkable strides in various tasks. Whether LLMs are competitive few-shot solvers for information extraction (IE) tasks, however, remains an open problem. In this work, we aim to provide a thorough answer to this question. Through extensive experiments on nine datasets across four IE tasks, we demonstrate that current advanced LLMs consistently exhibit inferior performance, higher latency, and increased budget requirements compared to fine-tuned SLMs under most settings. Therefore, we conclude that LLMs are not effective few-shot information extractors in general. Nonetheless, we illustrate that with appropriate prompting strategies, LLMs can effectively complement SLMs and tackle challenging samples that SLMs struggle with. And moreover, we propose an adaptive filter-then-rerank paradigm to combine the strengths of LLMs and SLMs. In this paradigm, SLMs serve as filters and LLMs serve as rerankers. By prompting LLMs to rerank a small portion of difficult samples identified by SLMs, our preliminary system consistently achieves promising improvements (2.4{\%} F1-gain on average) on various IE tasks, with an acceptable time and cost investment.",
}
| Large Language Models (LLMs) have made remarkable strides in various tasks. Whether LLMs are competitive few-shot solvers for information extraction (IE) tasks, however, remains an open problem. In this work, we aim to provide a thorough answer to this question. Through extensive experiments on nine datasets across four IE tasks, we demonstrate that current advanced LLMs consistently exhibit inferior performance, higher latency, and increased budget requirements compared to fine-tuned SLMs under most settings. Therefore, we conclude that LLMs are not effective few-shot information extractors in general. Nonetheless, we illustrate that with appropriate prompting strategies, LLMs can effectively complement SLMs and tackle challenging samples that SLMs struggle with. And moreover, we propose an adaptive filter-then-rerank paradigm to combine the strengths of LLMs and SLMs. In this paradigm, SLMs serve as filters and LLMs serve as rerankers. By prompting LLMs to rerank a small portion of difficult samples identified by SLMs, our preliminary system consistently achieves promising improvements (2.4{\%} F1-gain on average) on various IE tasks, with an acceptable time and cost investment. | [
"Ma, Yubo",
"Cao, Yixin",
"Hong, Yong",
"Sun, Aixin"
] | Large Language Model Is Not a Good Few-shot Information Extractor, but a Good Reranker for Hard Samples! | findings-emnlp.710 | 2303.08559 | [
"https://github.com/mayubo2333/llm-ie"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.711.bib | https://aclanthology.org/2023.findings-emnlp.711/ | @inproceedings{deng-etal-2023-prompting,
title = "Prompting and Evaluating Large Language Models for Proactive Dialogues: Clarification, Target-guided, and Non-collaboration",
author = "Deng, Yang and
Liao, Lizi and
Chen, Liang and
Wang, Hongru and
Lei, Wenqiang and
Chua, Tat-Seng",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.711",
doi = "10.18653/v1/2023.findings-emnlp.711",
pages = "10602--10621",
abstract = "Conversational systems based on Large Language Models (LLMs), such as ChatGPT, show exceptional proficiency in context understanding and response generation. However, they still possess limitations, such as failing to ask clarifying questions to ambiguous queries or refuse users{'} unreasonable requests, both of which are considered as key aspects of a conversational agent{'}s proactivity. This raises the question of whether LLM-based conversational systems are equipped to handle proactive dialogue problems. In this work, we conduct a comprehensive analysis of LLM-based conversational systems, specifically focusing on three key aspects of proactive dialogues: clarification, target-guided, and non-collaborative dialogues. To trigger the proactivity of LLMs, we propose the Proactive Chain-of-Thought prompting scheme, which augments LLMs with the goal planning capability over descriptive reasoning chains. Empirical findings are discussed to promote future studies on LLM-based proactive dialogue systems.",
}
| Conversational systems based on Large Language Models (LLMs), such as ChatGPT, show exceptional proficiency in context understanding and response generation. However, they still possess limitations, such as failing to ask clarifying questions to ambiguous queries or refuse users{'} unreasonable requests, both of which are considered as key aspects of a conversational agent{'}s proactivity. This raises the question of whether LLM-based conversational systems are equipped to handle proactive dialogue problems. In this work, we conduct a comprehensive analysis of LLM-based conversational systems, specifically focusing on three key aspects of proactive dialogues: clarification, target-guided, and non-collaborative dialogues. To trigger the proactivity of LLMs, we propose the Proactive Chain-of-Thought prompting scheme, which augments LLMs with the goal planning capability over descriptive reasoning chains. Empirical findings are discussed to promote future studies on LLM-based proactive dialogue systems. | [
"Deng, Yang",
"Liao, Lizi",
"Chen, Liang",
"Wang, Hongru",
"Lei, Wenqiang",
"Chua, Tat-Seng"
] | Prompting and Evaluating Large Language Models for Proactive Dialogues: Clarification, Target-guided, and Non-collaboration | findings-emnlp.711 | 2305.13626 | [
"https://github.com/dengyang17/llm-proactive"
] | https://huggingface.co/papers/2305.13626 | 0 | 0 | 0 | 4 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.712.bib | https://aclanthology.org/2023.findings-emnlp.712/ | @inproceedings{jiang-etal-2023-ecologically,
title = "Ecologically Valid Explanations for Label Variation in {NLI}",
author = "Jiang, Nan-Jiang and
Tan, Chenhao and
de Marneffe, Marie-Catherine",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.712",
doi = "10.18653/v1/2023.findings-emnlp.712",
pages = "10622--10633",
abstract = "Human label variation, or annotation disagreement, exists in many natural language processing (NLP) tasks, including natural language inference (NLI). To gain direct evidence of how NLI label variation arises, we build LiveNLI, an English dataset of 1,415 ecologically valid explanations (annotators explain the NLI labels they chose) for 122 MNLI items (at least 10 explanations per item). The LiveNLI explanations confirm that people can systematically vary on their interpretation and highlight within-label variation: annotators sometimes choose the same label for different reasons. This suggests that explanations are crucial for navigating label interpretations in general. We few-shot prompt large language models to generate explanations but the results are inconsistent: they sometimes produces valid and informative explanations, but it also generates implausible ones that do not support the label, highlighting directions for improvement.",
}
| Human label variation, or annotation disagreement, exists in many natural language processing (NLP) tasks, including natural language inference (NLI). To gain direct evidence of how NLI label variation arises, we build LiveNLI, an English dataset of 1,415 ecologically valid explanations (annotators explain the NLI labels they chose) for 122 MNLI items (at least 10 explanations per item). The LiveNLI explanations confirm that people can systematically vary on their interpretation and highlight within-label variation: annotators sometimes choose the same label for different reasons. This suggests that explanations are crucial for navigating label interpretations in general. We few-shot prompt large language models to generate explanations but the results are inconsistent: they sometimes produces valid and informative explanations, but it also generates implausible ones that do not support the label, highlighting directions for improvement. | [
"Jiang, Nan-Jiang",
"Tan, Chenhao",
"de Marneffe, Marie-Catherine"
] | Ecologically Valid Explanations for Label Variation in NLI | findings-emnlp.712 | 2310.13850 | [
"https://github.com/njjiang/livenli"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.713.bib | https://aclanthology.org/2023.findings-emnlp.713/ | @inproceedings{kochsiek-gemulla-2023-benchmark,
title = "A Benchmark for Semi-Inductive Link Prediction in Knowledge Graphs",
author = "Kochsiek, Adrian and
Gemulla, Rainer",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.713",
doi = "10.18653/v1/2023.findings-emnlp.713",
pages = "10634--10643",
abstract = "Semi-inductive link prediction (LP) in knowledge graphs (KG) is the task of predicting facts for new, previously unseen entities based on context information. Although new entities can be integrated by retraining the model from scratch in principle, such an approach is infeasible for large-scale KGs, where retraining is expensive and new entities may arise frequently. In this paper, we propose and describe a large-scale benchmark to evaluate semi-inductive LP models. The benchmark is based on and extends Wikidata5M: It provides transductive, k-shot, and 0-shot LP tasks, each varying the available information from (i) only KG structure, to (ii) including textual mentions, and (iii) detailed descriptions of the entities. We report on a small study of recent approaches and found that semi-inductive LP performance is far from transductive performance on long-tail entities throughout all experiments. The benchmark provides a test bed for further research into integrating context and textual information in semi-inductive LP models.",
}
| Semi-inductive link prediction (LP) in knowledge graphs (KG) is the task of predicting facts for new, previously unseen entities based on context information. Although new entities can be integrated by retraining the model from scratch in principle, such an approach is infeasible for large-scale KGs, where retraining is expensive and new entities may arise frequently. In this paper, we propose and describe a large-scale benchmark to evaluate semi-inductive LP models. The benchmark is based on and extends Wikidata5M: It provides transductive, k-shot, and 0-shot LP tasks, each varying the available information from (i) only KG structure, to (ii) including textual mentions, and (iii) detailed descriptions of the entities. We report on a small study of recent approaches and found that semi-inductive LP performance is far from transductive performance on long-tail entities throughout all experiments. The benchmark provides a test bed for further research into integrating context and textual information in semi-inductive LP models. | [
"Kochsiek, Adrian",
"Gemulla, Rainer"
] | A Benchmark for Semi-Inductive Link Prediction in Knowledge Graphs | findings-emnlp.713 | 2310.11917 | [
"https://github.com/uma-pi1/wikidata5m-si"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.714.bib | https://aclanthology.org/2023.findings-emnlp.714/ | @inproceedings{zhang-etal-2023-summit,
title = "{S}umm{I}t: Iterative Text Summarization via {C}hat{GPT}",
author = "Zhang, Haopeng and
Liu, Xiao and
Zhang, Jiawei",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.714",
doi = "10.18653/v1/2023.findings-emnlp.714",
pages = "10644--10657",
abstract = "Existing text summarization systems have made significant progress in recent years, but typically generate summaries in a single step. The one-shot summarization setting is sometimes inadequate, however, as the generated summary may contain hallucinations or overlook important details related to the reader{'}s interests. In this paper, we address this limitation by proposing SummIt, an iterative text summarization framework based on large language models like ChatGPT. Our framework enables the model to refine the generated summary iteratively through self-evaluation and feedback, closely resembling the iterative process humans undertake when drafting and revising summaries. Furthermore, we explore the potential benefits of integrating knowledge and topic extractors into the framework to enhance summary faithfulness and controllability. We evaluate the performance of our framework on three benchmark summarization datasets through empirical and qualitative analyses. We also conduct a human evaluation to validate the effectiveness of the model{'}s refinements and find a potential issue of over-correction.",
}
| Existing text summarization systems have made significant progress in recent years, but typically generate summaries in a single step. The one-shot summarization setting is sometimes inadequate, however, as the generated summary may contain hallucinations or overlook important details related to the reader{'}s interests. In this paper, we address this limitation by proposing SummIt, an iterative text summarization framework based on large language models like ChatGPT. Our framework enables the model to refine the generated summary iteratively through self-evaluation and feedback, closely resembling the iterative process humans undertake when drafting and revising summaries. Furthermore, we explore the potential benefits of integrating knowledge and topic extractors into the framework to enhance summary faithfulness and controllability. We evaluate the performance of our framework on three benchmark summarization datasets through empirical and qualitative analyses. We also conduct a human evaluation to validate the effectiveness of the model{'}s refinements and find a potential issue of over-correction. | [
"Zhang, Haopeng",
"Liu, Xiao",
"Zhang, Jiawei"
] | SummIt: Iterative Text Summarization via ChatGPT | findings-emnlp.714 | 2305.14835 | [
"https://github.com/hpzhang94/summ_it"
] | https://huggingface.co/papers/2305.14835 | 1 | 1 | 0 | 3 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.715.bib | https://aclanthology.org/2023.findings-emnlp.715/ | @inproceedings{wang-etal-2023-orthogonal,
title = "Orthogonal Subspace Learning for Language Model Continual Learning",
author = "Wang, Xiao and
Chen, Tianze and
Ge, Qiming and
Xia, Han and
Bao, Rong and
Zheng, Rui and
Zhang, Qi and
Gui, Tao and
Huang, Xuanjing",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.715",
doi = "10.18653/v1/2023.findings-emnlp.715",
pages = "10658--10671",
abstract = "Benefiting from massive corpora and advanced hardware, large language models (LLMs) exhibit remarkable capabilities in language understanding and generation. However, their performance degrades in scenarios where multiple tasks are encountered sequentially, also known as catastrophic forgetting. In this paper, we propose orthogonal low-rank adaptation (O-LoRA), a simple and efficient approach for continual learning in language models, effectively mitigating catastrophic forgetting while learning new tasks. Specifically, O-LoRA learns tasks in different (low-rank) vector subspaces that are kept orthogonal to each other in order to minimize interference. Our method induces only marginal additional parameter costs and requires no user data storage for replay. Experimental results on continual learning benchmarks show that our method outperforms state-of-the-art methods. Furthermore, compared to previous approaches, our method excels in preserving the generalization ability of LLMs on unseen tasks.",
}
| Benefiting from massive corpora and advanced hardware, large language models (LLMs) exhibit remarkable capabilities in language understanding and generation. However, their performance degrades in scenarios where multiple tasks are encountered sequentially, also known as catastrophic forgetting. In this paper, we propose orthogonal low-rank adaptation (O-LoRA), a simple and efficient approach for continual learning in language models, effectively mitigating catastrophic forgetting while learning new tasks. Specifically, O-LoRA learns tasks in different (low-rank) vector subspaces that are kept orthogonal to each other in order to minimize interference. Our method induces only marginal additional parameter costs and requires no user data storage for replay. Experimental results on continual learning benchmarks show that our method outperforms state-of-the-art methods. Furthermore, compared to previous approaches, our method excels in preserving the generalization ability of LLMs on unseen tasks. | [
"Wang, Xiao",
"Chen, Tianze",
"Ge, Qiming",
"Xia, Han",
"Bao, Rong",
"Zheng, Rui",
"Zhang, Qi",
"Gui, Tao",
"Huang, Xuanjing"
] | Orthogonal Subspace Learning for Language Model Continual Learning | findings-emnlp.715 | 2310.14152 | [
"https://github.com/cmnfriend/o-lora"
] | https://huggingface.co/papers/2310.14152 | 0 | 2 | 0 | 9 | [] | [] | [] | 1 | Poster |
https://aclanthology.org/2023.findings-emnlp.716.bib | https://aclanthology.org/2023.findings-emnlp.716/ | @inproceedings{lyu-etal-2023-attention,
title = "Attention-Enhancing Backdoor Attacks Against {BERT}-based Models",
author = "Lyu, Weimin and
Zheng, Songzhu and
Pang, Lu and
Ling, Haibin and
Chen, Chao",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.716",
doi = "10.18653/v1/2023.findings-emnlp.716",
pages = "10672--10690",
abstract = "Recent studies have revealed that Backdoor Attacks can threaten the safety of natural language processing (NLP) models. Investigating the strategies of backdoor attacks will help to understand the model{'}s vulnerability. Most existing textual backdoor attacks focus on generating stealthy triggers or modifying model weights. In this paper, we directly target the interior structure of neural networks and the backdoor mechanism. We propose a novel Trojan Attention Loss (TAL), which enhances the Trojan behavior by directly manipulating the attention patterns. Our loss can be applied to different attacking methods to boost their attack efficacy in terms of attack successful rates and poisoning rates. It applies to not only traditional dirty-label attacks, but also the more challenging clean-label attacks. We validate our method on different backbone models (BERT, RoBERTa, and DistilBERT) and various tasks (Sentiment Analysis, Toxic Detection, and Topic Classification).",
}
| Recent studies have revealed that Backdoor Attacks can threaten the safety of natural language processing (NLP) models. Investigating the strategies of backdoor attacks will help to understand the model{'}s vulnerability. Most existing textual backdoor attacks focus on generating stealthy triggers or modifying model weights. In this paper, we directly target the interior structure of neural networks and the backdoor mechanism. We propose a novel Trojan Attention Loss (TAL), which enhances the Trojan behavior by directly manipulating the attention patterns. Our loss can be applied to different attacking methods to boost their attack efficacy in terms of attack successful rates and poisoning rates. It applies to not only traditional dirty-label attacks, but also the more challenging clean-label attacks. We validate our method on different backbone models (BERT, RoBERTa, and DistilBERT) and various tasks (Sentiment Analysis, Toxic Detection, and Topic Classification). | [
"Lyu, Weimin",
"Zheng, Songzhu",
"Pang, Lu",
"Ling, Haibin",
"Chen, Chao"
] | Attention-Enhancing Backdoor Attacks Against BERT-based Models | findings-emnlp.716 | 2310.14480 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.717.bib | https://aclanthology.org/2023.findings-emnlp.717/ | @inproceedings{wu-etal-2023-hi,
title = "Hi-{T}o{M}: A Benchmark for Evaluating Higher-Order Theory of Mind Reasoning in Large Language Models",
author = "Wu, Yufan and
He, Yinghui and
Jia, Yilin and
Mihalcea, Rada and
Chen, Yulong and
Deng, Naihao",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.717",
doi = "10.18653/v1/2023.findings-emnlp.717",
pages = "10691--10706",
abstract = "Theory of Mind (ToM) is the ability to reason about one{'}s own and others{'} mental states. ToM plays a critical role in the development of intelligence, language understanding, and cognitive processes. While previous work has primarily focused on first and second-order ToM, we explore higher-order ToM, which involves recursive reasoning on others{'} beliefs. {\%}We also incorporate a new deception mechanism in ToM reasoning. We introduce Hi-ToM, a Higher Order Theory of Mind benchmark. Our experimental evaluation using various Large Language Models (LLMs) indicates a decline in performance on higher-order ToM tasks, demonstrating the limitations of current LLMs. We conduct a thorough analysis of different failure cases of LLMs, and share our thoughts on the implications of our findings on the future of NLP.",
}
| Theory of Mind (ToM) is the ability to reason about one{'}s own and others{'} mental states. ToM plays a critical role in the development of intelligence, language understanding, and cognitive processes. While previous work has primarily focused on first and second-order ToM, we explore higher-order ToM, which involves recursive reasoning on others{'} beliefs. {\%}We also incorporate a new deception mechanism in ToM reasoning. We introduce Hi-ToM, a Higher Order Theory of Mind benchmark. Our experimental evaluation using various Large Language Models (LLMs) indicates a decline in performance on higher-order ToM tasks, demonstrating the limitations of current LLMs. We conduct a thorough analysis of different failure cases of LLMs, and share our thoughts on the implications of our findings on the future of NLP. | [
"Wu, Yufan",
"He, Yinghui",
"Jia, Yilin",
"Mihalcea, Rada",
"Chen, Yulong",
"Deng, Naihao"
] | Hi-ToM: A Benchmark for Evaluating Higher-Order Theory of Mind Reasoning in Large Language Models | findings-emnlp.717 | 2310.16755 | [
"https://github.com/ying-hui-he/hi-tom_dataset"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
|
https://aclanthology.org/2023.findings-emnlp.718.bib | https://aclanthology.org/2023.findings-emnlp.718/ | @inproceedings{meunier-etal-2023-image,
title = "Image and Text: Fighting the same Battle? Super Resolution Learning for Imbalanced Text Classification",
author = "Meunier, Romain and
Farah, Benamara and
Moriceau, V{\'e}ronique and
Stolf, Patricia",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.718",
doi = "10.18653/v1/2023.findings-emnlp.718",
pages = "10707--10720",
abstract = "In this paper, we propose SRL4NLP, a new approach for data augmentation by drawing an analogy between image and text processing: Super-resolution learning. This method is based on using high-resolution images to overcome the problem of low resolution images. While this technique is a common usage in image processing when images have a low resolution or are too noisy, it has never been used in NLP. We therefore propose the first adaptation of this method for text classification and evaluate its effectiveness on urgency detection from tweets posted in crisis situations, a very challenging task where messages are scarce and highly imbalanced. We show that this strategy is efficient when compared to competitive state-of-the-art data augmentation techniques on several benchmarks datasets in two languages.",
}
| In this paper, we propose SRL4NLP, a new approach for data augmentation by drawing an analogy between image and text processing: Super-resolution learning. This method is based on using high-resolution images to overcome the problem of low resolution images. While this technique is a common usage in image processing when images have a low resolution or are too noisy, it has never been used in NLP. We therefore propose the first adaptation of this method for text classification and evaluate its effectiveness on urgency detection from tweets posted in crisis situations, a very challenging task where messages are scarce and highly imbalanced. We show that this strategy is efficient when compared to competitive state-of-the-art data augmentation techniques on several benchmarks datasets in two languages. | [
"Meunier, Romain",
"Farah, Benamara",
"Moriceau, V{\\'e}ronique",
"Stolf, Patricia"
] | Image and Text: Fighting the same Battle? Super Resolution Learning for Imbalanced Text Classification | findings-emnlp.718 | [
""
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
||
https://aclanthology.org/2023.findings-emnlp.719.bib | https://aclanthology.org/2023.findings-emnlp.719/ | @inproceedings{mekala-etal-2023-selfood,
title = "{SELFOOD}: Self-Supervised Out-Of-Distribution Detection via Learning to Rank",
author = "Mekala, Dheeraj and
Samavedhi, Adithya and
Dong, Chengyu and
Shang, Jingbo",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.719",
doi = "10.18653/v1/2023.findings-emnlp.719",
pages = "10721--10734",
abstract = "Deep neural classifiers trained with cross-entropy loss (CE loss) often suffer from poor calibration, necessitating the task of out-of-distribution (OOD) detection. Traditional supervised OOD detection methods require expensive manual annotation of in-distribution and OOD samples. To address the annotation bottleneck, we introduce SELFOOD, a self-supervised OOD detection method that requires only in-distribution samples as supervision. We cast OOD detection as an inter-document intra-label (IDIL) ranking problem and train the classifier with our pairwise ranking loss, referred to as IDIL loss. Specifically, given a set of in-distribution documents and their labels, for each label, we train the classifier to rank the softmax scores of documents belonging to that label to be higher than the scores of documents that belong to other labels. Unlike CE loss, our IDIL loss function reaches zero when the desired confidence ranking is achieved and gradients are backpropagated to decrease probabilities associated with incorrect labels rather than continuously increasing the probability of the correct label. Extensive experiments with several classifiers on multiple classification datasets demonstrate the effectiveness of our method in both coarse- and fine-grained settings.",
}
| Deep neural classifiers trained with cross-entropy loss (CE loss) often suffer from poor calibration, necessitating the task of out-of-distribution (OOD) detection. Traditional supervised OOD detection methods require expensive manual annotation of in-distribution and OOD samples. To address the annotation bottleneck, we introduce SELFOOD, a self-supervised OOD detection method that requires only in-distribution samples as supervision. We cast OOD detection as an inter-document intra-label (IDIL) ranking problem and train the classifier with our pairwise ranking loss, referred to as IDIL loss. Specifically, given a set of in-distribution documents and their labels, for each label, we train the classifier to rank the softmax scores of documents belonging to that label to be higher than the scores of documents that belong to other labels. Unlike CE loss, our IDIL loss function reaches zero when the desired confidence ranking is achieved and gradients are backpropagated to decrease probabilities associated with incorrect labels rather than continuously increasing the probability of the correct label. Extensive experiments with several classifiers on multiple classification datasets demonstrate the effectiveness of our method in both coarse- and fine-grained settings. | [
"Mekala, Dheeraj",
"Samavedhi, Adithya",
"Dong, Chengyu",
"Shang, Jingbo"
] | SELFOOD: Self-Supervised Out-Of-Distribution Detection via Learning to Rank | findings-emnlp.719 | 2305.14696 | [
"https://github.com/dheeraj7596/selfood"
] | -1 | -1 | -1 | -1 | [] | [] | [] | 0 | Poster |
Subsets and Splits