Datasets:

NAACL

/

NAACL2024-papers

like 0

Follow

North American Chapter of the Association for Computational Linguistics 1

https://aclanthology.org/2024.naacl-long.401.bib

https://aclanthology.org/2024.naacl-long.401/

@inproceedings{chirkova-nikoulina-2024-key, title = "Key ingredients for effective zero-shot cross-lingual knowledge transfer in generative tasks", author = "Chirkova, Nadezhda and Nikoulina, Vassilina", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.401", doi = "10.18653/v1/2024.naacl-long.401", pages = "7222--7238", abstract = "Zero-shot cross-lingual transfer, which implies finetuning of the multilingual pretrained language model on input-output pairs in one language and using it to make task predictions for inputs in other languages, was widely studied for natural language understanding but is understudied for generation. Previous works notice a frequent problem of generation in a wrong language and propose approaches to address it, usually using mT5 as a backbone model. In this work we compare various approaches proposed from the literature in unified settings, also including alternative backbone models, namely mBART and NLLB-200. We first underline the importance of tuning learning rate used for finetuning, which helps to substantially alleviate the problem of generation in the wrong language. Then, we show that with careful learning rate tuning, the simple full finetuning of the model acts as a very strong baseline and alternative approaches bring only marginal improvements. Finally, we find that mBART performs similarly to mT5 of the same size, and NLLB-200 can be competitive in some cases. Our final zero-shot models reach the performance of the approach based on data translation which is usually considered as an upper baseline for zero-shot cross-lingual transfer in generation.", }

Zero-shot cross-lingual transfer, which implies finetuning of the multilingual pretrained language model on input-output pairs in one language and using it to make task predictions for inputs in other languages, was widely studied for natural language understanding but is understudied for generation. Previous works notice a frequent problem of generation in a wrong language and propose approaches to address it, usually using mT5 as a backbone model. In this work we compare various approaches proposed from the literature in unified settings, also including alternative backbone models, namely mBART and NLLB-200. We first underline the importance of tuning learning rate used for finetuning, which helps to substantially alleviate the problem of generation in the wrong language. Then, we show that with careful learning rate tuning, the simple full finetuning of the model acts as a very strong baseline and alternative approaches bring only marginal improvements. Finally, we find that mBART performs similarly to mT5 of the same size, and NLLB-200 can be competitive in some cases. Our final zero-shot models reach the performance of the approach based on data translation which is usually considered as an upper baseline for zero-shot cross-lingual transfer in generation.

Key ingredients for effective zero-shot cross-lingual knowledge transfer in generative tasks

naacl-long.401

Poster

2402.12279

https://aclanthology.org/2024.naacl-long.402.bib

https://aclanthology.org/2024.naacl-long.402/

@inproceedings{petty-etal-2024-impact, title = "The Impact of Depth on Compositional Generalization in Transformer Language Models", author = "Petty, Jackson and Steenkiste, Sjoerd and Dasgupta, Ishita and Sha, Fei and Garrette, Dan and Linzen, Tal", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.402", doi = "10.18653/v1/2024.naacl-long.402", pages = "7239--7252", abstract = "To process novel sentences, language models (LMs) must generalize compositionally{---}combine familiar elements in new ways. What aspects of a model{'}s structure promote compositional generalization? Focusing on transformers, we test the hypothesis, motivated by theoretical and empirical work, that deeper transformers generalize more compositionally. Simply adding layers increases the total number of parameters; to address this confound between depth and size, we construct three classes of models which trade off depth for width such that the total number of parameters is kept constant (41M, 134M and 374M parameters). We pretrain all models as LMs and fine-tune them on tasks that test for compositional generalization. We report three main conclusions: (1) after fine-tuning, deeper models generalize more compositionally than shallower models do, but the benefit of additional layers diminishes rapidly; (2) within each family, deeper models show better language modeling performance, but returns are similarly diminishing; (3) the benefits of depth for compositional generalization cannot be attributed solely to better performance on language modeling. Because model latency is approximately linear in the number of layers, these results lead us to the recommendation that, with a given total parameter budget, transformers can be made shallower than is typical without sacrificing performance.", }

To process novel sentences, language models (LMs) must generalize compositionally{---}combine familiar elements in new ways. What aspects of a model{'}s structure promote compositional generalization? Focusing on transformers, we test the hypothesis, motivated by theoretical and empirical work, that deeper transformers generalize more compositionally. Simply adding layers increases the total number of parameters; to address this confound between depth and size, we construct three classes of models which trade off depth for width such that the total number of parameters is kept constant (41M, 134M and 374M parameters). We pretrain all models as LMs and fine-tune them on tasks that test for compositional generalization. We report three main conclusions: (1) after fine-tuning, deeper models generalize more compositionally than shallower models do, but the benefit of additional layers diminishes rapidly; (2) within each family, deeper models show better language modeling performance, but returns are similarly diminishing; (3) the benefits of depth for compositional generalization cannot be attributed solely to better performance on language modeling. Because model latency is approximately linear in the number of layers, these results lead us to the recommendation that, with a given total parameter budget, transformers can be made shallower than is typical without sacrificing performance.

The Impact of Depth on Compositional Generalization in Transformer Language Models

naacl-long.402

Poster

2310.19956

https://huggingface.co/papers/2310.19956

https://aclanthology.org/2024.naacl-long.403.bib

https://aclanthology.org/2024.naacl-long.403/

@inproceedings{srikanth-etal-2024-pregnant, title = "Pregnant Questions: The Importance of Pragmatic Awareness in Maternal Health Question Answering", author = "Srikanth, Neha and Sarkar, Rupak and Mane, Heran and Aparicio, Elizabeth and Nguyen, Quynh and Rudinger, Rachel and Boyd-Graber, Jordan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.403", doi = "10.18653/v1/2024.naacl-long.403", pages = "7253--7268", abstract = "Questions posed by information-seeking users often contain implicit false or potentially harmful assumptions. In a high-risk domain such as maternal and infant health, a question-answering system must recognize these pragmatic constraints and go beyond simply answering user questions, examining them in context to respond helpfully. To achieve this, we study assumptions and implications, or pragmatic inferences, made when mothers ask questions about pregnancy and infant care by collecting a dataset of 2,727 inferences from 500 questions across three diverse sources. We study how health experts naturally address these inferences when writing answers, and illustrate that informing existing QA pipelines with pragmatic inferences produces responses that are more complete, mitigating the propagation of harmful beliefs.", }

Questions posed by information-seeking users often contain implicit false or potentially harmful assumptions. In a high-risk domain such as maternal and infant health, a question-answering system must recognize these pragmatic constraints and go beyond simply answering user questions, examining them in context to respond helpfully. To achieve this, we study assumptions and implications, or pragmatic inferences, made when mothers ask questions about pregnancy and infant care by collecting a dataset of 2,727 inferences from 500 questions across three diverse sources. We study how health experts naturally address these inferences when writing answers, and illustrate that informing existing QA pipelines with pragmatic inferences produces responses that are more complete, mitigating the propagation of harmful beliefs.

Pregnant Questions: The Importance of Pragmatic Awareness in Maternal Health Question Answering

naacl-long.403

Poster

2311.09542

https://aclanthology.org/2024.naacl-long.404.bib

https://aclanthology.org/2024.naacl-long.404/

@inproceedings{valvoda-cotterell-2024-towards, title = "Towards Explainability in Legal Outcome Prediction Models", author = "Valvoda, Josef and Cotterell, Ryan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.404", doi = "10.18653/v1/2024.naacl-long.404", pages = "7269--7289", abstract = "Current legal outcome prediction models - a staple of legal NLP - do not explain their reasoning. However, to employ these models in the real world, human legal actors need to be able to understand the model{'}s decisions. In the case of common law, legal practitioners reason towards the outcome of a case by referring to past case law, known as precedent. We contend that precedent is, therefore, a natural way of facilitating explainability for legal NLP models. In this paper, we contribute a novel method for identifying the precedent employed by legal outcome prediction models. Furthermore, by developing a taxonomy of legal precedent, we are able to compare human judges and neural models with respect to the different types of precedent they rely on. We find that while the models learn to predict outcomes reasonably well, their use of precedent is unlike that of human judges.", }

Current legal outcome prediction models - a staple of legal NLP - do not explain their reasoning. However, to employ these models in the real world, human legal actors need to be able to understand the model{'}s decisions. In the case of common law, legal practitioners reason towards the outcome of a case by referring to past case law, known as precedent. We contend that precedent is, therefore, a natural way of facilitating explainability for legal NLP models. In this paper, we contribute a novel method for identifying the precedent employed by legal outcome prediction models. Furthermore, by developing a taxonomy of legal precedent, we are able to compare human judges and neural models with respect to the different types of precedent they rely on. We find that while the models learn to predict outcomes reasonably well, their use of precedent is unlike that of human judges.

Towards Explainability in Legal Outcome Prediction Models

naacl-long.404

Poster

2403.16852

https://aclanthology.org/2024.naacl-long.405.bib

https://aclanthology.org/2024.naacl-long.405/

@inproceedings{li-etal-2024-steerability, title = "The steerability of large language models toward data-driven personas", author = "Li, Junyi and Peris, Charith and Mehrabi, Ninareh and Goyal, Palash and Chang, Kai-Wei and Galstyan, Aram and Zemel, Richard and Gupta, Rahul", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.405", doi = "10.18653/v1/2024.naacl-long.405", pages = "7290--7305", abstract = "Large language models (LLMs) are known to generate biased responses where the opinions of certain groups and populations are underrepresented. Here, we present a novel approach to achieve controllable generation of specific viewpoints using LLMs, that can be leveraged to produce multiple perspectives and to reflect the diverse opinions. Moving beyond the traditional reliance on demographics like age, gender, or party affiliation, we introduce a data-driven notion of persona grounded in collaborative filtering, which is defined as either a single individual or a cohort of individuals manifesting similar views across specific inquiries. As individuals in the same demographic group may have different personas, our data-driven persona definition allows for a more nuanced understanding of different (latent) social groups present in the population. In addition to this, we also explore an efficient method to steer LLMs toward the personas that we define. We show that our data-driven personas significantly enhance model steerability, with improvements of between 57{\%}-77{\%} over our best performing baselines.", }

Large language models (LLMs) are known to generate biased responses where the opinions of certain groups and populations are underrepresented. Here, we present a novel approach to achieve controllable generation of specific viewpoints using LLMs, that can be leveraged to produce multiple perspectives and to reflect the diverse opinions. Moving beyond the traditional reliance on demographics like age, gender, or party affiliation, we introduce a data-driven notion of persona grounded in collaborative filtering, which is defined as either a single individual or a cohort of individuals manifesting similar views across specific inquiries. As individuals in the same demographic group may have different personas, our data-driven persona definition allows for a more nuanced understanding of different (latent) social groups present in the population. In addition to this, we also explore an efficient method to steer LLMs toward the personas that we define. We show that our data-driven personas significantly enhance model steerability, with improvements of between 57{\%}-77{\%} over our best performing baselines.

The steerability of large language models toward data-driven personas

naacl-long.405

Oral

https://aclanthology.org/2024.naacl-long.406.bib

https://aclanthology.org/2024.naacl-long.406/

@inproceedings{jiang-dreyer-2024-ccsum, title = "{CCS}um: A Large-Scale and High-Quality Dataset for Abstractive News Summarization", author = "Jiang, Xiang and Dreyer, Markus", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.406", doi = "10.18653/v1/2024.naacl-long.406", pages = "7306--7336", abstract = "Training a supervised news summarization model requires large amounts of high-quality training data consisting of news articles paired with reference summaries. However, obtaining such data is costly, and existing datasets contain considerable amount of noise. We present a new large-scale and high-quality dataset for supervised abstractive news summarization containing 1.3 million training samples, which we call CCSum. In creating this dataset, we take advantage of the journalistic inverted-pyramid style in news writing: In some articles, the first sentence can be considered a summary of the reported story. Accordingly, among 35 million CommonCrawl News articles, we identify pairs of articles about the same news story and use one article{'}s first sentence as the summary for the other article. To ensure high quality, we apply strict filters whose parameters we optimize using Bayesian optimization. We show that the resulting dataset is more factual and informative than established summarization datasets; less than 1{\%} of the summaries have major factual inconsistencies with the corresponding news articles, compared to 5.5{\%} to 15.4{\%} in existing datasets, according to our human evaluation. Summarization models trained on our dataset are more favored compared to those trained on CNN/Daily Mail. The proposed dataset can open new opportunities for future research in abstractive summarization.", }

Training a supervised news summarization model requires large amounts of high-quality training data consisting of news articles paired with reference summaries. However, obtaining such data is costly, and existing datasets contain considerable amount of noise. We present a new large-scale and high-quality dataset for supervised abstractive news summarization containing 1.3 million training samples, which we call CCSum. In creating this dataset, we take advantage of the journalistic inverted-pyramid style in news writing: In some articles, the first sentence can be considered a summary of the reported story. Accordingly, among 35 million CommonCrawl News articles, we identify pairs of articles about the same news story and use one article{'}s first sentence as the summary for the other article. To ensure high quality, we apply strict filters whose parameters we optimize using Bayesian optimization. We show that the resulting dataset is more factual and informative than established summarization datasets; less than 1{\%} of the summaries have major factual inconsistencies with the corresponding news articles, compared to 5.5{\%} to 15.4{\%} in existing datasets, according to our human evaluation. Summarization models trained on our dataset are more favored compared to those trained on CNN/Daily Mail. The proposed dataset can open new opportunities for future research in abstractive summarization.

CCSum: A Large-Scale and High-Quality Dataset for Abstractive News Summarization

naacl-long.406

Oral

https://aclanthology.org/2024.naacl-long.407.bib

https://aclanthology.org/2024.naacl-long.407/

@inproceedings{mokhberian-etal-2024-capturing, title = "Capturing Perspectives of Crowdsourced Annotators in Subjective Learning Tasks", author = "Mokhberian, Negar and Marmarelis, Myrl and Hopp, Frederic and Basile, Valerio and Morstatter, Fred and Lerman, Kristina", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.407", doi = "10.18653/v1/2024.naacl-long.407", pages = "7337--7349", abstract = "Supervised classification heavily depends on datasets annotated by humans. However, in subjective tasks such as toxicity classification, these annotations often exhibit low agreement among raters. Annotations have commonly been aggregated by employing methods like majority voting to determine a single ground truth label. In subjective tasks, aggregating labels will result in biased labeling and, consequently, biased models that can overlook minority opinions. Previous studies have shed light on the pitfalls of label aggregation and have introduced a handful of practical approaches to tackle this issue. Recently proposed multi-annotator models, which predict labels individually per annotator, are vulnerable to under-determination for annotators with few samples. This problem is exacerbated in crowdsourced datasets. In this work, we propose Annotator Aware Representations for Texts (AART) for subjective classification tasks. Our approach involves learning representations of annotators, allowing for exploration of annotation behaviors. We show the improvement of our method on metrics that assess the performance on capturing individual annotators{'} perspectives. Additionally, we demonstrate fairness metrics to evaluate our model{'}s equability of performance for marginalized annotators compared to others.", }

Supervised classification heavily depends on datasets annotated by humans. However, in subjective tasks such as toxicity classification, these annotations often exhibit low agreement among raters. Annotations have commonly been aggregated by employing methods like majority voting to determine a single ground truth label. In subjective tasks, aggregating labels will result in biased labeling and, consequently, biased models that can overlook minority opinions. Previous studies have shed light on the pitfalls of label aggregation and have introduced a handful of practical approaches to tackle this issue. Recently proposed multi-annotator models, which predict labels individually per annotator, are vulnerable to under-determination for annotators with few samples. This problem is exacerbated in crowdsourced datasets. In this work, we propose Annotator Aware Representations for Texts (AART) for subjective classification tasks. Our approach involves learning representations of annotators, allowing for exploration of annotation behaviors. We show the improvement of our method on metrics that assess the performance on capturing individual annotators{'} perspectives. Additionally, we demonstrate fairness metrics to evaluate our model{'}s equability of performance for marginalized annotators compared to others.

Capturing Perspectives of Crowdsourced Annotators in Subjective Learning Tasks

naacl-long.407

Poster

2311.09743

https://aclanthology.org/2024.naacl-long.408.bib

https://aclanthology.org/2024.naacl-long.408/

@inproceedings{sundararajan-etal-2024-improving, title = "Improving Factual Accuracy of Neural Table-to-Text Output by Addressing Input Problems in {T}o{TT}o", author = "Sundararajan, Barkavi and Sripada, Yaji and Reiter, Ehud", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.408", doi = "10.18653/v1/2024.naacl-long.408", pages = "7350--7376", abstract = "Neural Table-to-Text models tend to hallucinate, producing texts that contain factual errors. We investigate whether such errors in the output can be traced back to problems with the input. We manually annotated 1,837 texts generated by multiple models in the politics domain of the ToTTo dataset. We identify the input problems that are responsible for many output errors and show that fixing these inputs reduces factual errors by between 52{\%} and 76{\%} (depending on the model). In addition, we observe that models struggle in processing tabular inputs that are structured in a non-standard way, particularly when the input lacks distinct row and column values or when the column headers are not correctly mapped to corresponding values.", }

Neural Table-to-Text models tend to hallucinate, producing texts that contain factual errors. We investigate whether such errors in the output can be traced back to problems with the input. We manually annotated 1,837 texts generated by multiple models in the politics domain of the ToTTo dataset. We identify the input problems that are responsible for many output errors and show that fixing these inputs reduces factual errors by between 52{\%} and 76{\%} (depending on the model). In addition, we observe that models struggle in processing tabular inputs that are structured in a non-standard way, particularly when the input lacks distinct row and column values or when the column headers are not correctly mapped to corresponding values.

Improving Factual Accuracy of Neural Table-to-Text Output by Addressing Input Problems in ToTTo

naacl-long.408

Poster

2404.04103

https://aclanthology.org/2024.naacl-long.409.bib

https://aclanthology.org/2024.naacl-long.409/

@inproceedings{cai-etal-2024-ceret, title = "{CERET}: Cost-Effective Extrinsic Refinement for Text Generation", author = "Cai, Jason and Su, Hang and Sunkara, Monica and Shalyminov, Igor and Mansour, Saab", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.409", doi = "10.18653/v1/2024.naacl-long.409", pages = "7377--7390", abstract = "Large Language Models (LLMs) are powerful models for generation tasks, but they may not generate good quality outputs in their first attempt. Apart from model fine-tuning, existing approaches to improve prediction accuracy and quality typically involve LLM self-improvement / self-reflection that incorporate feedback from models themselves. Despite their effectiveness, these methods are hindered by their high computational cost and lack of scalability. In this work, we propose CERET, a method for refining text generations by considering semantic stability, entailment and inter-sample uncertainty measures. Experimental results show that CERET outperforms Self-consistency and Self-rerank baselines consistently under various task setups, by 1.6{\%} in Rouge-1 for abstractive summarization and 3.5{\%} in hit rate for question answering. Compared to LLM Self-rerank method, our approach only requires 9.4{\%} of its latency and is more cost-effective.", }

Large Language Models (LLMs) are powerful models for generation tasks, but they may not generate good quality outputs in their first attempt. Apart from model fine-tuning, existing approaches to improve prediction accuracy and quality typically involve LLM self-improvement / self-reflection that incorporate feedback from models themselves. Despite their effectiveness, these methods are hindered by their high computational cost and lack of scalability. In this work, we propose CERET, a method for refining text generations by considering semantic stability, entailment and inter-sample uncertainty measures. Experimental results show that CERET outperforms Self-consistency and Self-rerank baselines consistently under various task setups, by 1.6{\%} in Rouge-1 for abstractive summarization and 3.5{\%} in hit rate for question answering. Compared to LLM Self-rerank method, our approach only requires 9.4{\%} of its latency and is more cost-effective.

CERET: Cost-Effective Extrinsic Refinement for Text Generation

naacl-long.409

Poster

2406.05588

https://aclanthology.org/2024.naacl-long.410.bib

https://aclanthology.org/2024.naacl-long.410/

@inproceedings{khatuya-etal-2024-parameter, title = "Parameter-Efficient Instruction Tuning of Large Language Models For Extreme Financial Numeral Labelling", author = "Khatuya, Subhendu and Mukherjee, Rajdeep and Ghosh, Akash and Hegde, Manjunath and Dasgupta, Koustuv and Ganguly, Niloy and Ghosh, Saptarshi and Goyal, Pawan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.410", doi = "10.18653/v1/2024.naacl-long.410", pages = "7391--7403", abstract = "We study the problem of automatically annotating relevant numerals (GAAP metrics) occurring in the financial documents with their corresponding XBRL tags. Different from prior works, we investigate the feasibility of solving this extreme classification problem using a generative paradigm through instruction tuning of Large Language Models (LLMs). To this end, we leverage metric metadata informationto frame our target outputs while proposing a parameter efficient solution for the task using LoRA. We perform experiments on two recently released financial numeric labeling datasets. Our proposed model, **FLAN-FinXC**, achieves new state-of-the-art performances on both the datasets, outperforming several strong baselines. We explain the better scores of our proposed model by demonstrating its capability for zero-shot as well as the least frequently occurring tags. Also, even when we fail to predict the XBRL tags correctly, our generated output has substantial overlap with the ground-truth in majority of the cases.", }

We study the problem of automatically annotating relevant numerals (GAAP metrics) occurring in the financial documents with their corresponding XBRL tags. Different from prior works, we investigate the feasibility of solving this extreme classification problem using a generative paradigm through instruction tuning of Large Language Models (LLMs). To this end, we leverage metric metadata informationto frame our target outputs while proposing a parameter efficient solution for the task using LoRA. We perform experiments on two recently released financial numeric labeling datasets. Our proposed model, **FLAN-FinXC**, achieves new state-of-the-art performances on both the datasets, outperforming several strong baselines. We explain the better scores of our proposed model by demonstrating its capability for zero-shot as well as the least frequently occurring tags. Also, even when we fail to predict the XBRL tags correctly, our generated output has substantial overlap with the ground-truth in majority of the cases.

Parameter-Efficient Instruction Tuning of Large Language Models For Extreme Financial Numeral Labelling

naacl-long.410

Oral

2405.06671

https://aclanthology.org/2024.naacl-long.411.bib

https://aclanthology.org/2024.naacl-long.411/

@inproceedings{davoodi-goldwasser-2024-analysis, title = "Analysis of State-Level Legislative Process in Enhanced Linguistic and Nationwide Network Contexts", author = "Davoodi, Maryam and Goldwasser, Dan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.411", doi = "10.18653/v1/2024.naacl-long.411", pages = "7404--7422", abstract = "State bills have a significant impact on various aspects of society, including health, education, and the economy. Consequently, it is crucial to conduct systematic research on state bills before and after they are enacted to evaluate their benefits and drawbacks, thereby guiding future decision-making. In this work, we developed the first state-level deep learning framework that (1) handles the complex and inconsistent language of policies across US states using generative large language models and (2) decodes legislators{'} behavior and implications of state policies by establishing a shared nationwide network, enriched with diverse contexts, such as information on interest groups influencing public policy and legislators{'} courage test results, which reflect their political positions.", }

State bills have a significant impact on various aspects of society, including health, education, and the economy. Consequently, it is crucial to conduct systematic research on state bills before and after they are enacted to evaluate their benefits and drawbacks, thereby guiding future decision-making. In this work, we developed the first state-level deep learning framework that (1) handles the complex and inconsistent language of policies across US states using generative large language models and (2) decodes legislators{'} behavior and implications of state policies by establishing a shared nationwide network, enriched with diverse contexts, such as information on interest groups influencing public policy and legislators{'} courage test results, which reflect their political positions.

Analysis of State-Level Legislative Process in Enhanced Linguistic and Nationwide Network Contexts

naacl-long.411

Poster

https://aclanthology.org/2024.naacl-long.412.bib

https://aclanthology.org/2024.naacl-long.412/

@inproceedings{khanuja-etal-2024-demux, title = "{D}e{M}u{X}: Data-efficient Multilingual Learning", author = "Khanuja, Simran and Gowriraj, Srinivas and Dery, Lucio and Neubig, Graham", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.412", doi = "10.18653/v1/2024.naacl-long.412", pages = "7423--7436", abstract = "Pre-trained multilingual models have enabled deployment of NLP technologies for multiple languages. However, optimally fine-tuning these models under an annotation budget, such that performance on desired target languages is jointly maximized, still remains an open question. In this paper, we introduce DeMuX, a framework that prescribes the exact data-points to label from vast amounts of unlabelled multilingual data, having unknown degrees of overlap with the target set. Unlike most prior works, our end-to-end framework is language-agnostic, accounts for model representations, and supports multilingual target configurations. Our active learning strategies rely upon distance and uncertainty measures to select task-specific neighbors that are most informative to label, given a model. DeMuX outperforms strong baselines in 84{\%} of the test cases, in the zero-shot setting of disjoint source and target language sets (including multilingual target pools), across three models and four tasks. Notably, in low-budget settings (5-100 examples), we observe gains of up to 8-11 F1 points. Our code is released here: https://github.com/simran-khanuja/demux.", }

Pre-trained multilingual models have enabled deployment of NLP technologies for multiple languages. However, optimally fine-tuning these models under an annotation budget, such that performance on desired target languages is jointly maximized, still remains an open question. In this paper, we introduce DeMuX, a framework that prescribes the exact data-points to label from vast amounts of unlabelled multilingual data, having unknown degrees of overlap with the target set. Unlike most prior works, our end-to-end framework is language-agnostic, accounts for model representations, and supports multilingual target configurations. Our active learning strategies rely upon distance and uncertainty measures to select task-specific neighbors that are most informative to label, given a model. DeMuX outperforms strong baselines in 84{\%} of the test cases, in the zero-shot setting of disjoint source and target language sets (including multilingual target pools), across three models and four tasks. Notably, in low-budget settings (5-100 examples), we observe gains of up to 8-11 F1 points. Our code is released here: https://github.com/simran-khanuja/demux.

DeMuX: Data-efficient Multilingual Learning

naacl-long.412

Poster

2311.06379

https://aclanthology.org/2024.naacl-long.413.bib

https://aclanthology.org/2024.naacl-long.413/

@inproceedings{chandradevan-etal-2024-duqgen, title = "{DUQG}en: Effective Unsupervised Domain Adaptation of Neural Rankers by Diversifying Synthetic Query Generation", author = "Chandradevan, Ramraj and Dhole, Kaustubh and Agichtein, Eugene", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.413", doi = "10.18653/v1/2024.naacl-long.413", pages = "7437--7451", abstract = "State-of-the-art neural rankers pre-trained on large task-specific training data such as MS-MARCO, have been shown to exhibit strong performance on various ranking tasks without domain adaptation, also called zero-shot. However, zero-shot neural ranking may be sub-optimal, as it does not take advantage of the target domain information. Unfortunately, acquiring sufficiently large and high quality target training data to improve a modern neural ranker can be costly and time-consuming. To address this problem, we propose a new approach to unsupervised domain adaptation for ranking, DUQGen, which addresses a critical gap in prior literature, namely how to automatically generate both effective and diverse synthetic training data to fine tune a modern neural ranker for a new domain. Specifically, DUQGen produces a more effective representation of the target domain by identifying clusters of similar documents; and generates a more diverse training dataset by probabilistic sampling over the resulting document clusters. Our extensive experiments, over the standard BEIR collection, demonstrate that DUQGen consistently outperforms all zero-shot baselines and substantially outperforms the SOTA baselines on 16 out of 18 datasets, for an average of 4{\%} relative improvement across all datasets. We complement our results with a thorough analysis for more in-depth understanding of the proposed method{'}s performance and to identify promising areas for further improvements.", }

State-of-the-art neural rankers pre-trained on large task-specific training data such as MS-MARCO, have been shown to exhibit strong performance on various ranking tasks without domain adaptation, also called zero-shot. However, zero-shot neural ranking may be sub-optimal, as it does not take advantage of the target domain information. Unfortunately, acquiring sufficiently large and high quality target training data to improve a modern neural ranker can be costly and time-consuming. To address this problem, we propose a new approach to unsupervised domain adaptation for ranking, DUQGen, which addresses a critical gap in prior literature, namely how to automatically generate both effective and diverse synthetic training data to fine tune a modern neural ranker for a new domain. Specifically, DUQGen produces a more effective representation of the target domain by identifying clusters of similar documents; and generates a more diverse training dataset by probabilistic sampling over the resulting document clusters. Our extensive experiments, over the standard BEIR collection, demonstrate that DUQGen consistently outperforms all zero-shot baselines and substantially outperforms the SOTA baselines on 16 out of 18 datasets, for an average of 4{\%} relative improvement across all datasets. We complement our results with a thorough analysis for more in-depth understanding of the proposed method{'}s performance and to identify promising areas for further improvements.

DUQGen: Effective Unsupervised Domain Adaptation of Neural Rankers by Diversifying Synthetic Query Generation

naacl-long.413

Oral

2404.02489

https://huggingface.co/papers/2404.02489

https://aclanthology.org/2024.naacl-long.414.bib

https://aclanthology.org/2024.naacl-long.414/

@inproceedings{hua-etal-2024-get, title = "How did we get here? Summarizing conversation dynamics", author = "Hua, Yilun and Chernogor, Nicholas and Gu, Yuzhe and Jeong, Seoyeon and Luo, Miranda and Danescu-Niculescu-Mizil, Cristian", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.414", doi = "10.18653/v1/2024.naacl-long.414", pages = "7452--7477", abstract = "Throughout a conversation, the way participants interact with each other is in constant flux: their tones may change, they may resort to different strategies to convey their points, or they might alter their interaction patterns. An understanding of these dynamics can complement that of the actual facts and opinions discussed, offering a more holistic view of the trajectory of the conversation: how it arrived at its current state and where it is likely heading.In this work, we introduce the task of summarizing the dynamics of conversations, by constructing a dataset of human-written summaries, and exploring several automated baselines. We evaluate whether such summaries can capture the trajectory of conversations via an established downstream task: forecasting whether an ongoing conversation will eventually derail into toxic behavior. We show that they help both humans and automated systems with this forecasting task. Humans make predictions three times faster, and with greater confidence, when reading the summaries than when reading the transcripts. Furthermore, automated forecasting systems are more accurate when constructing, and then predicting based on, summaries of conversation dynamics, compared to directly predicting on the transcripts.", }

Throughout a conversation, the way participants interact with each other is in constant flux: their tones may change, they may resort to different strategies to convey their points, or they might alter their interaction patterns. An understanding of these dynamics can complement that of the actual facts and opinions discussed, offering a more holistic view of the trajectory of the conversation: how it arrived at its current state and where it is likely heading.In this work, we introduce the task of summarizing the dynamics of conversations, by constructing a dataset of human-written summaries, and exploring several automated baselines. We evaluate whether such summaries can capture the trajectory of conversations via an established downstream task: forecasting whether an ongoing conversation will eventually derail into toxic behavior. We show that they help both humans and automated systems with this forecasting task. Humans make predictions three times faster, and with greater confidence, when reading the summaries than when reading the transcripts. Furthermore, automated forecasting systems are more accurate when constructing, and then predicting based on, summaries of conversation dynamics, compared to directly predicting on the transcripts.

How did we get here? Summarizing conversation dynamics

naacl-long.414

Poster

2404.19007

https://aclanthology.org/2024.naacl-long.415.bib

https://aclanthology.org/2024.naacl-long.415/

@inproceedings{cho-etal-2024-language, title = "Can Language Model Moderators Improve the Health of Online Discourse?", author = "Cho, Hyundong and Liu, Shuai and Shi, Taiwei and Jain, Darpan and Rizk, Basem and Huang, Yuyang and Lu, Zixun and Wen, Nuan and Gratch, Jonathan and Ferrara, Emilio and May, Jonathan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.415", doi = "10.18653/v1/2024.naacl-long.415", pages = "7478--7496", abstract = "Conversational moderation of online communities is crucial to maintaining civility for a constructive environment, but it is challenging to scale and harmful to moderators. The inclusion of sophisticated natural language generation modules as a force multiplier to aid human moderators is a tantalizing prospect, but adequate evaluation approaches have so far been elusive. In this paper, we establish a systematic definition of conversational moderation effectiveness grounded on moderation literature and establish design criteria for conducting realistic yet safe evaluation. We then propose a comprehensive evaluation framework to assess models{'} moderation capabilities independently of human intervention. With our framework, we conduct the first known study of language models as conversational moderators, finding that appropriately prompted models that incorporate insights from social science can provide specific and fair feedback on toxic behavior but struggle to influence users to increase their levels of respect and cooperation.", }

Conversational moderation of online communities is crucial to maintaining civility for a constructive environment, but it is challenging to scale and harmful to moderators. The inclusion of sophisticated natural language generation modules as a force multiplier to aid human moderators is a tantalizing prospect, but adequate evaluation approaches have so far been elusive. In this paper, we establish a systematic definition of conversational moderation effectiveness grounded on moderation literature and establish design criteria for conducting realistic yet safe evaluation. We then propose a comprehensive evaluation framework to assess models{'} moderation capabilities independently of human intervention. With our framework, we conduct the first known study of language models as conversational moderators, finding that appropriately prompted models that incorporate insights from social science can provide specific and fair feedback on toxic behavior but struggle to influence users to increase their levels of respect and cooperation.

Can Language Model Moderators Improve the Health of Online Discourse?

naacl-long.415

Poster

2311.10781

https://aclanthology.org/2024.naacl-long.416.bib

https://aclanthology.org/2024.naacl-long.416/

@inproceedings{jiang-etal-2024-leanreasoner, title = "{L}ean{R}easoner: Boosting Complex Logical Reasoning with Lean", author = "Jiang, Dongwei and Fonseca, Marcio and Cohen, Shay", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.416", doi = "10.18653/v1/2024.naacl-long.416", pages = "7497--7510", abstract = "Large language models (LLMs) often struggle with complex logical reasoning due to logical inconsistencies and the inherent difficulty ofsuch reasoning. We use Lean, a theorem proving framework, to address these challenges. By formalizing logical reasoning problems intotheorems within Lean, we can solve them by proving or disproving the corresponding theorems. This method reduces the risk of logical inconsistencies with the help of Lean{'}s symbolic solver. It also enhances our ability to treat complex reasoning tasks using Lean{'}s extensive library of theorem proofs. Our method achieves state-of-the-art performance on the FOLIO dataset and achieves performance near this level on ProofWriter. Notably, these results were accomplished by fine-tuning on fewer than 100 in-domain samples for each dataset", }

Large language models (LLMs) often struggle with complex logical reasoning due to logical inconsistencies and the inherent difficulty ofsuch reasoning. We use Lean, a theorem proving framework, to address these challenges. By formalizing logical reasoning problems intotheorems within Lean, we can solve them by proving or disproving the corresponding theorems. This method reduces the risk of logical inconsistencies with the help of Lean{'}s symbolic solver. It also enhances our ability to treat complex reasoning tasks using Lean{'}s extensive library of theorem proofs. Our method achieves state-of-the-art performance on the FOLIO dataset and achieves performance near this level on ProofWriter. Notably, these results were accomplished by fine-tuning on fewer than 100 in-domain samples for each dataset

LeanReasoner: Boosting Complex Logical Reasoning with Lean

naacl-long.416

Poster

2403.13312

https://aclanthology.org/2024.naacl-long.417.bib

https://aclanthology.org/2024.naacl-long.417/

@inproceedings{wu-2024-uicoder, title = "{UIC}oder: Finetuning Large Language Models to Generate User Interface Code through Automated Feedback", author = "Wu, Jason and Schoop, Eldon and Leung, Alan and Barik, Titus and Bigham, Jeffrey and Nichols, Jeffrey", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.417", doi = "10.18653/v1/2024.naacl-long.417", pages = "7511--7525", abstract = "Many large language models (LLMs) struggle to consistently generate UI code that compiles and produces visually relevant designs. Existing approaches to improve generation rely either on expensive human feedback or distilling a proprietary model. In this paper, we explore the use of automated feedback (compilers and multi-modal models) to guide LLMs to generate high-quality UI code. Our method starts with an existing LLM and iteratively produces improved models by self-generating a large synthetic dataset using an original model, applying automated tools to aggressively filter, score, and de-duplicate the data into a refined higher quality dataset, and producing a new LLM by finetuning the original on the refined dataset.We applied our approach to several open-source LLMs and compared the resulting performance to baseline models with both automated metrics and human preferences.Our results show the resulting models outperform all other downloadable baselines and approach the performance of larger proprietary models.", }

Many large language models (LLMs) struggle to consistently generate UI code that compiles and produces visually relevant designs. Existing approaches to improve generation rely either on expensive human feedback or distilling a proprietary model. In this paper, we explore the use of automated feedback (compilers and multi-modal models) to guide LLMs to generate high-quality UI code. Our method starts with an existing LLM and iteratively produces improved models by self-generating a large synthetic dataset using an original model, applying automated tools to aggressively filter, score, and de-duplicate the data into a refined higher quality dataset, and producing a new LLM by finetuning the original on the refined dataset.We applied our approach to several open-source LLMs and compared the resulting performance to baseline models with both automated metrics and human preferences.Our results show the resulting models outperform all other downloadable baselines and approach the performance of larger proprietary models.

UICoder: Finetuning Large Language Models to Generate User Interface Code through Automated Feedback

naacl-long.417

Poster

2406.07739

https://huggingface.co/papers/2406.07739

https://aclanthology.org/2024.naacl-long.418.bib

https://aclanthology.org/2024.naacl-long.418/

@inproceedings{de-souza-etal-2024-measuring, title = "Measuring Cross-lingual Transfer in Bytes", author = "De Souza, Leandro and Almeida, Thales and Lotufo, Roberto and Frassetto Nogueira, Rodrigo", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.418", doi = "10.18653/v1/2024.naacl-long.418", pages = "7526--7537", abstract = "Multilingual pretraining has been a successful solution to the challenges posed by the lack of resources for languages. These models can transfer knowledge to target languages with minimal or no examples. Recent research suggests that monolingual models also have a similar capability, but the mechanisms behind this transfer remain unclear. Some studies have explored factors like language contamination and syntactic similarity. An emerging line of research suggests that the representations learned by language models contain two components: a language-specific and a language-agnostic component. The latter is responsible for transferring a more universal knowledge. However, there is a lack of comprehensive exploration of these properties across diverse target languages. To investigate this hypothesis, we conducted an experiment inspired by the work on the Scaling Laws for Transfer. We measured the amount of data transferred from a source language to a target language and found that models initialized from diverse languages perform similarly to a target language in a cross-lingual setting. This was surprising because the amount of data transferred to 10 diverse target languages, such as Spanish, Korean, and Finnish, was quite similar. We also found evidence that this transfer is not related to language contamination or language proximity, which strengthens the hypothesis that the model also relies on language-agnostic knowledge. Our experiments have opened up new possibilities for measuring how much data represents the language-agnostic representations learned during pretraining.", }

Multilingual pretraining has been a successful solution to the challenges posed by the lack of resources for languages. These models can transfer knowledge to target languages with minimal or no examples. Recent research suggests that monolingual models also have a similar capability, but the mechanisms behind this transfer remain unclear. Some studies have explored factors like language contamination and syntactic similarity. An emerging line of research suggests that the representations learned by language models contain two components: a language-specific and a language-agnostic component. The latter is responsible for transferring a more universal knowledge. However, there is a lack of comprehensive exploration of these properties across diverse target languages. To investigate this hypothesis, we conducted an experiment inspired by the work on the Scaling Laws for Transfer. We measured the amount of data transferred from a source language to a target language and found that models initialized from diverse languages perform similarly to a target language in a cross-lingual setting. This was surprising because the amount of data transferred to 10 diverse target languages, such as Spanish, Korean, and Finnish, was quite similar. We also found evidence that this transfer is not related to language contamination or language proximity, which strengthens the hypothesis that the model also relies on language-agnostic knowledge. Our experiments have opened up new possibilities for measuring how much data represents the language-agnostic representations learned during pretraining.

Measuring Cross-lingual Transfer in Bytes

naacl-long.418

Poster

2404.08191

https://aclanthology.org/2024.naacl-long.419.bib

https://aclanthology.org/2024.naacl-long.419/

@inproceedings{hossain-etal-2024-misgendermender, title = "{M}isgender{M}ender: A Community-Informed Approach to Interventions for Misgendering", author = "Hossain, Tamanna and Dev, Sunipa and Singh, Sameer", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.419", doi = "10.18653/v1/2024.naacl-long.419", pages = "7538--7558", abstract = "Content Warning: This paper contains examples of misgendering and erasure that could be offensive and potentially triggering.Misgendering, the act of incorrectly addressing someone{'}s gender, inflicts serious harm and is pervasive in everyday technologies, yet there is a notable lack of research to combat it. We are the first to address this lack of research into interventions for misgendering by conducting a survey of gender-diverse individuals in the US to understand perspectives about automated interventions for text-based misgendering. Based on survey insights on the prevalence of misgendering, desired solutions, and associated concerns, we introduce a misgendering interventions task and evaluation dataset, MisgenderMender. We define the task with two sub-tasks: (i) detecting misgendering, followed by (ii) correcting misgendering where misgendering is present, in domains where editing is appropriate. MisgenderMender comprises 3790 instances of social media content and LLM-generations about non-cisgender public figures, annotated for the presence of misgendering, with additional annotations for correcting misgendering in LLM-generated text. Using this dataset, we set initial benchmarks by evaluating existing NLP systems and highlighting challenges for future models to address. We release the full dataset, code, and demo at https://tamannahossainkay.github.io/misgendermender/", }

Content Warning: This paper contains examples of misgendering and erasure that could be offensive and potentially triggering.Misgendering, the act of incorrectly addressing someone{'}s gender, inflicts serious harm and is pervasive in everyday technologies, yet there is a notable lack of research to combat it. We are the first to address this lack of research into interventions for misgendering by conducting a survey of gender-diverse individuals in the US to understand perspectives about automated interventions for text-based misgendering. Based on survey insights on the prevalence of misgendering, desired solutions, and associated concerns, we introduce a misgendering interventions task and evaluation dataset, MisgenderMender. We define the task with two sub-tasks: (i) detecting misgendering, followed by (ii) correcting misgendering where misgendering is present, in domains where editing is appropriate. MisgenderMender comprises 3790 instances of social media content and LLM-generations about non-cisgender public figures, annotated for the presence of misgendering, with additional annotations for correcting misgendering in LLM-generated text. Using this dataset, we set initial benchmarks by evaluating existing NLP systems and highlighting challenges for future models to address. We release the full dataset, code, and demo at https://tamannahossainkay.github.io/misgendermender/

MisgenderMender: A Community-Informed Approach to Interventions for Misgendering

naacl-long.419

Poster

2404.14695

https://aclanthology.org/2024.naacl-long.420.bib

https://aclanthology.org/2024.naacl-long.420/

@inproceedings{chen-etal-2024-interplay, title = "Interplay of Machine Translation, Diacritics, and Diacritization", author = "Chen, Wei-Rui and Adebara, Ife and Abdul-Mageed, Muhammad", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.420", doi = "10.18653/v1/2024.naacl-long.420", pages = "7559--7601", abstract = "We investigate two research questions: (1) how do machine translation (MT) and diacritization influence the performance of each other in a multi-task learning setting (2) the effect of keeping (vs. removing) diacritics on MT performance. We examine these two questions in both high-resource (HR) and low-resource (LR) settings across 55 different languages (36 African languages and 19 European languages). For (1), results show that diacritization significantly benefits MT in the LR scenario, doubling or even tripling performance for some languages, but harms MT in the HR scenario. We find that MT harms diacritization in LR but benefits significantly in HR for some languages. For (2), MT performance is similar regardless of diacritics being kept or removed. In addition, we propose two classes of metrics to measure the complexity of a diacritical system, finding these metrics to correlate positively with the performance of our diacritization models. Overall, our work provides insights for developing MT and diacritization systems under different data size conditions and may have implications that generalize beyond the 55 languages we investigate.", }

We investigate two research questions: (1) how do machine translation (MT) and diacritization influence the performance of each other in a multi-task learning setting (2) the effect of keeping (vs. removing) diacritics on MT performance. We examine these two questions in both high-resource (HR) and low-resource (LR) settings across 55 different languages (36 African languages and 19 European languages). For (1), results show that diacritization significantly benefits MT in the LR scenario, doubling or even tripling performance for some languages, but harms MT in the HR scenario. We find that MT harms diacritization in LR but benefits significantly in HR for some languages. For (2), MT performance is similar regardless of diacritics being kept or removed. In addition, we propose two classes of metrics to measure the complexity of a diacritical system, finding these metrics to correlate positively with the performance of our diacritization models. Overall, our work provides insights for developing MT and diacritization systems under different data size conditions and may have implications that generalize beyond the 55 languages we investigate.

Interplay of Machine Translation, Diacritics, and Diacritization

naacl-long.420

Poster

2404.05943

https://aclanthology.org/2024.naacl-long.421.bib

https://aclanthology.org/2024.naacl-long.421/

@inproceedings{li-etal-2024-quantity, title = "From Quantity to Quality: Boosting {LLM} Performance with Self-Guided Data Selection for Instruction Tuning", author = "Li, Ming and Zhang, Yong and Li, Zhitao and Chen, Jiuhai and Chen, Lichang and Cheng, Ning and Wang, Jianzong and Zhou, Tianyi and Xiao, Jing", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.421", doi = "10.18653/v1/2024.naacl-long.421", pages = "7602--7635", abstract = "In the realm of Large Language Models (LLMs), the balance between instruction data quality and quantity is a focal point. Recognizing this, we introduce a self-guided methodology for LLMs to autonomously discern and select cherry samples from open-source datasets, effectively minimizing manual curation and potential cost for instruction tuning an LLM. Our key innovation, the Instruction-Following Difficulty (IFD) metric, emerges as a pivotal metric to identify discrepancies between a model{'}s expected responses and its intrinsic generation capability. Through the application of IFD, cherry samples can be pinpointed, leading to a marked uptick in model training efficiency. Empirical validations on datasets like Alpaca and WizardLM underpin our findings; with a mere 10{\%} of original data input, our strategy showcases improved results. This synthesis of self-guided cherry-picking and the IFD metric signifies a transformative leap in the instruction tuning of LLMs, promising both efficiency and resource-conscious advancements. Codes, data, and models are available.", }

In the realm of Large Language Models (LLMs), the balance between instruction data quality and quantity is a focal point. Recognizing this, we introduce a self-guided methodology for LLMs to autonomously discern and select cherry samples from open-source datasets, effectively minimizing manual curation and potential cost for instruction tuning an LLM. Our key innovation, the Instruction-Following Difficulty (IFD) metric, emerges as a pivotal metric to identify discrepancies between a model{'}s expected responses and its intrinsic generation capability. Through the application of IFD, cherry samples can be pinpointed, leading to a marked uptick in model training efficiency. Empirical validations on datasets like Alpaca and WizardLM underpin our findings; with a mere 10{\%} of original data input, our strategy showcases improved results. This synthesis of self-guided cherry-picking and the IFD metric signifies a transformative leap in the instruction tuning of LLMs, promising both efficiency and resource-conscious advancements. Codes, data, and models are available.

From Quantity to Quality: Boosting LLM Performance with Self-Guided Data Selection for Instruction Tuning

naacl-long.421

Poster

2308.12032

https://aclanthology.org/2024.naacl-long.422.bib

https://aclanthology.org/2024.naacl-long.422/

@inproceedings{shi-etal-2024-safer, title = "Safer-Instruct: Aligning Language Models with Automated Preference Data", author = "Shi, Taiwei and Chen, Kai and Zhao, Jieyu", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.422", doi = "10.18653/v1/2024.naacl-long.422", pages = "7636--7651", abstract = "Reinforcement learning from human feedback (RLHF) is a vital strategy for enhancing model capability in language models. However, annotating preference data for RLHF is a resource-intensive and creativity-demanding process, while existing automatic generation methods face limitations in data diversity and quality. In response, we present Safer-Instruct, a novel pipeline for automatically constructing large-scale preference data. Our approach leverages reversed instruction tuning, instruction induction, and expert model evaluation to efficiently generate high-quality preference data without human annotators. To verify the effectiveness of Safer-Instruct, we apply the pipeline to construct a safety preference dataset as a case study. Finetuning an Alpaca model on this synthetic dataset not only demonstrates improved harmlessness but also outperforms models fine-tuned on human-annotated safety preference data, all the while maintaining a competitive edge in downstream tasks. Importantly, our Safer-Instruct framework is versatile and can be applied to generate preference data across various domains, extending its utility beyond safety preferences. It addresses the challenges in preference data acquisition and advances the development of more capable and responsible AI systems. For dataset and code implementation, see https://github.com/uscnlp-lime/safer-instruct/.", }

Reinforcement learning from human feedback (RLHF) is a vital strategy for enhancing model capability in language models. However, annotating preference data for RLHF is a resource-intensive and creativity-demanding process, while existing automatic generation methods face limitations in data diversity and quality. In response, we present Safer-Instruct, a novel pipeline for automatically constructing large-scale preference data. Our approach leverages reversed instruction tuning, instruction induction, and expert model evaluation to efficiently generate high-quality preference data without human annotators. To verify the effectiveness of Safer-Instruct, we apply the pipeline to construct a safety preference dataset as a case study. Finetuning an Alpaca model on this synthetic dataset not only demonstrates improved harmlessness but also outperforms models fine-tuned on human-annotated safety preference data, all the while maintaining a competitive edge in downstream tasks. Importantly, our Safer-Instruct framework is versatile and can be applied to generate preference data across various domains, extending its utility beyond safety preferences. It addresses the challenges in preference data acquisition and advances the development of more capable and responsible AI systems. For dataset and code implementation, see https://github.com/uscnlp-lime/safer-instruct/.

Safer-Instruct: Aligning Language Models with Automated Preference Data

naacl-long.422

Oral

2311.08685

https://huggingface.co/papers/2311.08685

https://aclanthology.org/2024.naacl-long.423.bib

https://aclanthology.org/2024.naacl-long.423/

@inproceedings{peper-etal-2024-pelms, title = "{PELMS}: Pre-training for Effective Low-Shot Multi-Document Summarization", author = "Peper, Joseph and Qiu, Wenzhao and Wang, Lu", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.423", doi = "10.18653/v1/2024.naacl-long.423", pages = "7652--7674", abstract = "We investigate pre-training techniques for abstractive multi-document summarization (MDS), which is much less studied than summarizing single documents. Though recent work has demonstrated the effectiveness of highlighting information salience for pre-training strategy design, they struggle to generate abstractive and reflective summaries, which are critical properties for MDS. To this end, we present **PELMS**, a pre-trained model that uses pre-training objectives based on semantic coherence heuristics and faithfulness constraints together with unlabeled multi-document inputs, to promote the generation of concise, fluent, and faithful summaries. To support the training of PELMS, we compile **MultiPT**, a multi-document pre-training corpus containing over 93 million documents to form more than 3million unlabeled topic-centric document clusters, covering diverse genres such as product reviews, news, and general knowledge. We perform extensive evaluation of PELMS in low-shot settings on a wide range of MDS datasets. Our approach consistently outperforms competitive comparisons with respect to overall informativeness, abstractiveness, coherence, and faithfulness, and with minimal fine-tuning can match performance of language models at a much larger scale (e.g., GPT-4).", }

We investigate pre-training techniques for abstractive multi-document summarization (MDS), which is much less studied than summarizing single documents. Though recent work has demonstrated the effectiveness of highlighting information salience for pre-training strategy design, they struggle to generate abstractive and reflective summaries, which are critical properties for MDS. To this end, we present **PELMS**, a pre-trained model that uses pre-training objectives based on semantic coherence heuristics and faithfulness constraints together with unlabeled multi-document inputs, to promote the generation of concise, fluent, and faithful summaries. To support the training of PELMS, we compile **MultiPT**, a multi-document pre-training corpus containing over 93 million documents to form more than 3million unlabeled topic-centric document clusters, covering diverse genres such as product reviews, news, and general knowledge. We perform extensive evaluation of PELMS in low-shot settings on a wide range of MDS datasets. Our approach consistently outperforms competitive comparisons with respect to overall informativeness, abstractiveness, coherence, and faithfulness, and with minimal fine-tuning can match performance of language models at a much larger scale (e.g., GPT-4).

PELMS: Pre-training for Effective Low-Shot Multi-Document Summarization

naacl-long.423

Poster

2311.09836

https://aclanthology.org/2024.naacl-long.424.bib

https://aclanthology.org/2024.naacl-long.424/

@inproceedings{li-etal-2024-deceptive, title = "Deceptive Semantic Shortcuts on Reasoning Chains: How Far Can Models Go without Hallucination?", author = "Li, Bangzheng and Zhou, Ben and Wang, Fei and Fu, Xingyu and Roth, Dan and Chen, Muhao", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.424", doi = "10.18653/v1/2024.naacl-long.424", pages = "7675--7688", abstract = "Despite the high performances of large language models (LLMs) across numerous benchmarks, recent research has unveiled their suffering from hallucinations and unfaithful reasoning. This work studies a type of hallucination induced by semantic associations. We investigate to what extent LLMs take shortcuts from certain keyword/entity biases in the prompt instead of following correct reasoning paths. To quantify this phenomenon, we propose a novel probing method and benchmark called EUREQA. EUREQA is an entity-searching task where a model finds a missing entity based on described multi-hop relations with other entities. These deliberately designed multi-hop relations create deceptive semantic associations, and models must stick to the correct reasoning path instead of incorrect shortcuts to find the correct answer.Experiments show that existing LLMs cannot follow correct reasoning paths and resist the attempt of greedy shortcuts, with GPT-4 only achieving 62{\%} accuracy. Analyses provide further evidence that LLMs rely on semantic biases to solve the task instead of proper reasoning, questioning the validity and generalizability of current LLMs{'} high performances.", }

Despite the high performances of large language models (LLMs) across numerous benchmarks, recent research has unveiled their suffering from hallucinations and unfaithful reasoning. This work studies a type of hallucination induced by semantic associations. We investigate to what extent LLMs take shortcuts from certain keyword/entity biases in the prompt instead of following correct reasoning paths. To quantify this phenomenon, we propose a novel probing method and benchmark called EUREQA. EUREQA is an entity-searching task where a model finds a missing entity based on described multi-hop relations with other entities. These deliberately designed multi-hop relations create deceptive semantic associations, and models must stick to the correct reasoning path instead of incorrect shortcuts to find the correct answer.Experiments show that existing LLMs cannot follow correct reasoning paths and resist the attempt of greedy shortcuts, with GPT-4 only achieving 62{\%} accuracy. Analyses provide further evidence that LLMs rely on semantic biases to solve the task instead of proper reasoning, questioning the validity and generalizability of current LLMs{'} high performances.

Deceptive Semantic Shortcuts on Reasoning Chains: How Far Can Models Go without Hallucination?

naacl-long.424

Poster

2311.09702

https://aclanthology.org/2024.naacl-long.425.bib

https://aclanthology.org/2024.naacl-long.425/

@inproceedings{kumar-etal-2024-indisentiment140, title = "{I}ndi{S}entiment140: Sentiment Analysis Dataset for {I}ndian Languages with Emphasis on Low-Resource Languages using Machine Translation", author = "Kumar, Saurabh and Sanasam, Ranbir and Nandi, Sukumar", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.425", doi = "10.18653/v1/2024.naacl-long.425", pages = "7689--7698", abstract = "Sentiment analysis, a fundamental aspect of Natural Language Processing (NLP), involves the classification of emotions, opinions, and attitudes in text data. In the context of India, with its vast linguistic diversity and low-resource languages, the challenge is to support sentiment analysis in numerous Indian languages. This study explores the use of machine translation to bridge this gap. The investigation examines the feasibility of machine translation for creating sentiment analysis datasets in 22 Indian languages. Google Translate, with its extensive language support, is employed for this purpose in translating the Sentiment140 dataset. The study aims to provide insights into the practicality of using machine translation in the context of India{'}s linguistic diversity for sentiment analysis datasets. Our findings indicate that a dataset generated using Google Translate has the potential to serve as a foundational framework for tackling the low-resource challenges commonly encountered in sentiment analysis for Indian languages.", }

Sentiment analysis, a fundamental aspect of Natural Language Processing (NLP), involves the classification of emotions, opinions, and attitudes in text data. In the context of India, with its vast linguistic diversity and low-resource languages, the challenge is to support sentiment analysis in numerous Indian languages. This study explores the use of machine translation to bridge this gap. The investigation examines the feasibility of machine translation for creating sentiment analysis datasets in 22 Indian languages. Google Translate, with its extensive language support, is employed for this purpose in translating the Sentiment140 dataset. The study aims to provide insights into the practicality of using machine translation in the context of India{'}s linguistic diversity for sentiment analysis datasets. Our findings indicate that a dataset generated using Google Translate has the potential to serve as a foundational framework for tackling the low-resource challenges commonly encountered in sentiment analysis for Indian languages.

IndiSentiment140: Sentiment Analysis Dataset for Indian Languages with Emphasis on Low-Resource Languages using Machine Translation

naacl-long.425

Poster

https://aclanthology.org/2024.naacl-long.426.bib

https://aclanthology.org/2024.naacl-long.426/

@inproceedings{thakur-etal-2024-leveraging, title = "Leveraging {LLM}s for Synthesizing Training Data Across Many Languages in Multilingual Dense Retrieval", author = "Thakur, Nandan and Ni, Jianmo and Hernandez Abrego, Gustavo and Wieting, John and Lin, Jimmy and Cer, Daniel", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.426", doi = "10.18653/v1/2024.naacl-long.426", pages = "7699--7724", abstract = "There has been limited success for dense retrieval models in multilingual retrieval, due to uneven and scarce training data available across multiple languages. Synthetic training data generation is promising (e.g., InPars or Promptagator), but has been investigated only for English. Therefore, to study model capabilities across both cross-lingual and monolingual retrieval tasks, we develop **SWIM-IR**, a synthetic retrieval training dataset containing 33 (high to very-low resource) languages for fine-tuning multilingual dense retrievers without requiring any human supervision. To construct SWIM-IR, we propose SAP (summarize-then-ask prompting), where the large language model (LLM) generates a textual summary prior to the query generation step. SAP assists the LLM in generating informative queries in the target language. Using SWIM-IR, we explore synthetic fine-tuning of multilingual dense retrieval models and evaluate them robustly on three retrieval benchmarks: XOR-Retrieve (cross-lingual), MIRACL (monolingual) and XTREME-UP (cross-lingual). Our models, called SWIM-X, are competitive with human-supervised dense retrieval models, e.g., mContriever-X, finding that SWIM-IR can cheaply substitute for expensive human-labeled retrieval training data. SWIM-IR dataset and SWIM-X models are available at: https://github.com/google-research-datasets/SWIM-IR.", }

There has been limited success for dense retrieval models in multilingual retrieval, due to uneven and scarce training data available across multiple languages. Synthetic training data generation is promising (e.g., InPars or Promptagator), but has been investigated only for English. Therefore, to study model capabilities across both cross-lingual and monolingual retrieval tasks, we develop **SWIM-IR**, a synthetic retrieval training dataset containing 33 (high to very-low resource) languages for fine-tuning multilingual dense retrievers without requiring any human supervision. To construct SWIM-IR, we propose SAP (summarize-then-ask prompting), where the large language model (LLM) generates a textual summary prior to the query generation step. SAP assists the LLM in generating informative queries in the target language. Using SWIM-IR, we explore synthetic fine-tuning of multilingual dense retrieval models and evaluate them robustly on three retrieval benchmarks: XOR-Retrieve (cross-lingual), MIRACL (monolingual) and XTREME-UP (cross-lingual). Our models, called SWIM-X, are competitive with human-supervised dense retrieval models, e.g., mContriever-X, finding that SWIM-IR can cheaply substitute for expensive human-labeled retrieval training data. SWIM-IR dataset and SWIM-X models are available at: https://github.com/google-research-datasets/SWIM-IR.

Leveraging LLMs for Synthesizing Training Data Across Many Languages in Multilingual Dense Retrieval

naacl-long.426

Poster

2311.05800

https://huggingface.co/papers/2311.05800

https://aclanthology.org/2024.naacl-long.427.bib

https://aclanthology.org/2024.naacl-long.427/

@inproceedings{ok-etal-2024-scanner, title = "{SCANNER}: Knowledge-Enhanced Approach for Robust Multi-modal Named Entity Recognition of Unseen Entities", author = "Ok, Hyunjong and Kil, Taeho and Seo, Sukmin and Lee, Jaeho", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.427", doi = "10.18653/v1/2024.naacl-long.427", pages = "7725--7737", abstract = "Recent advances in named entity recognition (NER) have pushed the boundary of the task to incorporate visual signals, leading to many variants, including multi-modal NER (MNER) or grounded MNER (GMNER). A key challenge to these tasks is that the model should be able to generalize to the entities unseen during the training, and should be able to handle the training samples with noisy annotations.To address this obstacle, we propose SCANNER (Span CANdidate detection and recognition for NER), a model capable of effectively handling all three NER variants.SCANNER is a two-stage structure; we extract entity candidates in the first stage and use it as a query to get knowledge, effectively pulling knowledge from various sources.We can boost our performance by utilizing this entity-centric extracted knowledge to address unseen entities.Furthermore, to tackle the challenges arising from noisy annotations in NER datasets, we introduce a novel self-distillation method, enhancing the robustness and accuracy of our model in processing training data with inherent uncertainties.Our approach demonstrates competitive performance on the NER benchmark and surpasses existing methods on both MNER and GMNER benchmarks.Further analysis shows that the proposed distillation and knowledge utilization methods improve the performance of our model on various benchmarks.", }

Recent advances in named entity recognition (NER) have pushed the boundary of the task to incorporate visual signals, leading to many variants, including multi-modal NER (MNER) or grounded MNER (GMNER). A key challenge to these tasks is that the model should be able to generalize to the entities unseen during the training, and should be able to handle the training samples with noisy annotations.To address this obstacle, we propose SCANNER (Span CANdidate detection and recognition for NER), a model capable of effectively handling all three NER variants.SCANNER is a two-stage structure; we extract entity candidates in the first stage and use it as a query to get knowledge, effectively pulling knowledge from various sources.We can boost our performance by utilizing this entity-centric extracted knowledge to address unseen entities.Furthermore, to tackle the challenges arising from noisy annotations in NER datasets, we introduce a novel self-distillation method, enhancing the robustness and accuracy of our model in processing training data with inherent uncertainties.Our approach demonstrates competitive performance on the NER benchmark and surpasses existing methods on both MNER and GMNER benchmarks.Further analysis shows that the proposed distillation and knowledge utilization methods improve the performance of our model on various benchmarks.

SCANNER: Knowledge-Enhanced Approach for Robust Multi-modal Named Entity Recognition of Unseen Entities

naacl-long.427

Poster

2404.01914

https://aclanthology.org/2024.naacl-long.428.bib

https://aclanthology.org/2024.naacl-long.428/

@inproceedings{tian-etal-2024-theory, title = "A Theory Guided Scaffolding Instruction Framework for {LLM}-Enabled Metaphor Reasoning", author = "Tian, Yuan and Xu, Nan and Mao, Wenji", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.428", doi = "10.18653/v1/2024.naacl-long.428", pages = "7738--7755", abstract = "Metaphor detection is a challenging task in figurative language processing, which aims to distinguish between metaphorical and literal expressions in text. Existing methods tackle metaphor detection via training or fine-tuning discriminative models on labeled data. However, these approaches struggle to explain the underlying reasoning process behind the metaphorical/literal judgment. Recently, large language models (LLMs) have shown promise in language reasoning tasks. Although promising, LLM-based methods for metaphor detection and reasoning are still faced with the challenging issue of bringing the explainable concepts for metaphor reasoning and their linguistic manifestation. To fill this gap, we propose a novel Theory guided Scaffolding Instruction (TSI) framework that instructs an LLM to infer the underlying reasoning process of metaphor detection guided by metaphor theories for the first time. Our work is inspired by a pedagogical strategy called scaffolding instruction, which encourages educators to provide questioning and support as scaffolding so as to assist learners in constructing the understanding of pedagogical goals step by step. We first construct a metaphor knowledge graph grounded in metaphor theory which serves as the instructional structure to obtain a series of scaffolding questions, directing the LLM to incrementally generate the reasoning process for metaphor understanding through dialogue interactions. During this theory guided instruction process, we explore the LLM{'}s mastery boundary and provide the relevant knowledge as scaffolding support when the question is beyond the LLM{'}s capability. Experimental results verify that our method significantly outperforms both the LLM-based reasoning methods and the SOTA methods in metaphor detection, indicating the facilitation of metaphor and instruction theories in guiding LLM-based reasoning process.", }

Metaphor detection is a challenging task in figurative language processing, which aims to distinguish between metaphorical and literal expressions in text. Existing methods tackle metaphor detection via training or fine-tuning discriminative models on labeled data. However, these approaches struggle to explain the underlying reasoning process behind the metaphorical/literal judgment. Recently, large language models (LLMs) have shown promise in language reasoning tasks. Although promising, LLM-based methods for metaphor detection and reasoning are still faced with the challenging issue of bringing the explainable concepts for metaphor reasoning and their linguistic manifestation. To fill this gap, we propose a novel Theory guided Scaffolding Instruction (TSI) framework that instructs an LLM to infer the underlying reasoning process of metaphor detection guided by metaphor theories for the first time. Our work is inspired by a pedagogical strategy called scaffolding instruction, which encourages educators to provide questioning and support as scaffolding so as to assist learners in constructing the understanding of pedagogical goals step by step. We first construct a metaphor knowledge graph grounded in metaphor theory which serves as the instructional structure to obtain a series of scaffolding questions, directing the LLM to incrementally generate the reasoning process for metaphor understanding through dialogue interactions. During this theory guided instruction process, we explore the LLM{'}s mastery boundary and provide the relevant knowledge as scaffolding support when the question is beyond the LLM{'}s capability. Experimental results verify that our method significantly outperforms both the LLM-based reasoning methods and the SOTA methods in metaphor detection, indicating the facilitation of metaphor and instruction theories in guiding LLM-based reasoning process.

A Theory Guided Scaffolding Instruction Framework for LLM-Enabled Metaphor Reasoning

naacl-long.428

Poster

https://aclanthology.org/2024.naacl-long.429.bib

https://aclanthology.org/2024.naacl-long.429/

@inproceedings{chuang-etal-2024-learning, title = "Learning to Compress Prompt in Natural Language Formats", author = "Chuang, Yu-Neng and Xing, Tianwei and Chang, Chia-Yuan and Liu, Zirui and Chen, Xun and Hu, Xia", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.429", doi = "10.18653/v1/2024.naacl-long.429", pages = "7756--7767", abstract = "Large language models (LLMs) are great at processing multiple natural language processing tasks, but their abilities are constrained by inferior performance with long context, slow inference speed, and the high cost of computing the results. Deploying LLMs with precise and informative context helps users process large-scale datasets more effectively and cost-efficiently. Existing works rely on compressing long prompt contexts into soft prompts. However, soft prompt compression encounters limitations in transferability across different LLMs, especially API-based LLMs. To this end, this work aims to compress lengthy prompts in the form of natural language with LLM transferability. This poses two challenges: (i) Natural Language (NL) prompts are incompatible with back-propagation, and (ii) NL prompts lack flexibility in imposing length constraints. In this work, we propose a Natural Language Prompt Encapsulation (Nano-Capsulator) framework compressing original prompts into NL formatted Capsule Prompt while maintaining prompt utility and transferability. Specifically, to tackle the first challenge, the Nano-Capsulator is optimized by a reward function that interacts with the proposed semantics preserving loss. To address the second question, the Nano-Capsulator is optimized by a reward function featuring length constraints. Experimental results demonstrate that the Capsule Prompt can reduce 81.4{\%} of the original length, decrease inference latency up to 4.5x, and save 80.1{\%} of budget overheads while providing transferability across diverse LLMs and different datasets.", }

Large language models (LLMs) are great at processing multiple natural language processing tasks, but their abilities are constrained by inferior performance with long context, slow inference speed, and the high cost of computing the results. Deploying LLMs with precise and informative context helps users process large-scale datasets more effectively and cost-efficiently. Existing works rely on compressing long prompt contexts into soft prompts. However, soft prompt compression encounters limitations in transferability across different LLMs, especially API-based LLMs. To this end, this work aims to compress lengthy prompts in the form of natural language with LLM transferability. This poses two challenges: (i) Natural Language (NL) prompts are incompatible with back-propagation, and (ii) NL prompts lack flexibility in imposing length constraints. In this work, we propose a Natural Language Prompt Encapsulation (Nano-Capsulator) framework compressing original prompts into NL formatted Capsule Prompt while maintaining prompt utility and transferability. Specifically, to tackle the first challenge, the Nano-Capsulator is optimized by a reward function that interacts with the proposed semantics preserving loss. To address the second question, the Nano-Capsulator is optimized by a reward function featuring length constraints. Experimental results demonstrate that the Capsule Prompt can reduce 81.4{\%} of the original length, decrease inference latency up to 4.5x, and save 80.1{\%} of budget overheads while providing transferability across diverse LLMs and different datasets.

Learning to Compress Prompt in Natural Language Formats

naacl-long.429

Poster

2402.18700

https://huggingface.co/papers/2402.18700

https://aclanthology.org/2024.naacl-long.430.bib

https://aclanthology.org/2024.naacl-long.430/

@inproceedings{zhuang-etal-2024-automatic, title = "Automatic, Meta and Human Evaluation for Multimodal Summarization with Multimodal Output", author = "Zhuang, Haojie and Zhang, Wei Emma and Xie, Leon and Chen, Weitong and Yang, Jian and Sheng, Quan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.430", doi = "10.18653/v1/2024.naacl-long.430", pages = "7768--7790", abstract = "Multimodal summarization with multimodal output (MSMO) has attracted increasing research interests recently as multimodal summary could provide more comprehensive information compared to text-only summary, effectively improving the user experience and satisfaction. As one of the most fundamental components for the development of MSMO, evaluation is an emerging yet underexplored research topic. In this paper, we fill this gap and propose a research framework that studies three research questions of MSMO evaluation: (1) Automatic Evaluation: We propose a novel metric mLLM-EVAL, which utilizes multimodal Large Language Model for MSMO EVALuation. (2) Meta-Evaluation: We create a meta-evaluation benchmark dataset by collecting human-annotated scores for multimodal summaries. With our benchmark, we conduct meta-evaluation analysis to assess the quality of different evaluation metrics and show the effectiveness of our proposed mLLM-EVAL. (3) Human Evaluation: To provide more objective and unbiased human annotations for meta-evaluation, we hypothesize and verify three types of cognitive biases in human evaluation. We also incorporate our findings into the human annotation process in the meta-evaluation benchmark. Overall, our research framework provides an evaluation metric, a meta-evaluation benchmark dataset annotated by humans and an analysis of cognitive biases in human evaluation, which we believe would serve as a valuable and comprehensive resource for the MSMO research community.", }

Multimodal summarization with multimodal output (MSMO) has attracted increasing research interests recently as multimodal summary could provide more comprehensive information compared to text-only summary, effectively improving the user experience and satisfaction. As one of the most fundamental components for the development of MSMO, evaluation is an emerging yet underexplored research topic. In this paper, we fill this gap and propose a research framework that studies three research questions of MSMO evaluation: (1) Automatic Evaluation: We propose a novel metric mLLM-EVAL, which utilizes multimodal Large Language Model for MSMO EVALuation. (2) Meta-Evaluation: We create a meta-evaluation benchmark dataset by collecting human-annotated scores for multimodal summaries. With our benchmark, we conduct meta-evaluation analysis to assess the quality of different evaluation metrics and show the effectiveness of our proposed mLLM-EVAL. (3) Human Evaluation: To provide more objective and unbiased human annotations for meta-evaluation, we hypothesize and verify three types of cognitive biases in human evaluation. We also incorporate our findings into the human annotation process in the meta-evaluation benchmark. Overall, our research framework provides an evaluation metric, a meta-evaluation benchmark dataset annotated by humans and an analysis of cognitive biases in human evaluation, which we believe would serve as a valuable and comprehensive resource for the MSMO research community.

Automatic, Meta and Human Evaluation for Multimodal Summarization with Multimodal Output

naacl-long.430

Poster

https://aclanthology.org/2024.naacl-long.431.bib

https://aclanthology.org/2024.naacl-long.431/

@inproceedings{su-etal-2024-naive, title = "Naive {B}ayes-based Context Extension for Large Language Models", author = "Su, Jianlin and Ahmed, Murtadha and Wen, Bo and Ao, Luo and Zhu, Mingren and Liu, Yunfeng", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.431", doi = "10.18653/v1/2024.naacl-long.431", pages = "7791--7807", abstract = "Large Language Models (LLMs) have shown promising in-context learning abilities. However, conventional In-Context Learning (ICL) approaches are often impeded by length limitations of transformer architecture, which pose challenges when attempting to effectively integrate supervision from a substantial number of demonstration examples. In this paper, we introduce a novel framework, called Naive Bayes-based Context Extension (NBCE), to enable existing LLMs to perform ICL with an increased number of demonstrations by significantly expanding their context size. Importantly, this expansion does not require fine-tuning or dependence on particular model architectures, all the while preserving linear efficiency. NBCE initially splits the context into equal-sized windows fitting the target LLM{'}s maximum length. Then, it introduces a voting mechanism to select the most relevant window, regarded as the posterior context. Finally, it employs Bayes{'} theorem to generate the test task. Our experimental results demonstrate that NBCE substantially enhances performance, particularly as the number of demonstration examples increases, consistently outperforming alternative methods. The NBCE code will be made publicly accessible. The code NBCE is available at: https://github.com/amurtadha/NBCE-master", }

Large Language Models (LLMs) have shown promising in-context learning abilities. However, conventional In-Context Learning (ICL) approaches are often impeded by length limitations of transformer architecture, which pose challenges when attempting to effectively integrate supervision from a substantial number of demonstration examples. In this paper, we introduce a novel framework, called Naive Bayes-based Context Extension (NBCE), to enable existing LLMs to perform ICL with an increased number of demonstrations by significantly expanding their context size. Importantly, this expansion does not require fine-tuning or dependence on particular model architectures, all the while preserving linear efficiency. NBCE initially splits the context into equal-sized windows fitting the target LLM{'}s maximum length. Then, it introduces a voting mechanism to select the most relevant window, regarded as the posterior context. Finally, it employs Bayes{'} theorem to generate the test task. Our experimental results demonstrate that NBCE substantially enhances performance, particularly as the number of demonstration examples increases, consistently outperforming alternative methods. The NBCE code will be made publicly accessible. The code NBCE is available at: https://github.com/amurtadha/NBCE-master

Naive Bayes-based Context Extension for Large Language Models

naacl-long.431

Oral

2403.17552

https://aclanthology.org/2024.naacl-long.432.bib

https://aclanthology.org/2024.naacl-long.432/

@inproceedings{mhamdi-may-2024-leitner, title = "Leitner-Guided Memory Replay for Cross-lingual Continual Learning", author = "M{'}hamdi, Meryem and May, Jonathan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.432", doi = "10.18653/v1/2024.naacl-long.432", pages = "7808--7821", abstract = "Cross-lingual continual learning aims to continuously fine-tune a downstream model on emerging data from new languages. One major challenge in cross-lingual continual learning is catastrophic forgetting: a stability-plasticity dilemma, where performance on previously seen languages decreases as the model learns to transfer to new languages. Experience replay, which revisits data from a fixed-size memory of old languages while training on new ones, is among the most successful approaches for solving this dilemma. Faced with the challenge of dynamically storing the memory with high-quality examples while complying with its fixed size limitations, we consider Leitner queuing, a human-inspired spaced-repetition technique, to determine what should be replayed at each phase of learning. Via a controlled set of quantitative and qualitative analyses across different memory strategies, we show that, just like humans, carefully picking informative examples to be prioritized in cross-lingual memory replay helps tame the stability-plasticity dilemma. Compared to vanilla and strong memory replay baselines, our Leitner-guided approach significantly and consistently decreases forgetting while maintaining accuracy across natural language understanding tasks, language orders, and languages.", }

Cross-lingual continual learning aims to continuously fine-tune a downstream model on emerging data from new languages. One major challenge in cross-lingual continual learning is catastrophic forgetting: a stability-plasticity dilemma, where performance on previously seen languages decreases as the model learns to transfer to new languages. Experience replay, which revisits data from a fixed-size memory of old languages while training on new ones, is among the most successful approaches for solving this dilemma. Faced with the challenge of dynamically storing the memory with high-quality examples while complying with its fixed size limitations, we consider Leitner queuing, a human-inspired spaced-repetition technique, to determine what should be replayed at each phase of learning. Via a controlled set of quantitative and qualitative analyses across different memory strategies, we show that, just like humans, carefully picking informative examples to be prioritized in cross-lingual memory replay helps tame the stability-plasticity dilemma. Compared to vanilla and strong memory replay baselines, our Leitner-guided approach significantly and consistently decreases forgetting while maintaining accuracy across natural language understanding tasks, language orders, and languages.

Leitner-Guided Memory Replay for Cross-lingual Continual Learning

naacl-long.432

Oral

https://aclanthology.org/2024.naacl-long.433.bib

https://aclanthology.org/2024.naacl-long.433/

@inproceedings{arps-etal-2024-multilingual, title = "Multilingual Nonce Dependency Treebanks: Understanding how Language Models Represent and Process Syntactic Structure", author = "Arps, David and Kallmeyer, Laura and Samih, Younes and Sajjad, Hassan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.433", doi = "10.18653/v1/2024.naacl-long.433", pages = "7822--7844", abstract = {We introduce SPUD (Semantically Perturbed Universal Dependencies), a framework for creating nonce treebanks for the multilingual Universal Dependencies (UD) corpora. SPUD data satisfies syntactic argument structure, provides syntactic annotations, and ensures grammaticality via language-specific rules. We create nonce data in Arabic, English, French, German, and Russian, and demonstrate two use cases of SPUD treebanks. First, we investigate the effect of nonce data on word co-occurrence statistics, as measured by perplexity scores of autoregressive (ALM) and masked language models (MLM). We find that ALM scores are significantly more affected by nonce data than MLM scores. Second, we show how nonce data affects the performance of syntactic dependency probes. We replicate the findings of M{\"u}ller-Eberstein et al. (2022) on nonce test data and show that the performance declines on both MLMs and ALMs wrt. original test data. However, a majority of the performance is kept, suggesting that the probe indeed learns syntax independently from semantics.}, }

We introduce SPUD (Semantically Perturbed Universal Dependencies), a framework for creating nonce treebanks for the multilingual Universal Dependencies (UD) corpora. SPUD data satisfies syntactic argument structure, provides syntactic annotations, and ensures grammaticality via language-specific rules. We create nonce data in Arabic, English, French, German, and Russian, and demonstrate two use cases of SPUD treebanks. First, we investigate the effect of nonce data on word co-occurrence statistics, as measured by perplexity scores of autoregressive (ALM) and masked language models (MLM). We find that ALM scores are significantly more affected by nonce data than MLM scores. Second, we show how nonce data affects the performance of syntactic dependency probes. We replicate the findings of M{\"u}ller-Eberstein et al. (2022) on nonce test data and show that the performance declines on both MLMs and ALMs wrt. original test data. However, a majority of the performance is kept, suggesting that the probe indeed learns syntax independently from semantics.

Multilingual Nonce Dependency Treebanks: Understanding how Language Models Represent and Process Syntactic Structure

naacl-long.433

Oral

2311.07497

https://aclanthology.org/2024.naacl-long.434.bib

https://aclanthology.org/2024.naacl-long.434/

@inproceedings{liang-etal-2024-actively, title = "Actively Learn from {LLM}s with Uncertainty Propagation for Generalized Category Discovery", author = "Liang, Jinggui and Liao, Lizi and Fei, Hao and Li, Bobo and Jiang, Jing", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.434", doi = "10.18653/v1/2024.naacl-long.434", pages = "7845--7858", abstract = "Generalized category discovery faces a key issue: the lack of supervision for new and unseen data categories. Traditional methods typically combine supervised pretraining with self-supervised learning to create models, and then employ clustering for category identification. However, these approaches tend to become overly tailored to known categories, failing to fully resolve the core issue. Hence, we propose to integrate the feedback from LLMs into an active learning paradigm. Specifically, our method innovatively employs uncertainty propagation to select data samples from high-uncertainty regions, which are then labeled using LLMs through a comparison-based prompting scheme. This not only eases the labeling task but also enhances accuracy in identifying new categories. Additionally, a soft feedback propagation mechanism is introduced to minimize the spread of inaccurate feedback. Experiments on various datasets demonstrate our framework{'}s efficacy and generalizability, significantly improving baseline models at a nominal average cost.", }

Generalized category discovery faces a key issue: the lack of supervision for new and unseen data categories. Traditional methods typically combine supervised pretraining with self-supervised learning to create models, and then employ clustering for category identification. However, these approaches tend to become overly tailored to known categories, failing to fully resolve the core issue. Hence, we propose to integrate the feedback from LLMs into an active learning paradigm. Specifically, our method innovatively employs uncertainty propagation to select data samples from high-uncertainty regions, which are then labeled using LLMs through a comparison-based prompting scheme. This not only eases the labeling task but also enhances accuracy in identifying new categories. Additionally, a soft feedback propagation mechanism is introduced to minimize the spread of inaccurate feedback. Experiments on various datasets demonstrate our framework{'}s efficacy and generalizability, significantly improving baseline models at a nominal average cost.

Actively Learn from LLMs with Uncertainty Propagation for Generalized Category Discovery

naacl-long.434

Poster

https://aclanthology.org/2024.naacl-long.435.bib

https://aclanthology.org/2024.naacl-long.435/

@inproceedings{vasileiou-eberle-2024-explaining, title = "Explaining Text Similarity in Transformer Models", author = "Vasileiou, Alexandros and Eberle, Oliver", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.435", doi = "10.18653/v1/2024.naacl-long.435", pages = "7859--7873", abstract = "As Transformers have become state-of-the-art models for natural language processing (NLP) tasks, the need to understand and explain their predictions is increasingly apparent. Especially in unsupervised applications, such as information retrieval tasks, similarity models built on top of foundation model representations have been widely applied. However, their inner prediction mechanisms have mostly remained opaque. Recent advances in explainable AI have made it possible to mitigate these limitations by leveraging improved explanations for Transformers through layer-wise relevance propagation (LRP). Using BiLRP, an extension developed for computing second-order explanations in bilinear similarity models, we investigate which feature interactions drive similarity in NLP models. We validate the resulting explanations and demonstrate their utility in three corpus-level use cases, analyzing grammatical interactions, multilingual semantics, and biomedical text retrieval. Our findings contribute to a deeper understanding of different semantic similarity tasks and models, highlighting how novel explainable AI methods enable in-depth analyses and corpus-level insights.", }

As Transformers have become state-of-the-art models for natural language processing (NLP) tasks, the need to understand and explain their predictions is increasingly apparent. Especially in unsupervised applications, such as information retrieval tasks, similarity models built on top of foundation model representations have been widely applied. However, their inner prediction mechanisms have mostly remained opaque. Recent advances in explainable AI have made it possible to mitigate these limitations by leveraging improved explanations for Transformers through layer-wise relevance propagation (LRP). Using BiLRP, an extension developed for computing second-order explanations in bilinear similarity models, we investigate which feature interactions drive similarity in NLP models. We validate the resulting explanations and demonstrate their utility in three corpus-level use cases, analyzing grammatical interactions, multilingual semantics, and biomedical text retrieval. Our findings contribute to a deeper understanding of different semantic similarity tasks and models, highlighting how novel explainable AI methods enable in-depth analyses and corpus-level insights.

Explaining Text Similarity in Transformer Models

naacl-long.435

Poster

2405.06604

https://huggingface.co/papers/2405.06604

https://aclanthology.org/2024.naacl-long.436.bib

https://aclanthology.org/2024.naacl-long.436/

@inproceedings{wang-etal-2024-large-language, title = "Large Language Models can Contrastively Refine their Generation for Better Sentence Representation Learning", author = "Wang, Huiming and Li, Zhaodonghui and Cheng, Liying and Soh, De Wen and Bing, Lidong", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.436", doi = "10.18653/v1/2024.naacl-long.436", pages = "7874--7891", abstract = "Recently, large language models (LLMs) have emerged as a groundbreaking technology and their unparalleled text generation capabilities have sparked interest in their application to the fundamental sentence representation learning task. Existing methods have explored utilizing LLMs as data annotators to generate synthesized data for training contrastive learning based sentence embedding models such as SimCSE. However, since contrastive learning models are sensitive to the quality of sentence pairs, the effectiveness of these methods is largely influenced by the content generated from LLMs, highlighting the need for more refined generation in the context of sentence representation learning. Building upon this premise, we propose MultiCSR, a multi-level contrastive sentence representation learning framework that decomposes the process of prompting LLMs to generate a corpus for training base sentence embedding models into three stages (i.e., sentence generation, sentence pair construction, in-batch training) and refines the generated content at these three distinct stages, ensuring only high-quality sentence pairs are utilized to train a base contrastive learning model. Our extensive experiments reveal that MultiCSR enables a less advanced LLM to surpass the performance of ChatGPT, while applying it to ChatGPT achieves better state-of-the-art results. Comprehensive analyses further underscore the potential of our framework in various application scenarios and achieving better sentence representation learning with LLMs.", }

Recently, large language models (LLMs) have emerged as a groundbreaking technology and their unparalleled text generation capabilities have sparked interest in their application to the fundamental sentence representation learning task. Existing methods have explored utilizing LLMs as data annotators to generate synthesized data for training contrastive learning based sentence embedding models such as SimCSE. However, since contrastive learning models are sensitive to the quality of sentence pairs, the effectiveness of these methods is largely influenced by the content generated from LLMs, highlighting the need for more refined generation in the context of sentence representation learning. Building upon this premise, we propose MultiCSR, a multi-level contrastive sentence representation learning framework that decomposes the process of prompting LLMs to generate a corpus for training base sentence embedding models into three stages (i.e., sentence generation, sentence pair construction, in-batch training) and refines the generated content at these three distinct stages, ensuring only high-quality sentence pairs are utilized to train a base contrastive learning model. Our extensive experiments reveal that MultiCSR enables a less advanced LLM to surpass the performance of ChatGPT, while applying it to ChatGPT achieves better state-of-the-art results. Comprehensive analyses further underscore the potential of our framework in various application scenarios and achieving better sentence representation learning with LLMs.

Large Language Models can Contrastively Refine their Generation for Better Sentence Representation Learning

naacl-long.436

Poster

2310.10962

https://aclanthology.org/2024.naacl-long.437.bib

https://aclanthology.org/2024.naacl-long.437/

@inproceedings{kim-etal-2024-hil, title = "{HIL}: Hybrid Isotropy Learning for Zero-shot Performance in Dense retrieval", author = "Kim, Jaeyoung and Lee, Dohyeon and Hwang, Seung-won", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.437", doi = "10.18653/v1/2024.naacl-long.437", pages = "7892--7903", abstract = "Advancements in dense retrieval models have brought ColBERT to prominence in Information Retrieval (IR) with its advanced interaction techniques.However, ColBERT is reported to frequently underperform in zero-shot scenarios, where traditional techniques such as BM25 still exceed it.Addressing this, we propose to balance representation isotropy and anisotropy for zero-shot model performance, based on our observations that isotropy can enhance cosine similarity computations and anisotropy may aid in generalizing to unseen data.Striking a balance between these isotropic and anisotropic qualities stands as a critical objective to refine model efficacy.Based on this, we present ours, a Hybrid Isotropy Learning (HIL) architecture that integrates isotropic and anisotropic representations.Our experiments with the BEIR benchmark show that our model significantly outperforms the baseline ColBERT model, highlighting the importance of harmonized isotropy in improving zero-shot retrieval performance.", }

Advancements in dense retrieval models have brought ColBERT to prominence in Information Retrieval (IR) with its advanced interaction techniques.However, ColBERT is reported to frequently underperform in zero-shot scenarios, where traditional techniques such as BM25 still exceed it.Addressing this, we propose to balance representation isotropy and anisotropy for zero-shot model performance, based on our observations that isotropy can enhance cosine similarity computations and anisotropy may aid in generalizing to unseen data.Striking a balance between these isotropic and anisotropic qualities stands as a critical objective to refine model efficacy.Based on this, we present ours, a Hybrid Isotropy Learning (HIL) architecture that integrates isotropic and anisotropic representations.Our experiments with the BEIR benchmark show that our model significantly outperforms the baseline ColBERT model, highlighting the importance of harmonized isotropy in improving zero-shot retrieval performance.

HIL: Hybrid Isotropy Learning for Zero-shot Performance in Dense retrieval

naacl-long.437

Poster

https://aclanthology.org/2024.naacl-long.438.bib

https://aclanthology.org/2024.naacl-long.438/

@inproceedings{pfister-hotho-2024-supergleber, title = "{S}uper{GLEB}er: {G}erman Language Understanding Evaluation Benchmark", author = "Pfister, Jan and Hotho, Andreas", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.438", doi = "10.18653/v1/2024.naacl-long.438", pages = "7904--7923", abstract = "We assemble a broad Natural Language Understanding benchmark suite for the German language and consequently evaluate a wide array of existing German-capable models in order to create a better understanding of the current state of German LLMs. Our benchmark consists of 29 different tasks ranging over different types such as document classification, sequence tagging, sentence similarity, and question answering, on which we evaluate 10 different German-pretrained models, thereby charting the landscape of German LLMs. In our comprehensive evaluation we find that encoder models are a good choice for most tasks, but also that the largest encoder model does not necessarily perform best for all tasks. We make our benchmark suite and a leaderboard publically available at https://supergleber.professor-x.de and encourage the community to contribute new tasks and evaluate more models on it (https://github.com/LSX-UniWue/SuperGLEBer).", }

We assemble a broad Natural Language Understanding benchmark suite for the German language and consequently evaluate a wide array of existing German-capable models in order to create a better understanding of the current state of German LLMs. Our benchmark consists of 29 different tasks ranging over different types such as document classification, sequence tagging, sentence similarity, and question answering, on which we evaluate 10 different German-pretrained models, thereby charting the landscape of German LLMs. In our comprehensive evaluation we find that encoder models are a good choice for most tasks, but also that the largest encoder model does not necessarily perform best for all tasks. We make our benchmark suite and a leaderboard publically available at https://supergleber.professor-x.de and encourage the community to contribute new tasks and evaluate more models on it (https://github.com/LSX-UniWue/SuperGLEBer).

SuperGLEBer: German Language Understanding Evaluation Benchmark

naacl-long.438

Poster

https://aclanthology.org/2024.naacl-long.439.bib

https://aclanthology.org/2024.naacl-long.439/

@inproceedings{bagdon-etal-2024-expert, title = "{``}You are an expert annotator{''}: Automatic Best{--}Worst-Scaling Annotations for Emotion Intensity Modeling", author = "Bagdon, Christopher and Karmalkar, Prathamesh and Gurulingappa, Harsha and Klinger, Roman", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.439", doi = "10.18653/v1/2024.naacl-long.439", pages = "7924--7936", abstract = "Labeling corpora constitutes a bottleneck to create models for new tasks or domains. Large language models mitigate the issue with automatic corpus labeling methods, particularly for categorical annotations. Some NLP tasks such as emotion intensity prediction, however, require text regression, but there is no work on automating annotations for continuous label assignments. Regression is considered more challenging than classification: The fact that humans perform worse when tasked to choose values from a rating scale lead to comparative annotation methods, including best{--}worst scaling. This raises the question if large language model-based annotation methods show similar patterns, namely that they perform worse on rating scale annotation tasks than on comparative annotation tasks. To study this, we automate emotion intensity predictions and compare direct rating scale predictions, pairwise comparisons and best{--}worst scaling. We find that the latter shows the highest reliability. A transformer regressor fine-tuned on these data performs nearly on par with a model trained on the original manual annotations.", }

Labeling corpora constitutes a bottleneck to create models for new tasks or domains. Large language models mitigate the issue with automatic corpus labeling methods, particularly for categorical annotations. Some NLP tasks such as emotion intensity prediction, however, require text regression, but there is no work on automating annotations for continuous label assignments. Regression is considered more challenging than classification: The fact that humans perform worse when tasked to choose values from a rating scale lead to comparative annotation methods, including best{--}worst scaling. This raises the question if large language model-based annotation methods show similar patterns, namely that they perform worse on rating scale annotation tasks than on comparative annotation tasks. To study this, we automate emotion intensity predictions and compare direct rating scale predictions, pairwise comparisons and best{--}worst scaling. We find that the latter shows the highest reliability. A transformer regressor fine-tuned on these data performs nearly on par with a model trained on the original manual annotations.

“You are an expert annotator”: Automatic Best–Worst-Scaling Annotations for Emotion Intensity Modeling

naacl-long.439

Poster

https://aclanthology.org/2024.naacl-long.440.bib

https://aclanthology.org/2024.naacl-long.440/

@inproceedings{zeng-etal-2024-matters, title = "What Matters in Training a {GPT}4-Style Language Model with Multimodal Inputs?", author = "Zeng, Yan and Zhang, Hanbo and Zheng, Jiani and Xia, Jiangnan and Wei, Guoqiang and Wei, Yang and Zhang, Yuchen and Kong, Tao and Song, Ruihua", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.440", doi = "10.18653/v1/2024.naacl-long.440", pages = "7937--7964", abstract = "Recent advancements in GPT-4V have displayed remarkable multi-modal capabilities in processing image inputs and following open-ended instructions. Despite these advancements, there is considerable scope for enhancing open-source multi-modal LLMs, especially in terms of multi-modal understanding accuracy and instruction-following proficiency. In this paper, we conduct a comprehensive study on training GPT4-style models. We introduce Lynx a multi-modal LLM developed through a series of controlled experiments comparing various model variants. This process allowed us to identify and implement an optimal training strategy tailored for multi-modal LLMs. In addition to our model development, we propose a plug-and-play technique designed to augment the instruction-following capabilities of multi-modal LLMs. We have validated the performance of Lynx on multiple benchmarks. Results demonstrate that Lynx not only achieves strong image understanding accuracy but also excels in instruction-following tasks, paving the path for ongoing enhancements in multi-modal LLMs.", }

Recent advancements in GPT-4V have displayed remarkable multi-modal capabilities in processing image inputs and following open-ended instructions. Despite these advancements, there is considerable scope for enhancing open-source multi-modal LLMs, especially in terms of multi-modal understanding accuracy and instruction-following proficiency. In this paper, we conduct a comprehensive study on training GPT4-style models. We introduce Lynx a multi-modal LLM developed through a series of controlled experiments comparing various model variants. This process allowed us to identify and implement an optimal training strategy tailored for multi-modal LLMs. In addition to our model development, we propose a plug-and-play technique designed to augment the instruction-following capabilities of multi-modal LLMs. We have validated the performance of Lynx on multiple benchmarks. Results demonstrate that Lynx not only achieves strong image understanding accuracy but also excels in instruction-following tasks, paving the path for ongoing enhancements in multi-modal LLMs.

What Matters in Training a GPT4-Style Language Model with Multimodal Inputs?

naacl-long.440

Oral

2307.02469

https://aclanthology.org/2024.naacl-long.441.bib

https://aclanthology.org/2024.naacl-long.441/

@inproceedings{ruan-etal-2024-defining, title = "Defining and Detecting Vulnerability in Human Evaluation Guidelines: A Preliminary Study Towards Reliable {NLG} Evaluation", author = "Ruan, Jie and Wang, Wenqing and Wan, Xiaojun", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.441", doi = "10.18653/v1/2024.naacl-long.441", pages = "7965--7989", abstract = "Human evaluation serves as the gold standard for assessing the quality of Natural Language Generation (NLG) systems. Nevertheless, the evaluation guideline, as a pivotal element ensuring reliable and reproducible human assessment, has received limited attention. Our investigation revealed that only 29.84{\%} of recent papers involving human evaluation at top conferences release their evaluation guidelines, with vulnerabilities identified in 77.09{\%} of these guidelines. Unreliable evaluation guidelines can yield inaccurate assessment outcomes, potentially impeding the advancement of NLG in the right direction. To address these challenges, we take an initial step towards reliable evaluation guidelines and propose the first human evaluation guideline dataset by collecting annotations of guidelines extracted from existing papers as well as generated via Large Language Models (LLMs). We then introduce a taxonomy of eight vulnerabilities and formulate a principle for composing evaluation guidelines. Furthermore, a method for detecting guideline vulnerabilities has been explored using LLMs, and we offer a set of recommendations to enhance reliability in human evaluation. The annotated human evaluation guideline dataset and code for the vulnerability detection method are publicly available online.", }

Human evaluation serves as the gold standard for assessing the quality of Natural Language Generation (NLG) systems. Nevertheless, the evaluation guideline, as a pivotal element ensuring reliable and reproducible human assessment, has received limited attention. Our investigation revealed that only 29.84{\%} of recent papers involving human evaluation at top conferences release their evaluation guidelines, with vulnerabilities identified in 77.09{\%} of these guidelines. Unreliable evaluation guidelines can yield inaccurate assessment outcomes, potentially impeding the advancement of NLG in the right direction. To address these challenges, we take an initial step towards reliable evaluation guidelines and propose the first human evaluation guideline dataset by collecting annotations of guidelines extracted from existing papers as well as generated via Large Language Models (LLMs). We then introduce a taxonomy of eight vulnerabilities and formulate a principle for composing evaluation guidelines. Furthermore, a method for detecting guideline vulnerabilities has been explored using LLMs, and we offer a set of recommendations to enhance reliability in human evaluation. The annotated human evaluation guideline dataset and code for the vulnerability detection method are publicly available online.

Defining and Detecting Vulnerability in Human Evaluation Guidelines: A Preliminary Study Towards Reliable NLG Evaluation

naacl-long.441

Poster

2406.07935

https://aclanthology.org/2024.naacl-long.442.bib

https://aclanthology.org/2024.naacl-long.442/

@inproceedings{conia-etal-2024-mosaico, title = "{MOSAIC}o: a Multilingual Open-text Semantically Annotated Interlinked Corpus", author = "Conia, Simone and Barba, Edoardo and Martinez Lorenzo, Abelardo Carlos and Huguet Cabot, Pere-Llu{\'\i}s and Orlando, Riccardo and Procopio, Luigi and Navigli, Roberto", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.442", doi = "10.18653/v1/2024.naacl-long.442", pages = "7990--8004", abstract = "Several Natural Language Understanding (NLU) tasks focus on linking text to explicit knowledge, including Word Sense Disambiguation, Semantic Role Labeling, Semantic Parsing, and Relation Extraction. In addition to the importance of connecting raw text with explicit knowledge bases, the integration of such carefully curated knowledge into deep learning models has been shown to be beneficial across a diverse range of applications, including Language Modeling and Machine Translation. Nevertheless, the scarcity of semantically-annotated corpora across various tasks and languages limits the potential advantages significantly. To address this issue, we put forward MOSAICo, the first endeavor aimed at equipping the research community with the key ingredients to model explicit semantic knowledge at a large scale, providing hundreds of millions of silver yet high-quality annotations for four NLU tasks across five languages. We describe the creation process of MOSAICo, demonstrate its quality and variety, and analyze the interplay between different types of semantic information. MOSAICo, available at https://github.com/SapienzaNLP/mosaico, aims to drop the requirement of closed, licensed datasets and represents a step towards a level playing field across languages and tasks in NLU.", }

Several Natural Language Understanding (NLU) tasks focus on linking text to explicit knowledge, including Word Sense Disambiguation, Semantic Role Labeling, Semantic Parsing, and Relation Extraction. In addition to the importance of connecting raw text with explicit knowledge bases, the integration of such carefully curated knowledge into deep learning models has been shown to be beneficial across a diverse range of applications, including Language Modeling and Machine Translation. Nevertheless, the scarcity of semantically-annotated corpora across various tasks and languages limits the potential advantages significantly. To address this issue, we put forward MOSAICo, the first endeavor aimed at equipping the research community with the key ingredients to model explicit semantic knowledge at a large scale, providing hundreds of millions of silver yet high-quality annotations for four NLU tasks across five languages. We describe the creation process of MOSAICo, demonstrate its quality and variety, and analyze the interplay between different types of semantic information. MOSAICo, available at https://github.com/SapienzaNLP/mosaico, aims to drop the requirement of closed, licensed datasets and represents a step towards a level playing field across languages and tasks in NLU.

MOSAICo: a Multilingual Open-text Semantically Annotated Interlinked Corpus

naacl-long.442

Oral

https://aclanthology.org/2024.naacl-long.443.bib

https://aclanthology.org/2024.naacl-long.443/

@inproceedings{formento-etal-2024-semrode, title = "{S}em{R}o{D}e: Macro Adversarial Training to Learn Representations that are Robust to Word-Level Attacks", author = "Formento, Brian and Feng, Wenjie and Foo, Chuan-Sheng and Luu, Anh Tuan and Ng, See-Kiong", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.443", doi = "10.18653/v1/2024.naacl-long.443", pages = "8005--8028", abstract = "Language models (LMs) are indispensable tools for natural language processing tasks, but their vulnerability to adversarial attacks remains a concern. While current research has explored adversarial training techniques, their improvements to defend against word-level attacks have been limited. In this work, we propose a novel approach called Semantic Robust Defence (SemRoDe), a Macro Adversarial Training strategy to enhance the robustness of LMs. Drawing inspiration from recent studies in the image domain, we investigate and later confirm that in a discrete data setting such as language, adversarial samples generated via word substitutions do indeed belong to an adversarial domain exhibiting a high Wasserstein distance from the base domain. Our method learns a robust representation that bridges these two domains. We hypothesize that if samples were not projected into an adversarial domain, but instead to a domain with minimal shift, it would improve attack robustness. We align the domains by incorporating a new distance-based objective. With this, our model is able to learn more generalized representations by aligning the model{'}s high-level output features and therefore better handling unseen adversarial samples. This method can be generalized across word embeddings, even when they share minimal overlap at both vocabulary and word-substitution levels. To evaluate the effectiveness of our approach, we conduct experiments on BERT and RoBERTa models on three datasets. The results demonstrate promising state-of-the-art robustness.", }

Language models (LMs) are indispensable tools for natural language processing tasks, but their vulnerability to adversarial attacks remains a concern. While current research has explored adversarial training techniques, their improvements to defend against word-level attacks have been limited. In this work, we propose a novel approach called Semantic Robust Defence (SemRoDe), a Macro Adversarial Training strategy to enhance the robustness of LMs. Drawing inspiration from recent studies in the image domain, we investigate and later confirm that in a discrete data setting such as language, adversarial samples generated via word substitutions do indeed belong to an adversarial domain exhibiting a high Wasserstein distance from the base domain. Our method learns a robust representation that bridges these two domains. We hypothesize that if samples were not projected into an adversarial domain, but instead to a domain with minimal shift, it would improve attack robustness. We align the domains by incorporating a new distance-based objective. With this, our model is able to learn more generalized representations by aligning the model{'}s high-level output features and therefore better handling unseen adversarial samples. This method can be generalized across word embeddings, even when they share minimal overlap at both vocabulary and word-substitution levels. To evaluate the effectiveness of our approach, we conduct experiments on BERT and RoBERTa models on three datasets. The results demonstrate promising state-of-the-art robustness.

SemRoDe: Macro Adversarial Training to Learn Representations that are Robust to Word-Level Attacks

naacl-long.443

Oral

2403.18423

https://aclanthology.org/2024.naacl-long.444.bib

https://aclanthology.org/2024.naacl-long.444/

@inproceedings{cornelius-etal-2024-bust, title = "{BUST}: Benchmark for the evaluation of detectors of {LLM}-Generated Text", author = "Cornelius, Joseph and Lithgow-Serrano, Oscar and Mitrovic, Sandra and Dolamic, Ljiljana and Rinaldi, Fabio", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.444", doi = "10.18653/v1/2024.naacl-long.444", pages = "8029--8057", abstract = "We introduce BUST, a comprehensive benchmark designed to evaluate detectors of texts generated by instruction-tuned large language models (LLMs). Unlike previous benchmarks, our focus lies on evaluating the performance of detector systems, acknowledging the inevitable influence of the underlying tasks and different LLM generators. Our benchmark dataset consists of 25K texts from humans and 7 LLMs responding to instructions across 10 tasks from 3 diverse sources. Using the benchmark, we evaluated 5 detectors and found substantial performance variance across tasks. A meta-analysis of the dataset characteristics was conducted to guide the examination of detector performance. The dataset was analyzed using diverse metrics assessing linguistic features like fluency and coherence, readability scores, and writer attitudes, such as emotions, convincingness, and persuasiveness. Features impacting detector performance were investigated with surrogate models, revealing emotional content in texts enhanced some detectors, yet the most effective detector demonstrated consistent performance, irrespective of writer{'}s attitudes and text styles. Our approach focused on investigating relationships between the detectors{'} performance and two key factors: text characteristics and LLM generators. We believe BUST will provide valuable insights into selecting detectors tailored to specific text styles and tasks and facilitate a more practical and in-depth investigation of detection systems for LLM-generated text.", }

We introduce BUST, a comprehensive benchmark designed to evaluate detectors of texts generated by instruction-tuned large language models (LLMs). Unlike previous benchmarks, our focus lies on evaluating the performance of detector systems, acknowledging the inevitable influence of the underlying tasks and different LLM generators. Our benchmark dataset consists of 25K texts from humans and 7 LLMs responding to instructions across 10 tasks from 3 diverse sources. Using the benchmark, we evaluated 5 detectors and found substantial performance variance across tasks. A meta-analysis of the dataset characteristics was conducted to guide the examination of detector performance. The dataset was analyzed using diverse metrics assessing linguistic features like fluency and coherence, readability scores, and writer attitudes, such as emotions, convincingness, and persuasiveness. Features impacting detector performance were investigated with surrogate models, revealing emotional content in texts enhanced some detectors, yet the most effective detector demonstrated consistent performance, irrespective of writer{'}s attitudes and text styles. Our approach focused on investigating relationships between the detectors{'} performance and two key factors: text characteristics and LLM generators. We believe BUST will provide valuable insights into selecting detectors tailored to specific text styles and tasks and facilitate a more practical and in-depth investigation of detection systems for LLM-generated text.

BUST: Benchmark for the evaluation of detectors of LLM-Generated Text

naacl-long.444

Poster

https://aclanthology.org/2024.naacl-long.445.bib

https://aclanthology.org/2024.naacl-long.445/

@inproceedings{li-etal-2024-improving-context, title = "Improving In-context Learning of Multilingual Generative Language Models with Cross-lingual Alignment", author = "Li, Chong and Wang, Shaonan and Zhang, Jiajun and Zong, Chengqing", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.445", doi = "10.18653/v1/2024.naacl-long.445", pages = "8058--8076", abstract = "Multilingual generative models obtain remarkable cross-lingual in-context learning capabilities through pre-training on large-scale corpora. However, they still exhibit a performance bias toward high-resource languages and learn isolated distributions of multilingual sentence representations, which may hinder knowledge transfer across languages. To bridge this gap, we propose a simple yet effective cross-lingual alignment framework exploiting pairs of translation sentences. It aligns the internal sentence representations across different languages via multilingual contrastive learning and aligns outputs by following cross-lingual instructions in the target language. Experimental results show that even with less than 0.1${\textperthousand}$ of pre-training tokens, our alignment framework significantly boosts the cross-lingual abilities of generative language models and mitigates the performance gap. Further analyses reveal that it results in a better internal multilingual representation distribution of multilingual models.", }

Multilingual generative models obtain remarkable cross-lingual in-context learning capabilities through pre-training on large-scale corpora. However, they still exhibit a performance bias toward high-resource languages and learn isolated distributions of multilingual sentence representations, which may hinder knowledge transfer across languages. To bridge this gap, we propose a simple yet effective cross-lingual alignment framework exploiting pairs of translation sentences. It aligns the internal sentence representations across different languages via multilingual contrastive learning and aligns outputs by following cross-lingual instructions in the target language. Experimental results show that even with less than 0.1${\textperthousand}$ of pre-training tokens, our alignment framework significantly boosts the cross-lingual abilities of generative language models and mitigates the performance gap. Further analyses reveal that it results in a better internal multilingual representation distribution of multilingual models.

Improving In-context Learning of Multilingual Generative Language Models with Cross-lingual Alignment

naacl-long.445

Poster

2311.08089

https://huggingface.co/papers/2311.08089

https://aclanthology.org/2024.naacl-long.446.bib

https://aclanthology.org/2024.naacl-long.446/

@inproceedings{zhuang-etal-2024-macsc, title = "{M}a{CSC}: Towards Multimodal-augmented Pre-trained Language Models via Conceptual Prototypes and Self-balancing Calibration", author = "Zhuang, Xianwei and Wang, Zhichang and Cheng, Xuxin and Xie, Yuxin and Liang, Liming and Zou, Yuexian", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.446", doi = "10.18653/v1/2024.naacl-long.446", pages = "8077--8090", abstract = "Pre-trained language models (PLMs) that rely solely on textual data may exhibit limitations in multimodal semantics comprehension. Existing solutions attempt to alleviate this issue by incorporating explicit image retrieval or generation techniques.However, these methods: (1) focus exclusively on the static image modality; (2) inevitably encounter modality gaps and noise; (3) indiscriminately treat all modalities.In this paper, we propose a novel multimodal-augmented framework termed MaCSC, which can infuse multimodal semantics into PLMs and facilitate a self-balancing calibration of information allocation.Specifically, MaCSC obtains modal-specific conceptual prototypes from contrastive pre-training models (e.g., CLIP),and aggregates the intra- and inter-modal semantics of the conceptual prototype to enhance PLMs.In addition, we utilize a novel self-balancing contrastive loss to achieve multi-scale self-balancing calibration of multimodal information during fine-tuning PLMs.Experimental results show that MaCSC consistently improves the performance of PLMs across various architectures and scales, and outperforms competitive baselines on multiple NLP tasks.", }

Pre-trained language models (PLMs) that rely solely on textual data may exhibit limitations in multimodal semantics comprehension. Existing solutions attempt to alleviate this issue by incorporating explicit image retrieval or generation techniques.However, these methods: (1) focus exclusively on the static image modality; (2) inevitably encounter modality gaps and noise; (3) indiscriminately treat all modalities.In this paper, we propose a novel multimodal-augmented framework termed MaCSC, which can infuse multimodal semantics into PLMs and facilitate a self-balancing calibration of information allocation.Specifically, MaCSC obtains modal-specific conceptual prototypes from contrastive pre-training models (e.g., CLIP),and aggregates the intra- and inter-modal semantics of the conceptual prototype to enhance PLMs.In addition, we utilize a novel self-balancing contrastive loss to achieve multi-scale self-balancing calibration of multimodal information during fine-tuning PLMs.Experimental results show that MaCSC consistently improves the performance of PLMs across various architectures and scales, and outperforms competitive baselines on multiple NLP tasks.

MaCSC: Towards Multimodal-augmented Pre-trained Language Models via Conceptual Prototypes and Self-balancing Calibration

naacl-long.446

Oral

https://aclanthology.org/2024.naacl-long.447.bib

https://aclanthology.org/2024.naacl-long.447/

@inproceedings{sakai-etal-2024-pre, title = "Does Pre-trained Language Model Actually Infer Unseen Links in Knowledge Graph Completion?", author = "Sakai, Yusuke and Kamigaito, Hidetaka and Hayashi, Katsuhiko and Watanabe, Taro", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.447", doi = "10.18653/v1/2024.naacl-long.447", pages = "8091--8106", abstract = "Knowledge graphs (KGs) consist of links that describe relationships between entities. Due to the difficulty of manually enumerating all relationships between entities, automatically completing them is essential for KGs. Knowledge Graph Completion (KGC) is a task that infers unseen relationships between entities in a KG. Traditional embedding-based KGC methods (e.g. RESCAL, TransE, DistMult, ComplEx, RotatE, HAKE, HousE, etc.) infer missing links using only the knowledge from training data. In contrast, the recent Pre-trained Language Model (PLM)-based KGC utilizes knowledge obtained during pre-training, which means it can estimate missing links between entities by reusing memorized knowledge from pre-training without inference. This part is problematic because building KGC models aims to infer unseen links between entities. However, conventional evaluations in KGC do not consider inference and memorization abilities separately. Thus, a PLM-based KGC method, which achieves high performance in current KGC evaluations, may be ineffective in practical applications. To address this issue, we analyze whether PLM-based KGC methods make inferences or merely access memorized knowledge. For this purpose, we propose a method for constructing synthetic datasets specified in this analysis and conclude that PLMs acquire the inference abilities required for KGC through pre-training, even though the performance improvements mostly come from textual information of entities and relations.", }

Knowledge graphs (KGs) consist of links that describe relationships between entities. Due to the difficulty of manually enumerating all relationships between entities, automatically completing them is essential for KGs. Knowledge Graph Completion (KGC) is a task that infers unseen relationships between entities in a KG. Traditional embedding-based KGC methods (e.g. RESCAL, TransE, DistMult, ComplEx, RotatE, HAKE, HousE, etc.) infer missing links using only the knowledge from training data. In contrast, the recent Pre-trained Language Model (PLM)-based KGC utilizes knowledge obtained during pre-training, which means it can estimate missing links between entities by reusing memorized knowledge from pre-training without inference. This part is problematic because building KGC models aims to infer unseen links between entities. However, conventional evaluations in KGC do not consider inference and memorization abilities separately. Thus, a PLM-based KGC method, which achieves high performance in current KGC evaluations, may be ineffective in practical applications. To address this issue, we analyze whether PLM-based KGC methods make inferences or merely access memorized knowledge. For this purpose, we propose a method for constructing synthetic datasets specified in this analysis and conclude that PLMs acquire the inference abilities required for KGC through pre-training, even though the performance improvements mostly come from textual information of entities and relations.

Does Pre-trained Language Model Actually Infer Unseen Links in Knowledge Graph Completion?

naacl-long.447

Oral

2311.09109

https://huggingface.co/papers/2311.09109

https://aclanthology.org/2024.naacl-long.448.bib

https://aclanthology.org/2024.naacl-long.448/

@inproceedings{suresh-etal-2024-discovering, title = "Discovering Lobby-Parliamentarian Alignments through {NLP}", author = "Suresh, Aswin and Radojevi{\'c}, Lazar and Salvi, Francesco and Magron, Antoine and Kristof, Victor and Grossglauser, Matthias", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.448", doi = "10.18653/v1/2024.naacl-long.448", pages = "8107--8120", abstract = "We discover alignments of views between interest groups (lobbies) and members of the European Parliament (MEPs) by automatically analyzing their texts. Specifically, we do so by collecting novel datasets of lobbies{'} position papers and MEPs{'} speeches, and comparing these texts on the basis of semantic similarity and entailment. In the absence of ground-truth, we perform an indirect validation by comparing the discovered alignments with a dataset, which we curate, of retweet links between MEPs and lobbies, and with the publicly disclosed meetings of MEPs. Our best method performs significantly better than several baselines. Moreover, an aggregate analysis of the discovered alignments, between groups of related lobbies and political groups of MEPs, correspond to the expectations from the ideology of the groups (e.g., groups on the political left are more aligned with humanitarian and environmental organisations). We believe that this work is a step towards enhancing the transparency of the intricate decision-making processes within democratic institutions.", }

We discover alignments of views between interest groups (lobbies) and members of the European Parliament (MEPs) by automatically analyzing their texts. Specifically, we do so by collecting novel datasets of lobbies{'} position papers and MEPs{'} speeches, and comparing these texts on the basis of semantic similarity and entailment. In the absence of ground-truth, we perform an indirect validation by comparing the discovered alignments with a dataset, which we curate, of retweet links between MEPs and lobbies, and with the publicly disclosed meetings of MEPs. Our best method performs significantly better than several baselines. Moreover, an aggregate analysis of the discovered alignments, between groups of related lobbies and political groups of MEPs, correspond to the expectations from the ideology of the groups (e.g., groups on the political left are more aligned with humanitarian and environmental organisations). We believe that this work is a step towards enhancing the transparency of the intricate decision-making processes within democratic institutions.

Discovering Lobby-Parliamentarian Alignments through NLP

naacl-long.448

Poster

https://aclanthology.org/2024.naacl-long.449.bib

https://aclanthology.org/2024.naacl-long.449/

@inproceedings{jang-etal-2024-itercqr, title = "{I}ter{CQR}: Iterative Conversational Query Reformulation with Retrieval Guidance", author = "Jang, Yunah and Lee, Kang-il and Bae, Hyunkyung and Lee, Hwanhee and Jung, Kyomin", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.449", doi = "10.18653/v1/2024.naacl-long.449", pages = "8121--8138", abstract = "Conversational search aims to retrieve passages containing essential information to answer queries in a multi-turn conversation. In conversational search, reformulating context-dependent conversational queries into stand-alone forms is imperative to effectively utilize off-the-shelf retrievers. Previous methodologies for conversational query reformulation frequently depend on human-annotated rewrites.However, these manually crafted queries often result in sub-optimal retrieval performance and require high collection costs.To address these challenges, we propose **Iter**ative **C**onversational **Q**uery **R**eformulation (**IterCQR**), a methodology that conducts query reformulation without relying on human rewrites. IterCQR iteratively trains the conversational query reformulation (CQR) model by directly leveraging information retrieval (IR) signals as a reward.Our IterCQR training guides the CQR model such that generated queries contain necessary information from the previous dialogue context.Our proposed method shows state-of-the-art performance on two widely-used datasets, demonstrating its effectiveness on both sparse and dense retrievers. Moreover, IterCQR exhibits superior performance in challenging settings such as generalization on unseen datasets and low-resource scenarios.", }

Conversational search aims to retrieve passages containing essential information to answer queries in a multi-turn conversation. In conversational search, reformulating context-dependent conversational queries into stand-alone forms is imperative to effectively utilize off-the-shelf retrievers. Previous methodologies for conversational query reformulation frequently depend on human-annotated rewrites.However, these manually crafted queries often result in sub-optimal retrieval performance and require high collection costs.To address these challenges, we propose **Iter**ative **C**onversational **Q**uery **R**eformulation (**IterCQR**), a methodology that conducts query reformulation without relying on human rewrites. IterCQR iteratively trains the conversational query reformulation (CQR) model by directly leveraging information retrieval (IR) signals as a reward.Our IterCQR training guides the CQR model such that generated queries contain necessary information from the previous dialogue context.Our proposed method shows state-of-the-art performance on two widely-used datasets, demonstrating its effectiveness on both sparse and dense retrievers. Moreover, IterCQR exhibits superior performance in challenging settings such as generalization on unseen datasets and low-resource scenarios.

IterCQR: Iterative Conversational Query Reformulation with Retrieval Guidance

naacl-long.449

Poster

2311.09820

https://aclanthology.org/2024.naacl-long.450.bib

https://aclanthology.org/2024.naacl-long.450/

@inproceedings{huang-etal-2024-acegpt, title = "{A}ce{GPT}, Localizing Large Language Models in {A}rabic", author = "Huang, Huang and Yu, Fei and Zhu, Jianqing and Sun, Xuening and Cheng, Hao and Dingjie, Song and Chen, Zhihong and Alharthi, Mosen and An, Bang and He, Juncai and Liu, Ziche and Chen, Junying and Li, Jianquan and Wang, Benyou and Zhang, Lian and Sun, Ruoyu and Wan, Xiang and Li, Haizhou and Xu, Jinchao", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.450", doi = "10.18653/v1/2024.naacl-long.450", pages = "8139--8163", abstract = "This paper is devoted to the development of a localized Large Language Model (LLM) specifically for Arabic, a language imbued with unique cultural characteristics inadequately addressed by current mainstream models. Significant concerns emerge when addressing cultural sensitivity and local values. To address this, the paper proposes a comprehensive solution that includes further pre-training with Arabic texts, Supervised Fine-Tuning (SFT) utilizing native Arabic instructions, and GPT-4 responses in Arabic, alongside Reinforcement Learning with AI Feedback (RLAIF) employing a reward model attuned to local culture and values. The goal is to cultivate culturally cognizant and value-aligned Arabic LLMs capable of accommodating the diverse, application-specific needs of Arabic-speaking communities. Comprehensive evaluations reveal that the resulting model, dubbed {`}AceGPT{'}, sets the state-of-the-art standard for open Arabic LLMs across various benchmarks. Codes, data, and models are in https://github.com/FreedomIntelligence/AceGPT.", }

This paper is devoted to the development of a localized Large Language Model (LLM) specifically for Arabic, a language imbued with unique cultural characteristics inadequately addressed by current mainstream models. Significant concerns emerge when addressing cultural sensitivity and local values. To address this, the paper proposes a comprehensive solution that includes further pre-training with Arabic texts, Supervised Fine-Tuning (SFT) utilizing native Arabic instructions, and GPT-4 responses in Arabic, alongside Reinforcement Learning with AI Feedback (RLAIF) employing a reward model attuned to local culture and values. The goal is to cultivate culturally cognizant and value-aligned Arabic LLMs capable of accommodating the diverse, application-specific needs of Arabic-speaking communities. Comprehensive evaluations reveal that the resulting model, dubbed {`}AceGPT{'}, sets the state-of-the-art standard for open Arabic LLMs across various benchmarks. Codes, data, and models are in https://github.com/FreedomIntelligence/AceGPT.

AceGPT, Localizing Large Language Models in Arabic

naacl-long.450

Poster

2309.12053

https://huggingface.co/papers/2309.12053

https://aclanthology.org/2024.naacl-long.451.bib

https://aclanthology.org/2024.naacl-long.451/

@inproceedings{he-etal-2024-improving, title = "Improving Machine Translation with Human Feedback: An Exploration of Quality Estimation as a Reward Model", author = "He, Zhiwei and Wang, Xing and Jiao, Wenxiang and Zhang, Zhuosheng and Wang, Rui and Shi, Shuming and Tu, Zhaopeng", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.451", doi = "10.18653/v1/2024.naacl-long.451", pages = "8164--8180", abstract = "Insufficient modeling of human preferences within the reward model is a major obstacle for leveraging human feedback to improve translation quality. Fortunately, quality estimation (QE), which predicts the quality of a given translation without reference, has achieved impressive alignment with human evaluations in the last two years. In this work, we investigate the potential of employing the QE model as the reward model to predict human preferences for feedback training. We first identify the overoptimization problem during QE-based feedback training, manifested as an increase in reward while translation quality declines. We examine the problem and argue that the vulnerability of the QE model might lead to high rewards for incorrect translations, resulting in overoptimization and error propagation. To address the problem, we adopt a simple yet effective method that uses heuristic rules to detect the incorrect translations and assigns a penalty term to the reward scores of them. Experimental results show that the proposed QE-based feedback training achieves consistent and significant improvements across various settings, further verified through human preference studies. Our subsequent analysis demonstrates the high data efficiency of the proposed QE-based feedback training: it outperforms systems using larger parallel corpora by a small amount of monolingual data. Our code is available at: https://github.com/zwhe99/FeedbackMT", }

Insufficient modeling of human preferences within the reward model is a major obstacle for leveraging human feedback to improve translation quality. Fortunately, quality estimation (QE), which predicts the quality of a given translation without reference, has achieved impressive alignment with human evaluations in the last two years. In this work, we investigate the potential of employing the QE model as the reward model to predict human preferences for feedback training. We first identify the overoptimization problem during QE-based feedback training, manifested as an increase in reward while translation quality declines. We examine the problem and argue that the vulnerability of the QE model might lead to high rewards for incorrect translations, resulting in overoptimization and error propagation. To address the problem, we adopt a simple yet effective method that uses heuristic rules to detect the incorrect translations and assigns a penalty term to the reward scores of them. Experimental results show that the proposed QE-based feedback training achieves consistent and significant improvements across various settings, further verified through human preference studies. Our subsequent analysis demonstrates the high data efficiency of the proposed QE-based feedback training: it outperforms systems using larger parallel corpora by a small amount of monolingual data. Our code is available at: https://github.com/zwhe99/FeedbackMT

Improving Machine Translation with Human Feedback: An Exploration of Quality Estimation as a Reward Model

naacl-long.451

Poster

2401.12873

https://aclanthology.org/2024.naacl-long.452.bib

https://aclanthology.org/2024.naacl-long.452/

@inproceedings{chen-etal-2024-depression, title = "Depression Detection in Clinical Interviews with {LLM}-Empowered Structural Element Graph", author = "Chen, Zhuang and Deng, Jiawen and Zhou, Jinfeng and Wu, Jincenzi and Qian, Tieyun and Huang, Minlie", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.452", doi = "10.18653/v1/2024.naacl-long.452", pages = "8181--8194", abstract = "Depression is a widespread mental health disorder affecting millions globally. Clinical interviews are the gold standard for assessing depression, but they heavily rely on scarce professional clinicians, highlighting the need for automated detection systems. However, existing methods only capture part of the relevant elements in clinical interviews, unable to incorporate all depressive cues. Moreover, the scarcity of participant data, due to privacy concerns and collection challenges, intrinsically constrains interview modeling. To address these limitations, in this paper, we propose a structural element graph (SEGA), which transforms the clinical interview into an expertise-inspired directed acyclic graph for comprehensive modeling. Additionally, we further empower SEGA by devising novel principle-guided data augmentation with large language models (LLMs) to supplement high-quality synthetic data and enable graph contrastive learning. Extensive evaluations on two real-world clinical datasets, in both English and Chinese, show that SEGA significantly outperforms baseline methods and powerful LLMs like GPT-3.5 and GPT-4.", }

Depression is a widespread mental health disorder affecting millions globally. Clinical interviews are the gold standard for assessing depression, but they heavily rely on scarce professional clinicians, highlighting the need for automated detection systems. However, existing methods only capture part of the relevant elements in clinical interviews, unable to incorporate all depressive cues. Moreover, the scarcity of participant data, due to privacy concerns and collection challenges, intrinsically constrains interview modeling. To address these limitations, in this paper, we propose a structural element graph (SEGA), which transforms the clinical interview into an expertise-inspired directed acyclic graph for comprehensive modeling. Additionally, we further empower SEGA by devising novel principle-guided data augmentation with large language models (LLMs) to supplement high-quality synthetic data and enable graph contrastive learning. Extensive evaluations on two real-world clinical datasets, in both English and Chinese, show that SEGA significantly outperforms baseline methods and powerful LLMs like GPT-3.5 and GPT-4.

Depression Detection in Clinical Interviews with LLM-Empowered Structural Element Graph

naacl-long.452

Poster

https://aclanthology.org/2024.naacl-long.453.bib

https://aclanthology.org/2024.naacl-long.453/

@inproceedings{razumovskaia-etal-2024-sqatin, title = "{SQATIN}: Supervised Instruction Tuning Meets Question Answering for Improved Dialogue {NLU}", author = "Razumovskaia, Evgeniia and Glava{\v{s}}, Goran and Korhonen, Anna and Vuli{\'c}, Ivan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.453", doi = "10.18653/v1/2024.naacl-long.453", pages = "8195--8211", abstract = "Task-oriented dialogue (TOD) systems help users execute well-defined tasks across a variety of domains (e.g., \textit{flight booking} or \textit{food ordering}), with their Natural Language Understanding (NLU) components being dedicated to the analysis of user utterances, predicting users{'} intents (\textit{Intent Detection}, ID) and extracting values for informational slots (\textit{Value Extraction}, VE). In most domains, labelled NLU data is scarce, making sample-efficient learning {--} enabled with effective transfer paradigms {--} paramount. In this work, we introduce SQATIN, a new framework for dialog NLU based on (i) instruction tuning and (ii) question-answering-based formulation of ID and VE tasks. According to the evaluation on established NLU benchmarks, SQATIN sets the new state of the art in dialogue NLU, substantially surpassing the performance of current models based on standard fine-tuning objectives in both in-domain training and cross-domain transfer, and it also surpasses off-the-shelf large language models for the same task, both in terms of performance and inference efficiency. Furthermore, SQATIN yields particularly large performance gains in cross-domain transfer, owing to the fact that our QA-based instruction tuning leverages similarities between natural language descriptions of classes (i.e., slots and intents) across domains.", }

Task-oriented dialogue (TOD) systems help users execute well-defined tasks across a variety of domains (e.g., \textit{flight booking} or \textit{food ordering}), with their Natural Language Understanding (NLU) components being dedicated to the analysis of user utterances, predicting users{'} intents (\textit{Intent Detection}, ID) and extracting values for informational slots (\textit{Value Extraction}, VE). In most domains, labelled NLU data is scarce, making sample-efficient learning {--} enabled with effective transfer paradigms {--} paramount. In this work, we introduce SQATIN, a new framework for dialog NLU based on (i) instruction tuning and (ii) question-answering-based formulation of ID and VE tasks. According to the evaluation on established NLU benchmarks, SQATIN sets the new state of the art in dialogue NLU, substantially surpassing the performance of current models based on standard fine-tuning objectives in both in-domain training and cross-domain transfer, and it also surpasses off-the-shelf large language models for the same task, both in terms of performance and inference efficiency. Furthermore, SQATIN yields particularly large performance gains in cross-domain transfer, owing to the fact that our QA-based instruction tuning leverages similarities between natural language descriptions of classes (i.e., slots and intents) across domains.

SQATIN: Supervised Instruction Tuning Meets Question Answering for Improved Dialogue NLU

naacl-long.453

Poster

2311.09502

https://aclanthology.org/2024.naacl-long.454.bib

https://aclanthology.org/2024.naacl-long.454/

@inproceedings{khosravani-etal-2024-enhancing, title = "Enhancing Argument Summarization: Prioritizing Exhaustiveness in Key Point Generation and Introducing an Automatic Coverage Evaluation Metric", author = "Khosravani, Mohammad and Huang, Chenyang and Trabelsi, Amine", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.454", doi = "10.18653/v1/2024.naacl-long.454", pages = "8212--8224", abstract = "The proliferation of social media platforms has given rise to the amount of online debates and arguments. Consequently, the need for automatic summarization methods for such debates is imperative, however this area of summarization is rather understudied. The Key Point Analysis (KPA) task formulates argument summarization as representing the summary of a large collection of arguments in the form of concise sentences in bullet-style format, called key points. A sub-task of KPA, called Key Point Generation (KPG), focuses on generating these key points given the arguments. This paper introduces a novel extractive approach for key point generation, that outperforms previous state-of-the-art methods for the task. Our method utilizes an extractive clustering based approach that offers concise, high quality generated key points with higher coverage of reference summaries, and less redundant outputs. In addition, we show that the existing evaluation metrics for summarization such as ROUGE are incapable of differentiating between generated key points of different qualities. To this end, we propose a new evaluation metric for assessing the generated key points by their coverage. Our code can be accessed online.", }

The proliferation of social media platforms has given rise to the amount of online debates and arguments. Consequently, the need for automatic summarization methods for such debates is imperative, however this area of summarization is rather understudied. The Key Point Analysis (KPA) task formulates argument summarization as representing the summary of a large collection of arguments in the form of concise sentences in bullet-style format, called key points. A sub-task of KPA, called Key Point Generation (KPG), focuses on generating these key points given the arguments. This paper introduces a novel extractive approach for key point generation, that outperforms previous state-of-the-art methods for the task. Our method utilizes an extractive clustering based approach that offers concise, high quality generated key points with higher coverage of reference summaries, and less redundant outputs. In addition, we show that the existing evaluation metrics for summarization such as ROUGE are incapable of differentiating between generated key points of different qualities. To this end, we propose a new evaluation metric for assessing the generated key points by their coverage. Our code can be accessed online.

Enhancing Argument Summarization: Prioritizing Exhaustiveness in Key Point Generation and Introducing an Automatic Coverage Evaluation Metric

naacl-long.454

Oral

2404.11793

https://aclanthology.org/2024.naacl-long.455.bib

https://aclanthology.org/2024.naacl-long.455/

@inproceedings{pang-etal-2024-arm, title = "{ARM}: Alignment with Residual Energy-Based Model", author = "Pang, Bo and Xiong, Caiming and Zhou, Yingbo", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.455", doi = "10.18653/v1/2024.naacl-long.455", pages = "8225--8236", abstract = "While large language models (LLMs) trained with large-scale unsupervised learning acquire a wide variety of world knowledge and skills, its behavior does not necessarily align with human preferences. RLHF methods achieve successes in aligning LLM responses with human preferences and improving the controllability of LLM behavior with human instruction. However, RLHF methods are considerably complicated to implement, computationally expensive to train, and notoriously tricky to tune. In this work, we propose Alignment with Residual Energy-Based Model (ARM), as a simple and flexible alternative to RLHF methods. Our method is driven by an observation that we can learn an aligned policy by minimizing a forward Kullback{--}Leibler (KL) divergence from a target policy (in the form of a residual energy-based model) to a parameteric policy (LLM), instead of a reverse KL as in RLHF methods. With samples from the energy-based target policy, we can leverage the power of DPO (or other offline methods) to learn an aligned policy efficiently. ARM is simple to implement and applicable in various data settings. Our extensive experiments demonstrate its strong performance across multiple datasets, compared to strong baselines like PPO, DPO.", }

While large language models (LLMs) trained with large-scale unsupervised learning acquire a wide variety of world knowledge and skills, its behavior does not necessarily align with human preferences. RLHF methods achieve successes in aligning LLM responses with human preferences and improving the controllability of LLM behavior with human instruction. However, RLHF methods are considerably complicated to implement, computationally expensive to train, and notoriously tricky to tune. In this work, we propose Alignment with Residual Energy-Based Model (ARM), as a simple and flexible alternative to RLHF methods. Our method is driven by an observation that we can learn an aligned policy by minimizing a forward Kullback{--}Leibler (KL) divergence from a target policy (in the form of a residual energy-based model) to a parameteric policy (LLM), instead of a reverse KL as in RLHF methods. With samples from the energy-based target policy, we can leverage the power of DPO (or other offline methods) to learn an aligned policy efficiently. ARM is simple to implement and applicable in various data settings. Our extensive experiments demonstrate its strong performance across multiple datasets, compared to strong baselines like PPO, DPO.

ARM: Alignment with Residual Energy-Based Model

naacl-long.455

Oral

https://aclanthology.org/2024.naacl-long.456.bib

https://aclanthology.org/2024.naacl-long.456/

@inproceedings{gritta-etal-2024-humanrankeval, title = "{H}uman{R}ank{E}val: Automatic Evaluation of {LM}s as Conversational Assistants", author = "Gritta, Milan and Lampouras, Gerasimos and Iacobacci, Ignacio", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.456", doi = "10.18653/v1/2024.naacl-long.456", pages = "8237--8249", abstract = "Language models (LMs) as conversational assistants recently became popular tools that help people accomplish a variety of tasks. These typically result from adapting LMs pretrained on general domain text sequences through further instruction-tuning and possibly preference optimisation methods. The evaluation of such LMs would ideally be performed using human judgement, however, this is not scalable. On the other hand, automatic evaluation featuring auxiliary LMs as judges and/or knowledge-based tasks is scalable but struggles with assessing conversational ability and adherence to instructions. To help accelerate the development of LMs as conversational assistants, we propose a novel automatic evaluation task: HumanRankEval (HRE). It consists of a large-scale, diverse and high-quality set of questions, each with several answers authored and scored by humans. To perform evaluation, HRE ranks these answers based on their log-likelihood under the LM{'}s distribution, and subsequently calculates their correlation with the corresponding human rankings. We support HRE{'}s efficacy by investigating how efficiently it separates pretrained and instruction-tuned LMs of various sizes. We show that HRE correlates well with human judgements and is particularly responsive to model changes following instruction-tuning.", }

Language models (LMs) as conversational assistants recently became popular tools that help people accomplish a variety of tasks. These typically result from adapting LMs pretrained on general domain text sequences through further instruction-tuning and possibly preference optimisation methods. The evaluation of such LMs would ideally be performed using human judgement, however, this is not scalable. On the other hand, automatic evaluation featuring auxiliary LMs as judges and/or knowledge-based tasks is scalable but struggles with assessing conversational ability and adherence to instructions. To help accelerate the development of LMs as conversational assistants, we propose a novel automatic evaluation task: HumanRankEval (HRE). It consists of a large-scale, diverse and high-quality set of questions, each with several answers authored and scored by humans. To perform evaluation, HRE ranks these answers based on their log-likelihood under the LM{'}s distribution, and subsequently calculates their correlation with the corresponding human rankings. We support HRE{'}s efficacy by investigating how efficiently it separates pretrained and instruction-tuned LMs of various sizes. We show that HRE correlates well with human judgements and is particularly responsive to model changes following instruction-tuning.

HumanRankEval: Automatic Evaluation of LMs as Conversational Assistants

naacl-long.456

Poster

2405.09186

https://huggingface.co/papers/2405.09186

https://aclanthology.org/2024.naacl-long.457.bib

https://aclanthology.org/2024.naacl-long.457/

@inproceedings{vashishtha-etal-2024-famus, title = "{FAM}u{S}: Frames Across Multiple Sources", author = "Vashishtha, Siddharth and Martin, Alexander and Gantt, William and Van Durme, Benjamin and White, Aaron", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.457", doi = "10.18653/v1/2024.naacl-long.457", pages = "8250--8273", abstract = "Understanding event descriptions is a central aspect of language processing, but current approaches focus overwhelmingly on single sentences or documents. Aggregating information about an event across documents can offer a much richer understanding. To this end, we present FAMuS, a new corpus of Wikipedia passages that report on some event, paired with underlying, genre-diverse (non-Wikipedia) source articles for the same event. Events and (cross-sentence) arguments in both report and source are annotated against FrameNet, providing broad coverage of different event types. We present results on two key event understanding tasks enabled by FAMuS: source validation{---}determining whether a document is a valid source for a target report event{---}and cross-document argument extraction{---}full-document argument extraction for a target event from both its report and the correct source article.", }

Understanding event descriptions is a central aspect of language processing, but current approaches focus overwhelmingly on single sentences or documents. Aggregating information about an event across documents can offer a much richer understanding. To this end, we present FAMuS, a new corpus of Wikipedia passages that report on some event, paired with underlying, genre-diverse (non-Wikipedia) source articles for the same event. Events and (cross-sentence) arguments in both report and source are annotated against FrameNet, providing broad coverage of different event types. We present results on two key event understanding tasks enabled by FAMuS: source validation{---}determining whether a document is a valid source for a target report event{---}and cross-document argument extraction{---}full-document argument extraction for a target event from both its report and the correct source article.

FAMuS: Frames Across Multiple Sources

naacl-long.457

Poster

2311.05601

https://aclanthology.org/2024.naacl-long.458.bib

https://aclanthology.org/2024.naacl-long.458/

@inproceedings{li-chaturvedi-2024-rationale, title = "Rationale-based Opinion Summarization", author = "Li, Haoyuan and Chaturvedi, Snigdha", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.458", doi = "10.18653/v1/2024.naacl-long.458", pages = "8274--8292", abstract = "Opinion summarization aims to generate concise summaries that present popular opinions of a large group of reviews. However, these summaries can be too generic and lack supporting details. To address these issues, we propose a new paradigm for summarizing reviews, rationale-based opinion summarization. Rationale-based opinion summaries output the representative opinions as well as one or more corresponding rationales. To extract good rationales, we define four desirable properties: relatedness, specificity, popularity, and diversity and present a Gibbs-sampling-based method to extract rationales. Overall, we propose RATION, an unsupervised extractive system that has two components: an Opinion Extractor (to extract representative opinions) and Rationales Extractor (to extract corresponding rationales). We conduct automatic and human evaluations to show that rationales extracted by RATION have the proposed properties and its summaries are more useful than conventional summaries. The implementation of our work is available at https://github.com/leehaoyuan/RATION.", }

Opinion summarization aims to generate concise summaries that present popular opinions of a large group of reviews. However, these summaries can be too generic and lack supporting details. To address these issues, we propose a new paradigm for summarizing reviews, rationale-based opinion summarization. Rationale-based opinion summaries output the representative opinions as well as one or more corresponding rationales. To extract good rationales, we define four desirable properties: relatedness, specificity, popularity, and diversity and present a Gibbs-sampling-based method to extract rationales. Overall, we propose RATION, an unsupervised extractive system that has two components: an Opinion Extractor (to extract representative opinions) and Rationales Extractor (to extract corresponding rationales). We conduct automatic and human evaluations to show that rationales extracted by RATION have the proposed properties and its summaries are more useful than conventional summaries. The implementation of our work is available at https://github.com/leehaoyuan/RATION.

Rationale-based Opinion Summarization

naacl-long.458

Poster

2404.00217

https://aclanthology.org/2024.naacl-long.459.bib

https://aclanthology.org/2024.naacl-long.459/

@inproceedings{melechovsky-etal-2024-mustango, title = "Mustango: Toward Controllable Text-to-Music Generation", author = "Melechovsky, Jan and Guo, Zixun and Ghosal, Deepanway and Majumder, Navonil and Herremans, Dorien and Poria, Soujanya", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.459", doi = "10.18653/v1/2024.naacl-long.459", pages = "8293--8316", abstract = "The quality of the text-to-music models has reached new heights due to recent advancements in diffusion models. The controllability of various musical aspects, however, has barely been explored. In this paper, we propose Mustango: a music-domain-knowledge-inspired text-to-music system based on diffusion. Mustango aims to control the generated music, not only with general text captions, but with more rich captions that can include specific instructions related to chords, beats, tempo, and key. At the core of Mustango is MuNet, a Music-Domain-Knowledge-Informed UNet guidance module that steers the generated music to include the music-specific conditions, which we predict from the text prompt, as well as the general text embedding, during the reverse diffusion process. To overcome the limited availability of open datasets of music with text captions, we propose a novel data augmentation method that includes altering the harmonic, rhythmic, and dynamic aspects of music audio and using state-of-the-art Music Information Retrieval methods to extract the music features which will then be appended to the existing descriptions in text format. We release the resulting MusicBench dataset which contains over 52K instances and includes music-theory-based descriptions in the caption text. Through extensive experiments, we show that the quality of the music generated by Mustango is state-of-the-art, and the controllability through music-specific text prompts greatly outperforms other models such as MusicGen and AudioLDM2.", }

The quality of the text-to-music models has reached new heights due to recent advancements in diffusion models. The controllability of various musical aspects, however, has barely been explored. In this paper, we propose Mustango: a music-domain-knowledge-inspired text-to-music system based on diffusion. Mustango aims to control the generated music, not only with general text captions, but with more rich captions that can include specific instructions related to chords, beats, tempo, and key. At the core of Mustango is MuNet, a Music-Domain-Knowledge-Informed UNet guidance module that steers the generated music to include the music-specific conditions, which we predict from the text prompt, as well as the general text embedding, during the reverse diffusion process. To overcome the limited availability of open datasets of music with text captions, we propose a novel data augmentation method that includes altering the harmonic, rhythmic, and dynamic aspects of music audio and using state-of-the-art Music Information Retrieval methods to extract the music features which will then be appended to the existing descriptions in text format. We release the resulting MusicBench dataset which contains over 52K instances and includes music-theory-based descriptions in the caption text. Through extensive experiments, we show that the quality of the music generated by Mustango is state-of-the-art, and the controllability through music-specific text prompts greatly outperforms other models such as MusicGen and AudioLDM2.

Mustango: Toward Controllable Text-to-Music Generation

naacl-long.459

Oral

2311.08355

https://aclanthology.org/2024.naacl-long.460.bib

https://aclanthology.org/2024.naacl-long.460/

@inproceedings{cueva-etal-2024-adaptive, title = "Adaptive Cross-lingual Text Classification through In-Context One-Shot Demonstrations", author = "Cueva, Emilio and Lopez Monroy, Adrian and S{\'a}nchez-Vega, Fernando and Solorio, Thamar", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.460", doi = "10.18653/v1/2024.naacl-long.460", pages = "8317--8335", abstract = "Zero-Shot Cross-lingual Transfer (ZS-XLT) utilizes a model trained in a source language to make predictions in another language, often with a performance loss. To alleviate this, additional improvements can be achieved through subsequent adaptation using examples in the target language. In this paper, we exploit In-Context Tuning (ICT) for One-Shot Cross-lingual transfer in the classification task by introducing In-Context Cross-lingual Transfer (IC-XLT). The novel concept involves training a model to learn from context examples and subsequently adapting it during inference to a target language by prepending a One-Shot context demonstration in that language. Our results show that IC-XLT successfully leverages target-language examples to improve the cross-lingual capabilities of the evaluated mT5 model, outperforming prompt-based models in the Zero and Few-shot scenarios adapted through fine-tuning. Moreover, we show that when source-language data is limited, the fine-tuning framework employed for IC-XLT performs comparably to prompt-based fine-tuning with significantly more training data in the source language.", }

Zero-Shot Cross-lingual Transfer (ZS-XLT) utilizes a model trained in a source language to make predictions in another language, often with a performance loss. To alleviate this, additional improvements can be achieved through subsequent adaptation using examples in the target language. In this paper, we exploit In-Context Tuning (ICT) for One-Shot Cross-lingual transfer in the classification task by introducing In-Context Cross-lingual Transfer (IC-XLT). The novel concept involves training a model to learn from context examples and subsequently adapting it during inference to a target language by prepending a One-Shot context demonstration in that language. Our results show that IC-XLT successfully leverages target-language examples to improve the cross-lingual capabilities of the evaluated mT5 model, outperforming prompt-based models in the Zero and Few-shot scenarios adapted through fine-tuning. Moreover, we show that when source-language data is limited, the fine-tuning framework employed for IC-XLT performs comparably to prompt-based fine-tuning with significantly more training data in the source language.

Adaptive Cross-lingual Text Classification through In-Context One-Shot Demonstrations

naacl-long.460

Oral

2404.02452

https://aclanthology.org/2024.naacl-long.461.bib

https://aclanthology.org/2024.naacl-long.461/

@inproceedings{martinelli-etal-2024-cner, title = "{CNER}: Concept and Named Entity Recognition", author = "Martinelli, Giuliano and Molfese, Francesco and Tedeschi, Simone and Fern{\'a}ndez-Castro, Alberte and Navigli, Roberto", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.461", doi = "10.18653/v1/2024.naacl-long.461", pages = "8336--8351", abstract = "Named entities {--} typically expressed via proper nouns {--} play a key role in Natural Language Processing, as their identification and comprehension are crucial in tasks such as Relation Extraction, Coreference Resolution and Question Answering, among others. Tasks like these also often entail dealing with concepts {--} typically represented by common nouns {--} which, however, have not received as much attention. Indeed, the potential of their identification and understanding remains underexplored, as does the benefit of a synergistic formulation with named entities. To fill this gap, we introduce Concept and Named Entity Recognition (CNER), a new unified task that handles concepts and entities mentioned in unstructured texts seamlessly. We put forward a comprehensive set of categories that can be used to model concepts and named entities jointly, and propose new approaches for the creation of CNER datasets. We evaluate the benefits of performing CNER as a unified task extensively, showing that a CNER model gains up to +5.4 and +8 macro F1 points when compared to specialized named entity and concept recognition systems, respectively. Finally, to encourage the development of CNER systems, we release our datasets and models at https://github.com/Babelscape/cner.", }

Named entities {--} typically expressed via proper nouns {--} play a key role in Natural Language Processing, as their identification and comprehension are crucial in tasks such as Relation Extraction, Coreference Resolution and Question Answering, among others. Tasks like these also often entail dealing with concepts {--} typically represented by common nouns {--} which, however, have not received as much attention. Indeed, the potential of their identification and understanding remains underexplored, as does the benefit of a synergistic formulation with named entities. To fill this gap, we introduce Concept and Named Entity Recognition (CNER), a new unified task that handles concepts and entities mentioned in unstructured texts seamlessly. We put forward a comprehensive set of categories that can be used to model concepts and named entities jointly, and propose new approaches for the creation of CNER datasets. We evaluate the benefits of performing CNER as a unified task extensively, showing that a CNER model gains up to +5.4 and +8 macro F1 points when compared to specialized named entity and concept recognition systems, respectively. Finally, to encourage the development of CNER systems, we release our datasets and models at https://github.com/Babelscape/cner.

CNER: Concept and Named Entity Recognition

naacl-long.461

Poster

https://aclanthology.org/2024.naacl-long.462.bib

https://aclanthology.org/2024.naacl-long.462/

@inproceedings{saha-etal-2024-branch, title = "Branch-Solve-Merge Improves Large Language Model Evaluation and Generation", author = "Saha, Swarnadeep and Levy, Omer and Celikyilmaz, Asli and Bansal, Mohit and Weston, Jason and Li, Xian", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.462", doi = "10.18653/v1/2024.naacl-long.462", pages = "8352--8370", abstract = "Large Language Models (LLMs) are frequently used for multi-faceted language generation and evaluation tasks that involve satisfying intricate user constraints or taking into account multiple aspects and criteria. However, their performance can fall short, due to the model{'}s lack of coherence and inability to plan and decompose the problem. We propose Branch-Solve-Merge (BSM), a Large Language Model program (Schlag et al., 2023) for tackling such challenging natural language tasks. It consists of branch, solve, and merge modules that are parameterized with specific prompts to the base LLM. These three modules plan a decomposition of the task into multiple parallel sub-tasks, independently solve them, and fuse the solutions to the sub-tasks. We apply our method to the tasks of LLM response evaluation and constrained text generation and evaluate its effectiveness with multiple LLMs, including Vicuna, LLaMA-2-chat, and GPT-4. BSM improves the evaluation correctness and consistency for each LLM by enhancing human-LLM agreement by up to 26{\%}, reducing length and pairwise position biases by up to 50{\%}, and allowing LLaMA-2-chat to match or outperform GPT-4 on most domains. On a constraint story generation task, BSM improves the coherence of stories while also improving constraint satisfaction by 12{\%}.", }

Large Language Models (LLMs) are frequently used for multi-faceted language generation and evaluation tasks that involve satisfying intricate user constraints or taking into account multiple aspects and criteria. However, their performance can fall short, due to the model{'}s lack of coherence and inability to plan and decompose the problem. We propose Branch-Solve-Merge (BSM), a Large Language Model program (Schlag et al., 2023) for tackling such challenging natural language tasks. It consists of branch, solve, and merge modules that are parameterized with specific prompts to the base LLM. These three modules plan a decomposition of the task into multiple parallel sub-tasks, independently solve them, and fuse the solutions to the sub-tasks. We apply our method to the tasks of LLM response evaluation and constrained text generation and evaluate its effectiveness with multiple LLMs, including Vicuna, LLaMA-2-chat, and GPT-4. BSM improves the evaluation correctness and consistency for each LLM by enhancing human-LLM agreement by up to 26{\%}, reducing length and pairwise position biases by up to 50{\%}, and allowing LLaMA-2-chat to match or outperform GPT-4 on most domains. On a constraint story generation task, BSM improves the coherence of stories while also improving constraint satisfaction by 12{\%}.

Branch-Solve-Merge Improves Large Language Model Evaluation and Generation

naacl-long.462

Poster

2310.15123

https://aclanthology.org/2024.naacl-long.463.bib

https://aclanthology.org/2024.naacl-long.463/

@inproceedings{shi-etal-2024-replug, title = "{REPLUG}: Retrieval-Augmented Black-Box Language Models", author = "Shi, Weijia and Min, Sewon and Yasunaga, Michihiro and Seo, Minjoon and James, Richard and Lewis, Mike and Zettlemoyer, Luke and Yih, Wen-tau", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.463", doi = "10.18653/v1/2024.naacl-long.463", pages = "8371--8384", abstract = "We introduce REPLUG, a retrieval-augmented language modeling framework that treats the language model (LM) as a black box and augments it with a tuneable retrieval model. Unlike prior retrieval-augmented LMs that train language models with special cross-attention mechanisms to encode the retrieved text, REPLUG simply prepends retrieved documents to the input for the frozen black-box LM. This simple design can be easily applied to any existing language models. Furthermore, we show that the LM can be used to supervise the retrieval model, which can then find documents that help the LM make better predictions. Our experiments demonstrate that REPLUG with the tuned retriever significantly improves the performance of GPT-3 (175B) on language modeling by 6.3{\%}, as well as the performance of Codex on five-shot MMLU by 5.1{\%}. Code is publicly released at github.com/swj0419/REPLUG.", }

We introduce REPLUG, a retrieval-augmented language modeling framework that treats the language model (LM) as a black box and augments it with a tuneable retrieval model. Unlike prior retrieval-augmented LMs that train language models with special cross-attention mechanisms to encode the retrieved text, REPLUG simply prepends retrieved documents to the input for the frozen black-box LM. This simple design can be easily applied to any existing language models. Furthermore, we show that the LM can be used to supervise the retrieval model, which can then find documents that help the LM make better predictions. Our experiments demonstrate that REPLUG with the tuned retriever significantly improves the performance of GPT-3 (175B) on language modeling by 6.3{\%}, as well as the performance of Codex on five-shot MMLU by 5.1{\%}. Code is publicly released at github.com/swj0419/REPLUG.

REPLUG: Retrieval-Augmented Black-Box Language Models

naacl-long.463

Oral

2301.12652

https://aclanthology.org/2024.naacl-long.464.bib

https://aclanthology.org/2024.naacl-long.464/

@inproceedings{han-etal-2024-david, title = "{D}avid helps Goliath: Inference-Time Collaboration Between Small Specialized and Large General Diffusion {LM}s", author = "Han, Xiaochuang and Kumar, Sachin and Tsvetkov, Yulia and Ghazvininejad, Marjan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.464", doi = "10.18653/v1/2024.naacl-long.464", pages = "8385--8400", abstract = "Diffusion-based language models are emerging as a promising alternative to autoregressive LMs: they approach the competence of autoregressive LMs while offering nuanced controllability at inference time. While autoregressive LMs have benefited immensely from scaling and instruction-based learning, existing studies of diffusion LMs have been conducted on a smaller scale. Starting with a recently proposed diffusion model SSD-LM, in this work we first explore methods to scale it from 0.4B to 13B parameters, proposing techniques to improve its training and inference efficiency, and to finetune the model to follow instructions. Armed with a more powerful, general purpose diffusion LM, we introduce the primary contribution of this work {--} SSD-2 {--} an approach to easily ensemble at inference time a large general-purpose diffusion LM with smaller, but specialized and contextualized diffusion LMs. We show that SSD-2 facilitates novel ensembles with 100x smaller models that can be customized and deployed by individual users. We find that compared to autoregressive models, the collaboration between diffusion LMs is more effective, leading to higher-quality model responses due to their ability to dynamically incorporate bi-directional contexts.", }

Diffusion-based language models are emerging as a promising alternative to autoregressive LMs: they approach the competence of autoregressive LMs while offering nuanced controllability at inference time. While autoregressive LMs have benefited immensely from scaling and instruction-based learning, existing studies of diffusion LMs have been conducted on a smaller scale. Starting with a recently proposed diffusion model SSD-LM, in this work we first explore methods to scale it from 0.4B to 13B parameters, proposing techniques to improve its training and inference efficiency, and to finetune the model to follow instructions. Armed with a more powerful, general purpose diffusion LM, we introduce the primary contribution of this work {--} SSD-2 {--} an approach to easily ensemble at inference time a large general-purpose diffusion LM with smaller, but specialized and contextualized diffusion LMs. We show that SSD-2 facilitates novel ensembles with 100x smaller models that can be customized and deployed by individual users. We find that compared to autoregressive models, the collaboration between diffusion LMs is more effective, leading to higher-quality model responses due to their ability to dynamically incorporate bi-directional contexts.

David helps Goliath: Inference-Time Collaboration Between Small Specialized and Large General Diffusion LMs

naacl-long.464

Oral

2305.14771

https://aclanthology.org/2024.naacl-long.465.bib

https://aclanthology.org/2024.naacl-long.465/

@inproceedings{zhu-etal-2024-efficient, title = "Efficient End-to-End Visual Document Understanding with Rationale Distillation", author = "Zhu, Wang and Agarwal, Alekh and Joshi, Mandar and Jia, Robin and Thomason, Jesse and Toutanova, Kristina", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.465", doi = "10.18653/v1/2024.naacl-long.465", pages = "8401--8424", abstract = "Understanding visually situated language requires interpreting complex layouts of textual and visual elements. Pre-processing tools, such as optical character recognition (OCR), can map document image inputs to textual tokens, then large language models (LLMs) can reason over text.However, such methods have high computational and engineering complexity. Can small pretrained image-to-text models accurately understand visual documents through similar recognition and reasoning steps instead?We propose Rationale Distillation (RD), which incorporates the outputs of OCR tools, LLMs, and larger multimodal models as intermediate {``}rationales{''}, and trains a small student model to predict both rationales and answers. On three visual document understanding benchmarks representing infographics, scanned documents, and figures, our Pix2Struct (282M parameters) student model finetuned with RD outperforms the base model by 4-5{\%} absolute accuracy with only 1{\%} higher computational cost.", }

Understanding visually situated language requires interpreting complex layouts of textual and visual elements. Pre-processing tools, such as optical character recognition (OCR), can map document image inputs to textual tokens, then large language models (LLMs) can reason over text.However, such methods have high computational and engineering complexity. Can small pretrained image-to-text models accurately understand visual documents through similar recognition and reasoning steps instead?We propose Rationale Distillation (RD), which incorporates the outputs of OCR tools, LLMs, and larger multimodal models as intermediate {``}rationales{''}, and trains a small student model to predict both rationales and answers. On three visual document understanding benchmarks representing infographics, scanned documents, and figures, our Pix2Struct (282M parameters) student model finetuned with RD outperforms the base model by 4-5{\%} absolute accuracy with only 1{\%} higher computational cost.

Efficient End-to-End Visual Document Understanding with Rationale Distillation

naacl-long.465

Poster

2311.09612

https://aclanthology.org/2024.naacl-long.466.bib

https://aclanthology.org/2024.naacl-long.466/

@inproceedings{eisape-etal-2024-systematic, title = "A Systematic Comparison of Syllogistic Reasoning in Humans and Language Models", author = "Eisape, Tiwalayo and Tessler, Michael and Dasgupta, Ishita and Sha, Fei and Steenkiste, Sjoerd and Linzen, Tal", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.466", doi = "10.18653/v1/2024.naacl-long.466", pages = "8425--8444", abstract = "A central component of rational behavior is logical inference: the process of determining which conclusions follow from a set of premises. Psychologists have documented several ways in which humans{'} inferences deviate from the rules of logic. Do language models, which are trained on text generated by humans, replicate such human biases, or are they able to overcome them? Focusing on the case of syllogisms{---}inferences from two simple premises{---}we show that, within the PaLM 2 family of transformer language models, larger models are more logical than smaller ones, and also more logical than humans. At the same time, even the largest models make systematic errors, some of which mirror human reasoning biases: they show sensitivity to the (irrelevant) ordering of the variables in the syllogism, and draw confident but incorrect inferences from particular syllogisms (syllogistic fallacies). Overall, we find that language models often mimic the human biases included in their training data, but are able to overcome them in some cases.", }

A central component of rational behavior is logical inference: the process of determining which conclusions follow from a set of premises. Psychologists have documented several ways in which humans{'} inferences deviate from the rules of logic. Do language models, which are trained on text generated by humans, replicate such human biases, or are they able to overcome them? Focusing on the case of syllogisms{---}inferences from two simple premises{---}we show that, within the PaLM 2 family of transformer language models, larger models are more logical than smaller ones, and also more logical than humans. At the same time, even the largest models make systematic errors, some of which mirror human reasoning biases: they show sensitivity to the (irrelevant) ordering of the variables in the syllogism, and draw confident but incorrect inferences from particular syllogisms (syllogistic fallacies). Overall, we find that language models often mimic the human biases included in their training data, but are able to overcome them in some cases.

A Systematic Comparison of Syllogistic Reasoning in Humans and Language Models

naacl-long.466

Oral

2311.00445

https://aclanthology.org/2024.naacl-long.467.bib

https://aclanthology.org/2024.naacl-long.467/

@inproceedings{lesci-vlachos-2024-anchoral, title = "{A}nchor{AL}: Computationally Efficient Active Learning for Large and Imbalanced Datasets", author = "Lesci, Pietro and Vlachos, Andreas", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.467", doi = "10.18653/v1/2024.naacl-long.467", pages = "8445--8464", abstract = "Active learning for imbalanced classification tasks is challenging as the minority classes naturally occur rarely. Gathering a large pool of unlabelled data is thus essential to capture minority instances. Standard pool-based active learning is computationally expensive on large pools and often reaches low accuracy by overfitting the initial decision boundary, thus failing to explore the input space and find minority instances. To address these issues we propose AnchorAL. At each iteration, AnchorAL chooses class-specific instances from the labelled set, or *anchors*, and retrieves the most similar unlabelled instances from the pool. This resulting *subpool* is then used for active learning. Using a small, fixed-sized subpool AnchorAL allows scaling any active learning strategy to large pools. By dynamically selecting different anchors at each iteration it promotes class balance and prevents overfitting the initial decision boundary, thus promoting the discovery of new clusters of minority instances. Experiments across different classification tasks, active learning strategies, and model architectures AnchorAL is *(i)* faster, often reducing runtime from hours to minutes, *(ii)* trains more performant models, *(iii)* and returns more balanced datasets than competing methods.", }

Active learning for imbalanced classification tasks is challenging as the minority classes naturally occur rarely. Gathering a large pool of unlabelled data is thus essential to capture minority instances. Standard pool-based active learning is computationally expensive on large pools and often reaches low accuracy by overfitting the initial decision boundary, thus failing to explore the input space and find minority instances. To address these issues we propose AnchorAL. At each iteration, AnchorAL chooses class-specific instances from the labelled set, or *anchors*, and retrieves the most similar unlabelled instances from the pool. This resulting *subpool* is then used for active learning. Using a small, fixed-sized subpool AnchorAL allows scaling any active learning strategy to large pools. By dynamically selecting different anchors at each iteration it promotes class balance and prevents overfitting the initial decision boundary, thus promoting the discovery of new clusters of minority instances. Experiments across different classification tasks, active learning strategies, and model architectures AnchorAL is *(i)* faster, often reducing runtime from hours to minutes, *(ii)* trains more performant models, *(iii)* and returns more balanced datasets than competing methods.

AnchorAL: Computationally Efficient Active Learning for Large and Imbalanced Datasets

naacl-long.467

Poster

2404.05623

https://huggingface.co/papers/2404.05623

https://aclanthology.org/2024.naacl-long.468.bib

https://aclanthology.org/2024.naacl-long.468/

@inproceedings{li-ng-2024-icle, title = "{ICLE}++: Modeling Fine-Grained Traits for Holistic Essay Scoring", author = "Li, Shengjie and Ng, Vincent", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.468", doi = "10.18653/v1/2024.naacl-long.468", pages = "8465--8486", abstract = "The majority of the recently developed models for automated essay scoring (AES) are evaluated solely on the ASAP corpus. However, ASAP is not without its limitations. For instance, it is not clear whether models trained on ASAP can generalize well when evaluated on other corpora. In light of these limitations, we introduce ICLE++, a corpus of persuasive student essays annotated with both holistic scores and trait-specific scores. Not only can ICLE++ be used to test the generalizability of AES models trained on ASAP, but it can also facilitate the evaluation of models developed for newer AES problems such as multi-trait scoring and cross-prompt scoring. We believe that ICLE++, which represents a culmination of our long-term effort in annotating the essays in the ICLE corpus, contributes to the set of much-needed annotated corpora for AES research.", }

The majority of the recently developed models for automated essay scoring (AES) are evaluated solely on the ASAP corpus. However, ASAP is not without its limitations. For instance, it is not clear whether models trained on ASAP can generalize well when evaluated on other corpora. In light of these limitations, we introduce ICLE++, a corpus of persuasive student essays annotated with both holistic scores and trait-specific scores. Not only can ICLE++ be used to test the generalizability of AES models trained on ASAP, but it can also facilitate the evaluation of models developed for newer AES problems such as multi-trait scoring and cross-prompt scoring. We believe that ICLE++, which represents a culmination of our long-term effort in annotating the essays in the ICLE corpus, contributes to the set of much-needed annotated corpora for AES research.

ICLE++: Modeling Fine-Grained Traits for Holistic Essay Scoring

naacl-long.468

Poster

https://aclanthology.org/2024.naacl-long.469.bib

https://aclanthology.org/2024.naacl-long.469/

@inproceedings{zhao-etal-2024-uncommonsense, title = "{UN}commonsense Reasoning: Abductive Reasoning about Uncommon Situations", author = "Zhao, Wenting and Chiu, Justin and Hwang, Jena and Brahman, Faeze and Hessel, Jack and Choudhury, Sanjiban and Choi, Yejin and Li, Xiang and Suhr, Alane", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.469", doi = "10.18653/v1/2024.naacl-long.469", pages = "8487--8505", abstract = "Language technologies that accurately model the dynamics of events must perform commonsense reasoning. Existing work evaluating commonsense reasoning focuses on making inferences about common, everyday situations. To instead investigate the ability to model unusual, unexpected, and unlikely situations, we explore the task of uncommonsense abductive reasoning. Given a piece of context with an unexpected outcome, this task requires reasoning abductively to generate an explanation that makes the unexpected outcome more likely in the context. To this end, we curate and release a new English language corpus called UNcommonsense. We characterize the performance differences between human explainers and the best-performing large language models, finding that model-enhanced human-written explanations achieve the highest quality by trading off between specificity and diversity. Finally, we experiment with several imitation learning algorithms to train open and accessible language models on this task. When compared with the vanilla supervised fine-tuning approach, these methods consistently reduce lose rates on both common and uncommonsense abductive reasoning judged by human evaluators.", }

Language technologies that accurately model the dynamics of events must perform commonsense reasoning. Existing work evaluating commonsense reasoning focuses on making inferences about common, everyday situations. To instead investigate the ability to model unusual, unexpected, and unlikely situations, we explore the task of uncommonsense abductive reasoning. Given a piece of context with an unexpected outcome, this task requires reasoning abductively to generate an explanation that makes the unexpected outcome more likely in the context. To this end, we curate and release a new English language corpus called UNcommonsense. We characterize the performance differences between human explainers and the best-performing large language models, finding that model-enhanced human-written explanations achieve the highest quality by trading off between specificity and diversity. Finally, we experiment with several imitation learning algorithms to train open and accessible language models on this task. When compared with the vanilla supervised fine-tuning approach, these methods consistently reduce lose rates on both common and uncommonsense abductive reasoning judged by human evaluators.

UNcommonsense Reasoning: Abductive Reasoning about Uncommon Situations

naacl-long.469

Poster

2311.08469

https://huggingface.co/papers/2311.08469

https://aclanthology.org/2024.naacl-long.470.bib

https://aclanthology.org/2024.naacl-long.470/

@inproceedings{hazra-majumder-2024-tell, title = "To Tell The Truth: Language of Deception and Language Models", author = "Hazra, Sanchaita and Majumder, Bodhisattwa Prasad", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.470", doi = "10.18653/v1/2024.naacl-long.470", pages = "8506--8520", abstract = "Text-based false information permeates online discourses, yet evidence of people{'}s ability to discern truth from such deceptive textual content is scarce. We analyze a novel TV game show data where conversations in a high-stake environment between individuals with conflicting objectives result in lies. We investigate the manifestation of potentially verifiable language cues of deception in the presence of objective truth, a distinguishing feature absent in previous text-based deception datasets. We show that there exists a class of detectors (algorithms) that have similar truth detection performance compared to human subjects, even when the former accesses only the language cues while the latter engages in conversations with complete access to all potential sources of cues (language and audio-visual). Our model, built on a large language model, employs a bottleneck framework to learn discernible cues to determine truth, an act of reasoning in which human subjects often perform poorly, even with incentives. Our model detects novel but accurate language cues in many cases where humans failed to detect deception, opening up the possibility of humans collaborating with algorithms and ameliorating their ability to detect the truth.", }

Text-based false information permeates online discourses, yet evidence of people{'}s ability to discern truth from such deceptive textual content is scarce. We analyze a novel TV game show data where conversations in a high-stake environment between individuals with conflicting objectives result in lies. We investigate the manifestation of potentially verifiable language cues of deception in the presence of objective truth, a distinguishing feature absent in previous text-based deception datasets. We show that there exists a class of detectors (algorithms) that have similar truth detection performance compared to human subjects, even when the former accesses only the language cues while the latter engages in conversations with complete access to all potential sources of cues (language and audio-visual). Our model, built on a large language model, employs a bottleneck framework to learn discernible cues to determine truth, an act of reasoning in which human subjects often perform poorly, even with incentives. Our model detects novel but accurate language cues in many cases where humans failed to detect deception, opening up the possibility of humans collaborating with algorithms and ameliorating their ability to detect the truth.

To Tell The Truth: Language of Deception and Language Models

naacl-long.470

Oral

2311.07092

https://aclanthology.org/2024.naacl-long.471.bib

https://aclanthology.org/2024.naacl-long.471/

@inproceedings{coto-solano-etal-2024-multilingual, title = "Multilingual Models for {ASR} in Chibchan Languages", author = "Coto-Solano, Rolando and Kim, Tai Wan and Jones, Alexander and Lo{\'a}iciga, Sharid", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.471", doi = "10.18653/v1/2024.naacl-long.471", pages = "8521--8535", abstract = "We present experiments on Automatic Speech Recognition (ASR) for Bribri and Cab{\'e}car, two languages from the Chibchan family. We fine-tune four ASR algorithms (Wav2Vec2, Whisper, MMS {\&} WavLM) to create monolingual models, with the Wav2Vec2 model demonstrating the best performance. We then proceed to use Wav2Vec2 for (1) experiments on training joint and transfer learning models for both languages, and (2) an analysis of the errors, with a focus on the transcription of tone. Results show effective transfer learning for both Bribri and Cab{\'e}car, but especially for Bribri. A post-processing spell checking step further reduced character and word error rates. As for the errors, tone is where the Bribri models make the most errors, whereas the simpler tonal system of Cab{\'e}car is better transcribed by the model. Our work contributes to developing better ASR technology, an important tool that could facilitate transcription, one of the major bottlenecks in language documentation efforts. Our work also assesses how existing pre-trained models and algorithms perform for genuine extremely low resource-languages.", }

We present experiments on Automatic Speech Recognition (ASR) for Bribri and Cab{\'e}car, two languages from the Chibchan family. We fine-tune four ASR algorithms (Wav2Vec2, Whisper, MMS {\&} WavLM) to create monolingual models, with the Wav2Vec2 model demonstrating the best performance. We then proceed to use Wav2Vec2 for (1) experiments on training joint and transfer learning models for both languages, and (2) an analysis of the errors, with a focus on the transcription of tone. Results show effective transfer learning for both Bribri and Cab{\'e}car, but especially for Bribri. A post-processing spell checking step further reduced character and word error rates. As for the errors, tone is where the Bribri models make the most errors, whereas the simpler tonal system of Cab{\'e}car is better transcribed by the model. Our work contributes to developing better ASR technology, an important tool that could facilitate transcription, one of the major bottlenecks in language documentation efforts. Our work also assesses how existing pre-trained models and algorithms perform for genuine extremely low resource-languages.

Multilingual Models for ASR in Chibchan Languages

naacl-long.471

Poster

https://aclanthology.org/2024.naacl-long.472.bib

https://aclanthology.org/2024.naacl-long.472/

@inproceedings{spangher-etal-2024-legaldiscourse, title = "{L}egal{D}iscourse: Interpreting When Laws Apply and To Whom", author = "Spangher, Alexander and Xue, Zihan and Wu, Te-Lin and Hansen, Mark and May, Jonathan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.472", doi = "10.18653/v1/2024.naacl-long.472", pages = "8536--8559", abstract = "While legal AI has made strides in recent years, it still struggles with basic legal concepts: {\_}when{\_} does a law apply? {\_}Who{\_} does it applies to? {\_}What{\_} does it do? We take a {\_}discourse{\_} approach to addressing these problems and introduce a novel taxonomy for span-and-relation parsing of legal texts. We create a dataset, {\_}LegalDiscourse{\_} of 602 state-level law paragraphs consisting of 3,715 discourse spans and 1,671 relations. Our trained annotators have an agreement-rate $\kappa>.8$, yet few-shot GPT3.5 performs poorly at span identification and relation classification. Although fine-tuning improves performance, GPT3.5 still lags far below human level. We demonstrate the usefulness of our schema by creating a web application with journalists. We collect over 100,000 laws for 52 U.S. states and territories using 20 scrapers we built, and apply our trained models to 6,000 laws using U.S. Census population numbers. We describe two journalistic outputs stemming from this application: (1) an investigation into the increase in liquor licenses following population growth and (2) a decrease in applicable laws under different under-count projections.", }

While legal AI has made strides in recent years, it still struggles with basic legal concepts: {\_}when{\_} does a law apply? {\_}Who{\_} does it applies to? {\_}What{\_} does it do? We take a {\_}discourse{\_} approach to addressing these problems and introduce a novel taxonomy for span-and-relation parsing of legal texts. We create a dataset, {\_}LegalDiscourse{\_} of 602 state-level law paragraphs consisting of 3,715 discourse spans and 1,671 relations. Our trained annotators have an agreement-rate $\kappa>.8$, yet few-shot GPT3.5 performs poorly at span identification and relation classification. Although fine-tuning improves performance, GPT3.5 still lags far below human level. We demonstrate the usefulness of our schema by creating a web application with journalists. We collect over 100,000 laws for 52 U.S. states and territories using 20 scrapers we built, and apply our trained models to 6,000 laws using U.S. Census population numbers. We describe two journalistic outputs stemming from this application: (1) an investigation into the increase in liquor licenses following population growth and (2) a decrease in applicable laws under different under-count projections.

LegalDiscourse: Interpreting When Laws Apply and To Whom

naacl-long.472

Oral

https://aclanthology.org/2024.naacl-long.473.bib

https://aclanthology.org/2024.naacl-long.473/

@inproceedings{liu-etal-2024-x, title = "{X}-Eval: Generalizable Multi-aspect Text Evaluation via Augmented Instruction Tuning with Auxiliary Evaluation Aspects", author = "Liu, Minqian and Shen, Ying and Xu, Zhiyang and Cao, Yixin and Cho, Eunah and Kumar, Vaibhav and Ghanadan, Reza and Huang, Lifu", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.473", doi = "10.18653/v1/2024.naacl-long.473", pages = "8560--8579", abstract = "Natural Language Generation (NLG) typically involves evaluating the generated text in various aspects (e.g., consistency and naturalness) to obtain a comprehensive assessment. However, multi-aspect evaluation remains challenging as it may require the evaluator to generalize to any given evaluation aspect even if it{'}s absent during training. In this paper, we introduce X-Eval, a two-stage instruction tuning framework to evaluate text in both seen and unseen aspects customized by end users. X-Eval consists of two learning stages: the vanilla instruction tuning stage that improves the model{'}s ability to follow evaluation instructions, and an enhanced instruction tuning stage that exploits the connections between fine-grained evaluation aspects to better assess text quality. To support the training of X-Eval, we collect AspectInstruct, the first instruction tuning dataset tailored for multi-aspect NLG evaluation spanning 27 diverse evaluation aspects with 65 tasks. To enhance task diversity, we devise an augmentation strategy that converts human rating annotations into diverse forms of NLG evaluation tasks, including scoring, comparison, ranking, and Boolean question answering. Extensive experiments across three essential categories of NLG tasks: dialogue generation, summarization, and data-to-text coupled with 21 aspects in meta-evaluation, demonstrate that X-Eval enables even a lightweight language model to achieve a comparable if not higher correlation with human judgments compared to the state-of-the-art NLG evaluators like GPT-4.", }

Natural Language Generation (NLG) typically involves evaluating the generated text in various aspects (e.g., consistency and naturalness) to obtain a comprehensive assessment. However, multi-aspect evaluation remains challenging as it may require the evaluator to generalize to any given evaluation aspect even if it{'}s absent during training. In this paper, we introduce X-Eval, a two-stage instruction tuning framework to evaluate text in both seen and unseen aspects customized by end users. X-Eval consists of two learning stages: the vanilla instruction tuning stage that improves the model{'}s ability to follow evaluation instructions, and an enhanced instruction tuning stage that exploits the connections between fine-grained evaluation aspects to better assess text quality. To support the training of X-Eval, we collect AspectInstruct, the first instruction tuning dataset tailored for multi-aspect NLG evaluation spanning 27 diverse evaluation aspects with 65 tasks. To enhance task diversity, we devise an augmentation strategy that converts human rating annotations into diverse forms of NLG evaluation tasks, including scoring, comparison, ranking, and Boolean question answering. Extensive experiments across three essential categories of NLG tasks: dialogue generation, summarization, and data-to-text coupled with 21 aspects in meta-evaluation, demonstrate that X-Eval enables even a lightweight language model to achieve a comparable if not higher correlation with human judgments compared to the state-of-the-art NLG evaluators like GPT-4.

X-Eval: Generalizable Multi-aspect Text Evaluation via Augmented Instruction Tuning with Auxiliary Evaluation Aspects

naacl-long.473

Poster

2311.08788

https://aclanthology.org/2024.naacl-long.474.bib

https://aclanthology.org/2024.naacl-long.474/

@inproceedings{sheng-etal-2024-reference, title = "Is Reference Necessary in the Evaluation of {NLG} Systems? When and Where?", author = "Sheng, Shuqian and Xu, Yi and Fu, Luoyi and Ding, Jiaxin and Zhou, Lei and Wang, Xinbing and Zhou, Chenghu", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.474", doi = "10.18653/v1/2024.naacl-long.474", pages = "8580--8596", abstract = "The majority of automatic metrics for evaluating NLG systems are reference-based. However, the challenge of collecting human annotation results in a lack of reliable references in numerous application scenarios. Despite recent advancements in reference-free metrics, it has not been well understood when and where they can be used as an alternative to reference-based metrics. In this study, by employing diverse analytical approaches, we comprehensively assess the performance of both metrics across a wide range of NLG tasks, encompassing eight datasets and eight evaluation models. Based on solid experiments, the results show that reference-free metrics exhibit a higher correlation with human judgment and greater sensitivity to deficiencies in language quality. However, their effectiveness varies across tasks and is influenced by the quality of candidate texts. Therefore, it{'}s important to assess the performance of reference-free metrics before applying them to a new task, especially when inputs are in uncommon form or when the answer space is highly variable. Our study can provide insight into the appropriate application of automatic metrics and the impact of metric choice on evaluation performance.", }

The majority of automatic metrics for evaluating NLG systems are reference-based. However, the challenge of collecting human annotation results in a lack of reliable references in numerous application scenarios. Despite recent advancements in reference-free metrics, it has not been well understood when and where they can be used as an alternative to reference-based metrics. In this study, by employing diverse analytical approaches, we comprehensively assess the performance of both metrics across a wide range of NLG tasks, encompassing eight datasets and eight evaluation models. Based on solid experiments, the results show that reference-free metrics exhibit a higher correlation with human judgment and greater sensitivity to deficiencies in language quality. However, their effectiveness varies across tasks and is influenced by the quality of candidate texts. Therefore, it{'}s important to assess the performance of reference-free metrics before applying them to a new task, especially when inputs are in uncommon form or when the answer space is highly variable. Our study can provide insight into the appropriate application of automatic metrics and the impact of metric choice on evaluation performance.

Is Reference Necessary in the Evaluation of NLG Systems? When and Where?

naacl-long.474

Poster

2403.14275

https://aclanthology.org/2024.naacl-long.475.bib

https://aclanthology.org/2024.naacl-long.475/

@inproceedings{su-etal-2024-semi, title = "Semi-Structured Chain-of-Thought: Integrating Multiple Sources of Knowledge for Improved Language Model Reasoning", author = "Su, Xin and Le, Tiep and Bethard, Steven and Howard, Phillip", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.475", doi = "10.18653/v1/2024.naacl-long.475", pages = "8597--8613", abstract = "An important open question in the use of large language models for knowledge-intensive tasks is how to effectively integrate knowledge from three sources: the model{'}s parametric memory, external structured knowledge, and external unstructured knowledge. Most existing prompting methods either rely on one or two of these sources, or require repeatedly invoking large language models to generate similar or identical content. In this work, we overcome these limitations by introducing a novel semi-structured prompting approach that seamlessly integrates the model{'}s parametric memory with unstructured knowledge from text documents and structured knowledge from knowledge graphs. Experimental results on open-domain multi-hop question answering datasets demonstrate that our prompting method significantly surpasses existing techniques, even exceeding those that require fine-tuning.", }

An important open question in the use of large language models for knowledge-intensive tasks is how to effectively integrate knowledge from three sources: the model{'}s parametric memory, external structured knowledge, and external unstructured knowledge. Most existing prompting methods either rely on one or two of these sources, or require repeatedly invoking large language models to generate similar or identical content. In this work, we overcome these limitations by introducing a novel semi-structured prompting approach that seamlessly integrates the model{'}s parametric memory with unstructured knowledge from text documents and structured knowledge from knowledge graphs. Experimental results on open-domain multi-hop question answering datasets demonstrate that our prompting method significantly surpasses existing techniques, even exceeding those that require fine-tuning.

Semi-Structured Chain-of-Thought: Integrating Multiple Sources of Knowledge for Improved Language Model Reasoning

naacl-long.475

Poster

2311.08505

https://aclanthology.org/2024.naacl-long.476.bib

https://aclanthology.org/2024.naacl-long.476/

@inproceedings{seals-shalin-2024-evaluating, title = "Evaluating the Deductive Competence of Large Language Models", author = "Seals, S and Shalin, Valerie", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.476", doi = "10.18653/v1/2024.naacl-long.476", pages = "8614--8630", abstract = "The development of highly fluent large language models (LLMs) has prompted increased interest in assessing their reasoning and problem-solving capabilities. We investigate whether several LLMs can solve a classic type of deductive reasoning problem from the cognitive science literature. The tested LLMs have limited abilities to solve these problems in their conventional form. We performed follow up experiments to investigate if changes to the presentation format and content improve model performance. We do find performance differences between conditions; however, they do not improve overall performance. Moreover, we find that performance interacts with presentation format and content in unexpected ways that differ from human performance. Overall, our results suggest that LLMs have unique reasoning biases that are only partially predicted from human reasoning performance and the human-generated language corpora that informs them.", }

The development of highly fluent large language models (LLMs) has prompted increased interest in assessing their reasoning and problem-solving capabilities. We investigate whether several LLMs can solve a classic type of deductive reasoning problem from the cognitive science literature. The tested LLMs have limited abilities to solve these problems in their conventional form. We performed follow up experiments to investigate if changes to the presentation format and content improve model performance. We do find performance differences between conditions; however, they do not improve overall performance. Moreover, we find that performance interacts with presentation format and content in unexpected ways that differ from human performance. Overall, our results suggest that LLMs have unique reasoning biases that are only partially predicted from human reasoning performance and the human-generated language corpora that informs them.

Evaluating the Deductive Competence of Large Language Models

naacl-long.476

Poster

2309.05452

https://aclanthology.org/2024.naacl-long.477.bib

https://aclanthology.org/2024.naacl-long.477/

@inproceedings{soni-etal-2024-large, title = "Large Human Language Models: A Need and the Challenges", author = "Soni, Nikita and Schwartz, H. and Sedoc, Jo{\~a}o and Balasubramanian, Niranjan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.477", doi = "10.18653/v1/2024.naacl-long.477", pages = "8631--8646", abstract = "As research in human-centered NLP advances, there is a growing recognition of the importance of incorporating human and social factors into NLP models. At the same time, our NLP systems have become heavily reliant on LLMs, most of which do not model authors. To build NLP systems that can truly understand human language, we must better integrate human contexts into LLMs. This brings to the fore a range of design considerations and challenges in terms of what human aspects to capture, how to represent them, and what modeling strategies to pursue. To address these, we advocate for three positions toward creating large human language models (LHLMs) using concepts from psychological and behavioral sciences: First, LM training should include the human context. Second, LHLMs should recognize that people are more than their group(s). Third, LHLMs should be able to account for the dynamic and temporally-dependent nature of the human context. We refer to relevant advances and present open challenges that need to be addressed and their possible solutions in realizing these goals.", }

As research in human-centered NLP advances, there is a growing recognition of the importance of incorporating human and social factors into NLP models. At the same time, our NLP systems have become heavily reliant on LLMs, most of which do not model authors. To build NLP systems that can truly understand human language, we must better integrate human contexts into LLMs. This brings to the fore a range of design considerations and challenges in terms of what human aspects to capture, how to represent them, and what modeling strategies to pursue. To address these, we advocate for three positions toward creating large human language models (LHLMs) using concepts from psychological and behavioral sciences: First, LM training should include the human context. Second, LHLMs should recognize that people are more than their group(s). Third, LHLMs should be able to account for the dynamic and temporally-dependent nature of the human context. We refer to relevant advances and present open challenges that need to be addressed and their possible solutions in realizing these goals.

Large Human Language Models: A Need and the Challenges

naacl-long.477

Oral

2312.07751

https://aclanthology.org/2024.naacl-long.478.bib

https://aclanthology.org/2024.naacl-long.478/

@inproceedings{liu-etal-2024-learning, title = "On Learning to Summarize with Large Language Models as References", author = "Liu, Yixin and Shi, Kejian and He, Katherine and Ye, Longtian and Fabbri, Alexander and Liu, Pengfei and Radev, Dragomir and Cohan, Arman", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.478", doi = "10.18653/v1/2024.naacl-long.478", pages = "8647--8664", abstract = "Recent studies have found that summaries generated by large language models (LLMs) are favored by human annotators over the original reference summaries in commonly used summarization datasets. Therefore, we study an LLM-as-reference learning setting for smaller text summarization models to investigate whether their performance can be substantially improved. To this end, we use LLMs as both oracle summary generators for standard supervised fine-tuning and oracle summary evaluators for efficient contrastive learning that leverages the LLMs{'} supervision signals. We conduct comprehensive experiments with source news articles and find that (1) summarization models trained under the LLM-as-reference setting achieve significant performance improvement in both LLM and human evaluations; (2) contrastive learning outperforms standard supervised fine-tuning under both low and high resource settings. Our experimental results also enable a meta-analysis of LLMs{'} summary evaluation capacities under a challenging setting, showing that LLMs are not well-aligned with human evaluators. Particularly, our expert human evaluation reveals remaining nuanced performance gaps between LLMs and our fine-tuned models, which LLMs fail to capture. Thus, we call for further studies into both the potential and challenges of using LLMs in summarization model development.", }

Recent studies have found that summaries generated by large language models (LLMs) are favored by human annotators over the original reference summaries in commonly used summarization datasets. Therefore, we study an LLM-as-reference learning setting for smaller text summarization models to investigate whether their performance can be substantially improved. To this end, we use LLMs as both oracle summary generators for standard supervised fine-tuning and oracle summary evaluators for efficient contrastive learning that leverages the LLMs{'} supervision signals. We conduct comprehensive experiments with source news articles and find that (1) summarization models trained under the LLM-as-reference setting achieve significant performance improvement in both LLM and human evaluations; (2) contrastive learning outperforms standard supervised fine-tuning under both low and high resource settings. Our experimental results also enable a meta-analysis of LLMs{'} summary evaluation capacities under a challenging setting, showing that LLMs are not well-aligned with human evaluators. Particularly, our expert human evaluation reveals remaining nuanced performance gaps between LLMs and our fine-tuned models, which LLMs fail to capture. Thus, we call for further studies into both the potential and challenges of using LLMs in summarization model development.

On Learning to Summarize with Large Language Models as References

naacl-long.478

Poster

2305.14239

https://aclanthology.org/2024.naacl-long.479.bib

https://aclanthology.org/2024.naacl-long.479/

@inproceedings{xia-etal-2024-hallucination, title = "Hallucination Diversity-Aware Active Learning for Text Summarization", author = "Xia, Yu and Liu, Xu and Yu, Tong and Kim, Sungchul and Rossi, Ryan and Rao, Anup and Mai, Tung and Li, Shuai", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.479", doi = "10.18653/v1/2024.naacl-long.479", pages = "8665--8677", abstract = "Large Language Models (LLMs) have shown propensity to generate hallucinated outputs, i.e., texts that are factually incorrect or unsupported. Existing methods for alleviating hallucinations typically require costly human annotations to identify and correct hallucinations in LLM outputs. Moreover, most of these methods focus on a specific type of hallucination, e.g., entity or token errors, which limits their effectiveness in addressing various types of hallucinations exhibited in LLM outputs. To our best knowledge, in this paper we propose the first active learning framework to alleviate LLM hallucinations, reducing costly human annotations of hallucination needed. By measuring fine-grained hallucinations from errors in semantic frame, discourse and content verifiability in text summarization, we propose HAllucination Diversity-Aware Sampling (HADAS) to select diverse hallucinations for annotations in active learning for LLM finetuning. Extensive experiments on three datasets and different backbone models demonstrate advantages of our method in effectively and efficiently mitigating LLM hallucinations.", }

Large Language Models (LLMs) have shown propensity to generate hallucinated outputs, i.e., texts that are factually incorrect or unsupported. Existing methods for alleviating hallucinations typically require costly human annotations to identify and correct hallucinations in LLM outputs. Moreover, most of these methods focus on a specific type of hallucination, e.g., entity or token errors, which limits their effectiveness in addressing various types of hallucinations exhibited in LLM outputs. To our best knowledge, in this paper we propose the first active learning framework to alleviate LLM hallucinations, reducing costly human annotations of hallucination needed. By measuring fine-grained hallucinations from errors in semantic frame, discourse and content verifiability in text summarization, we propose HAllucination Diversity-Aware Sampling (HADAS) to select diverse hallucinations for annotations in active learning for LLM finetuning. Extensive experiments on three datasets and different backbone models demonstrate advantages of our method in effectively and efficiently mitigating LLM hallucinations.

Hallucination Diversity-Aware Active Learning for Text Summarization

naacl-long.479

Poster

2404.01588

https://aclanthology.org/2024.naacl-long.480.bib

https://aclanthology.org/2024.naacl-long.480/

@inproceedings{bao-carpuat-2024-keep, title = "{K}eep it {P}rivate: Unsupervised Privatization of Online Text", author = "Bao, Calvin and Carpuat, Marine", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.480", doi = "10.18653/v1/2024.naacl-long.480", pages = "8678--8693", abstract = "Authorship obfuscation techniques hold the promise of helping people protect their privacy in online communications by automatically rewriting text to hide the identity of the original author. However, obfuscation has been evaluated in narrow settings in the NLP literature and has primarily been addressed with superficial edit operations that can lead to unnatural outputs. In this work, we introduce an automatic text privatization framework that fine-tunes a large language model via reinforcement learning to produce rewrites that balance soundness, sense, and privacy. We evaluate it extensively on a large-scale test set of English Reddit posts by 68k authors composed of short-medium length texts. We study how the performance changes among evaluative conditions including authorial profile length and authorship detection strategy. Our method maintains high text quality according to both automated metrics and human evaluation, and successfully evades several automated authorship attacks.", }

Authorship obfuscation techniques hold the promise of helping people protect their privacy in online communications by automatically rewriting text to hide the identity of the original author. However, obfuscation has been evaluated in narrow settings in the NLP literature and has primarily been addressed with superficial edit operations that can lead to unnatural outputs. In this work, we introduce an automatic text privatization framework that fine-tunes a large language model via reinforcement learning to produce rewrites that balance soundness, sense, and privacy. We evaluate it extensively on a large-scale test set of English Reddit posts by 68k authors composed of short-medium length texts. We study how the performance changes among evaluative conditions including authorial profile length and authorship detection strategy. Our method maintains high text quality according to both automated metrics and human evaluation, and successfully evades several automated authorship attacks.

Keep it Private: Unsupervised Privatization of Online Text

naacl-long.480

Poster

2405.10260

https://aclanthology.org/2024.naacl-long.481.bib

https://aclanthology.org/2024.naacl-long.481/

@inproceedings{renduchintala-etal-2024-tied, title = "Tied-{L}o{RA}: Enhancing parameter efficiency of {L}o{RA} with Weight Tying", author = "Renduchintala, Adithya and Konuk, Tugrul and Kuchaiev, Oleksii", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.481", doi = "10.18653/v1/2024.naacl-long.481", pages = "8694--8705", abstract = "We introduce Tied-LoRA, a novel paradigm leveraging weight tying and selective training to enhance the parameter efficiency of Low-rank Adaptation (LoRA). Our exploration encompasses different plausible combinations of parameter training and freezing, coupled with weight tying, aimed at identifying the optimal trade-off between performance and the count of trainable parameters. Across 5 diverse tasks and two foundational language models with different parameter counts, our experiments provide comprehensive insights into the inherent trade-offs between efficiency and performance.Our findings reveal a specific Tied-LoRA configuration that distinguishes itself by showcasing comparable performance to LoRA across multiple tasks while utilizing only a fraction of the parameters employed by the standard LoRA method, particularly at elevated ranks. This underscores the efficacy of Tied-LoRA in achieving impressive results with significantly reduced model complexity.", }

We introduce Tied-LoRA, a novel paradigm leveraging weight tying and selective training to enhance the parameter efficiency of Low-rank Adaptation (LoRA). Our exploration encompasses different plausible combinations of parameter training and freezing, coupled with weight tying, aimed at identifying the optimal trade-off between performance and the count of trainable parameters. Across 5 diverse tasks and two foundational language models with different parameter counts, our experiments provide comprehensive insights into the inherent trade-offs between efficiency and performance.Our findings reveal a specific Tied-LoRA configuration that distinguishes itself by showcasing comparable performance to LoRA across multiple tasks while utilizing only a fraction of the parameters employed by the standard LoRA method, particularly at elevated ranks. This underscores the efficacy of Tied-LoRA in achieving impressive results with significantly reduced model complexity.

Tied-LoRA: Enhancing parameter efficiency of LoRA with Weight Tying

naacl-long.481

Poster

2311.09578

https://huggingface.co/papers/2311.09578

https://aclanthology.org/2024.naacl-long.482.bib

https://aclanthology.org/2024.naacl-long.482/

@inproceedings{deng-etal-2024-investigating, title = "Investigating Data Contamination in Modern Benchmarks for Large Language Models", author = "Deng, Chunyuan and Zhao, Yilun and Tang, Xiangru and Gerstein, Mark and Cohan, Arman", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.482", doi = "10.18653/v1/2024.naacl-long.482", pages = "8706--8719", abstract = "Recent observations have underscored a disparity between the inflated benchmark scores and the actual performance of LLMs, raising concerns about potential contamination of evaluation benchmarks. This issue is especially critical for closed-source models and certain open-source models where training data transparency is lacking. In this paper we study data contamination by proposing two methods tailored for both open-source and proprietary LLMs. We first introduce a retrieval-based system to explore potential overlaps between evaluation benchmarks and pretraining corpora. We further present a novel investigation protocol named Testset Slot Guessing (TS-Guessing), applicable to both open and proprietary models. This approach entails masking a wrong answer in a multiple-choice question and prompting the model to fill in the gap. Additionally, it involves obscuring an unlikely word in an evaluation example and asking the model to produce it. We find that certain commercial LLMs could surprisingly guess the missing option in various test sets. Specifically, in the MMLU benchmark, ChatGPT and GPT-4 demonstrated an exact match rate of 52{\%} and 57{\%}, respectively, in guessing the missing options in benchmark test data. We hope these results underscore the need for more robust evaluation methodologies and benchmarks in the field.", }

Recent observations have underscored a disparity between the inflated benchmark scores and the actual performance of LLMs, raising concerns about potential contamination of evaluation benchmarks. This issue is especially critical for closed-source models and certain open-source models where training data transparency is lacking. In this paper we study data contamination by proposing two methods tailored for both open-source and proprietary LLMs. We first introduce a retrieval-based system to explore potential overlaps between evaluation benchmarks and pretraining corpora. We further present a novel investigation protocol named Testset Slot Guessing (TS-Guessing), applicable to both open and proprietary models. This approach entails masking a wrong answer in a multiple-choice question and prompting the model to fill in the gap. Additionally, it involves obscuring an unlikely word in an evaluation example and asking the model to produce it. We find that certain commercial LLMs could surprisingly guess the missing option in various test sets. Specifically, in the MMLU benchmark, ChatGPT and GPT-4 demonstrated an exact match rate of 52{\%} and 57{\%}, respectively, in guessing the missing options in benchmark test data. We hope these results underscore the need for more robust evaluation methodologies and benchmarks in the field.

Investigating Data Contamination in Modern Benchmarks for Large Language Models

naacl-long.482

Poster

2311.09783

https://aclanthology.org/2024.naacl-long.483.bib

https://aclanthology.org/2024.naacl-long.483/

@inproceedings{wang-zhang-2024-pre, title = "Pre-trained Language Models for Entity Blocking: A Reproducibility Study", author = "Wang, Runhui and Zhang, Yongfeng", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.483", doi = "10.18653/v1/2024.naacl-long.483", pages = "8720--8730", abstract = "Entity Resolution (ER) is an essential task in data integration and its goal is to find records that represent the same entity in a dataset. Deep learning models, especially large pre-trained language models, have achieved state-of-the-art results on this task. A typical ER pipeline consists of Entity Blocking and Entity Matching: Entity Blocking finds candidate record pairs that potentially match and Entity Matching determines if the pairs match. The goal of the entity blocking step is to include as many matching pairs as possible while including as few non-matching pairs as possible. On the other hand, the blocking task can also be considered as an Information Retrieval (IR) task. However, state-of-the-art neural IR models that are based on large language models have not been evaluated on the ER task. What{'}s more, the generalization ability of state-of-the-art methods for entity blocking is not well-studied but an import aspect in real-world applications. In this work, we evaluate state-of-the-art models for Entity Blocking along with neural IR models on a wide range of real-world datasets, and also study their in-distribution and out-of-distribution generalization abilities.", }

Entity Resolution (ER) is an essential task in data integration and its goal is to find records that represent the same entity in a dataset. Deep learning models, especially large pre-trained language models, have achieved state-of-the-art results on this task. A typical ER pipeline consists of Entity Blocking and Entity Matching: Entity Blocking finds candidate record pairs that potentially match and Entity Matching determines if the pairs match. The goal of the entity blocking step is to include as many matching pairs as possible while including as few non-matching pairs as possible. On the other hand, the blocking task can also be considered as an Information Retrieval (IR) task. However, state-of-the-art neural IR models that are based on large language models have not been evaluated on the ER task. What{'}s more, the generalization ability of state-of-the-art methods for entity blocking is not well-studied but an import aspect in real-world applications. In this work, we evaluate state-of-the-art models for Entity Blocking along with neural IR models on a wide range of real-world datasets, and also study their in-distribution and out-of-distribution generalization abilities.

Pre-trained Language Models for Entity Blocking: A Reproducibility Study

naacl-long.483

Poster

https://aclanthology.org/2024.naacl-long.484.bib

https://aclanthology.org/2024.naacl-long.484/

@inproceedings{ramu-etal-2024-re2, title = "$RE^2$: Region-Aware Relation Extraction from Visually Rich Documents", author = "Ramu, Pritika and Wang, Sijia and Mouatadid, Lalla and Rimchala, Joy and Huang, Lifu", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.484", doi = "10.18653/v1/2024.naacl-long.484", pages = "8731--8747", abstract = "Current research in form understanding predominantly relies on large pre-trained language models, necessitating extensive data for pre-training. However, the importance of layout structure (i.e., the spatial relationship between the entity blocks in the visually rich document) to relation extraction has been overlooked. In this paper, we propose $\textbf{RE}$gion-Aware $\textbf{R}$elation $\textbf{E}$xtraction ($\bf{RE^2}$) that leverages region-level spatial structure among the entity blocks to improve their relation prediction. We design an edge-aware graph attention network to learn the interaction between entities while considering their spatial relationship defined by their region-level representations. We also introduce a constraint objective to regularize the model towards consistency with the inherent constraints of the relation extraction task. To support the research on relation extraction from visually rich documents and demonstrate the generalizability of $\bf{RE^2}$, we build a new benchmark dataset, ${DiverseForm}$, that covers a wide range of domains. Extensive experiments on ${DiverseForm}$ and several public benchmark datasets demonstrate significant superiority and transferability of $\bf{RE^2}$ across various domains and languages, with up to 18.88{\%} absolute F-score gain over all high-performing baselines", }

Current research in form understanding predominantly relies on large pre-trained language models, necessitating extensive data for pre-training. However, the importance of layout structure (i.e., the spatial relationship between the entity blocks in the visually rich document) to relation extraction has been overlooked. In this paper, we propose $\textbf{RE}$gion-Aware $\textbf{R}$elation $\textbf{E}$xtraction ($\bf{RE^2}$) that leverages region-level spatial structure among the entity blocks to improve their relation prediction. We design an edge-aware graph attention network to learn the interaction between entities while considering their spatial relationship defined by their region-level representations. We also introduce a constraint objective to regularize the model towards consistency with the inherent constraints of the relation extraction task. To support the research on relation extraction from visually rich documents and demonstrate the generalizability of $\bf{RE^2}$, we build a new benchmark dataset, ${DiverseForm}$, that covers a wide range of domains. Extensive experiments on ${DiverseForm}$ and several public benchmark datasets demonstrate significant superiority and transferability of $\bf{RE^2}$ across various domains and languages, with up to 18.88{\%} absolute F-score gain over all high-performing baselines

RE^2: Region-Aware Relation Extraction from Visually Rich Documents

naacl-long.484

Poster

https://aclanthology.org/2024.naacl-long.485.bib

https://aclanthology.org/2024.naacl-long.485/

@inproceedings{nie-etal-2024-mix, title = "Mix-Initiative Response Generation with Dynamic Prefix Tuning", author = "Nie, Yuxiang and Huang, Heyan and Mao, Xian-Ling and Liao, Lizi", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.485", doi = "10.18653/v1/2024.naacl-long.485", pages = "8748--8761", abstract = "Mixed initiative serves as one of the key factors in controlling conversation directions. For a speaker, responding passively or leading proactively would result in rather different responses. However, most dialogue systems focus on training a holistic response generation model without any distinction among different initiatives. It leads to the cross-contamination problem, where the model confuses different initiatives and generates inappropriate responses. Moreover, obtaining plenty of human annotations for initiative labels can be expensive. To address this issue, we propose a general mix-Initiative Dynamic Prefix Tuning framework (IDPT) to decouple different initiatives from the generation model, which learns initiative-aware prefixes in both supervised and unsupervised settings. Specifically, IDPT decouples initiative factors into different prefix parameters and uses the attention mechanism to adjust the selection of initiatives in guiding generation dynamically. The prefix parameters can be tuned towards accurate initiative prediction as well as mix-initiative response generation. Extensive experiments on two public dialogue datasets show that the proposed IDPT outperforms previous baselines on both automatic metrics and human evaluations. It also manages to generate appropriate responses with manipulated initiatives.", }

Mixed initiative serves as one of the key factors in controlling conversation directions. For a speaker, responding passively or leading proactively would result in rather different responses. However, most dialogue systems focus on training a holistic response generation model without any distinction among different initiatives. It leads to the cross-contamination problem, where the model confuses different initiatives and generates inappropriate responses. Moreover, obtaining plenty of human annotations for initiative labels can be expensive. To address this issue, we propose a general mix-Initiative Dynamic Prefix Tuning framework (IDPT) to decouple different initiatives from the generation model, which learns initiative-aware prefixes in both supervised and unsupervised settings. Specifically, IDPT decouples initiative factors into different prefix parameters and uses the attention mechanism to adjust the selection of initiatives in guiding generation dynamically. The prefix parameters can be tuned towards accurate initiative prediction as well as mix-initiative response generation. Extensive experiments on two public dialogue datasets show that the proposed IDPT outperforms previous baselines on both automatic metrics and human evaluations. It also manages to generate appropriate responses with manipulated initiatives.

Mix-Initiative Response Generation with Dynamic Prefix Tuning

naacl-long.485

Poster

2403.17636

https://aclanthology.org/2024.naacl-long.486.bib

https://aclanthology.org/2024.naacl-long.486/

@inproceedings{yao-etal-2024-value, title = "Value {FULCRA}: Mapping Large Language Models to the Multidimensional Spectrum of Basic Human Value", author = "Yao, Jing and Yi, Xiaoyuan and Gong, Yifan and Wang, Xiting and Xie, Xing", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.486", doi = "10.18653/v1/2024.naacl-long.486", pages = "8762--8785", abstract = "Value alignment is crucial for the responsible development of Large Language Models (LLMs). However, how to define values in this context remains largely unexplored. Existing work mainly specifies values as risk criteria formulated in the AI community, e.g., fairness and privacy protection, suffering from poor clarity, adaptability and transparency. Leveraging basic values established in humanity and social science that are compatible with values across cultures, this paper introduces a novel value space spanned by multiple basic value dimensions and proposes BaseAlign, a corresponding value alignment paradigm. Applying the representative Schwartz{'}s Theory of Basic Values as an instantiation, we construct FULCRA, a dataset consisting of 20k $($LLM output, value vector$)$ pairs. LLMs{'} outputs are mapped into the $K$-dim value space beyond simple binary labels, by identifying their underlying priorities for these value dimensions. Extensive analysis and experiments on FULCRA: (1) reveal the essential relation between basic values and LLMs{'} behaviors, (2) demonstrate that our paradigm with basic values not only covers existing risks but also anticipates the unidentified ones, and (3) manifest BaseAlign{'}s superiority in alignment performance with less data, paving the way for addressing the above three challenges.", }

Value alignment is crucial for the responsible development of Large Language Models (LLMs). However, how to define values in this context remains largely unexplored. Existing work mainly specifies values as risk criteria formulated in the AI community, e.g., fairness and privacy protection, suffering from poor clarity, adaptability and transparency. Leveraging basic values established in humanity and social science that are compatible with values across cultures, this paper introduces a novel value space spanned by multiple basic value dimensions and proposes BaseAlign, a corresponding value alignment paradigm. Applying the representative Schwartz{'}s Theory of Basic Values as an instantiation, we construct FULCRA, a dataset consisting of 20k $($LLM output, value vector$)$ pairs. LLMs{'} outputs are mapped into the $K$-dim value space beyond simple binary labels, by identifying their underlying priorities for these value dimensions. Extensive analysis and experiments on FULCRA: (1) reveal the essential relation between basic values and LLMs{'} behaviors, (2) demonstrate that our paradigm with basic values not only covers existing risks but also anticipates the unidentified ones, and (3) manifest BaseAlign{'}s superiority in alignment performance with less data, paving the way for addressing the above three challenges.

Value FULCRA: Mapping Large Language Models to the Multidimensional Spectrum of Basic Human Value

naacl-long.486

Poster

https://aclanthology.org/2024.naacl-long.487.bib

https://aclanthology.org/2024.naacl-long.487/

@inproceedings{sahoo-etal-2024-indibias, title = "{I}ndi{B}ias: A Benchmark Dataset to Measure Social Biases in Language Models for {I}ndian Context", author = "Sahoo, Nihar and Kulkarni, Pranamya and Ahmad, Arif and Goyal, Tanu and Asad, Narjis and Garimella, Aparna and Bhattacharyya, Pushpak", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-long.487", doi = "10.18653/v1/2024.naacl-long.487", pages = "8786--8806", abstract = "The pervasive influence of social biases in language data has sparked the need for benchmark datasets that capture and evaluate these biases in Large Language Models (LLMs). Existing efforts predominantly focus on English language and the Western context, leaving a void for a reliable dataset that encapsulates India{'}s unique socio-cultural nuances. To bridge this gap, we introduce IndiBias, a comprehensive benchmarking dataset designed specifically for evaluating social biases in the Indian context. We filter and translate the existing CrowS-Pairs dataset to create a benchmark dataset suited to the Indian context in Hindi language. Additionally, we leverage LLMs including ChatGPT and InstructGPT to augment our dataset with diverse societal biases and stereotypes prevalent in India. The included bias dimensions encompass gender, religion, caste, age, region, physical appearance, and occupation. We also build a resource to address intersectional biases along three intersectional dimensions. Our dataset contains 800 sentence pairs and 300 tuples for bias measurement across different demographics. The dataset is available in English and Hindi, providing a size comparable to existing benchmark datasets. Furthermore, using IndiBias we compare ten different language models on multiple bias measurement metrics. We observed that the language models exhibit more bias across a majority of the intersectional groups. All the scripts utilized and datasets created in this study are publicly available.", }

The pervasive influence of social biases in language data has sparked the need for benchmark datasets that capture and evaluate these biases in Large Language Models (LLMs). Existing efforts predominantly focus on English language and the Western context, leaving a void for a reliable dataset that encapsulates India{'}s unique socio-cultural nuances. To bridge this gap, we introduce IndiBias, a comprehensive benchmarking dataset designed specifically for evaluating social biases in the Indian context. We filter and translate the existing CrowS-Pairs dataset to create a benchmark dataset suited to the Indian context in Hindi language. Additionally, we leverage LLMs including ChatGPT and InstructGPT to augment our dataset with diverse societal biases and stereotypes prevalent in India. The included bias dimensions encompass gender, religion, caste, age, region, physical appearance, and occupation. We also build a resource to address intersectional biases along three intersectional dimensions. Our dataset contains 800 sentence pairs and 300 tuples for bias measurement across different demographics. The dataset is available in English and Hindi, providing a size comparable to existing benchmark datasets. Furthermore, using IndiBias we compare ten different language models on multiple bias measurement metrics. We observed that the language models exhibit more bias across a majority of the intersectional groups. All the scripts utilized and datasets created in this study are publicly available.

IndiBias: A Benchmark Dataset to Measure Social Biases in Language Models for Indian Context

naacl-long.487

Poster

2403.20147

https://huggingface.co/papers/2403.20147

https://aclanthology.org/2024.naacl-short.1.bib

https://aclanthology.org/2024.naacl-short.1/

@inproceedings{chhabra-etal-2024-revisiting, title = "Revisiting Zero-Shot Abstractive Summarization in the Era of Large Language Models from the Perspective of Position Bias", author = "Chhabra, Anshuman and Askari, Hadi and Mohapatra, Prasant", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.1", doi = "10.18653/v1/2024.naacl-short.1", pages = "1--11", abstract = "We characterize and study zero-shot abstractive summarization in Large Language Models (LLMs) by measuring position bias, which we propose as a general formulation of the more restrictive lead bias phenomenon studied previously in the literature. Position bias captures the tendency of a model unfairly prioritizing information from certain parts of the input text over others, leading to undesirable behavior. Through numerous experiments on four diverse real-world datasets, we study position bias in multiple LLM models such as GPT 3.5-Turbo, Llama-2, and Dolly-v2, as well as state-of-the-art pretrained encoder-decoder abstractive summarization models such as Pegasus and BART. Our findings lead to novel insights and discussion on performance and position bias of models for zero-shot summarization tasks.", }

We characterize and study zero-shot abstractive summarization in Large Language Models (LLMs) by measuring position bias, which we propose as a general formulation of the more restrictive lead bias phenomenon studied previously in the literature. Position bias captures the tendency of a model unfairly prioritizing information from certain parts of the input text over others, leading to undesirable behavior. Through numerous experiments on four diverse real-world datasets, we study position bias in multiple LLM models such as GPT 3.5-Turbo, Llama-2, and Dolly-v2, as well as state-of-the-art pretrained encoder-decoder abstractive summarization models such as Pegasus and BART. Our findings lead to novel insights and discussion on performance and position bias of models for zero-shot summarization tasks.

Revisiting Zero-Shot Abstractive Summarization in the Era of Large Language Models from the Perspective of Position Bias

naacl-short.1

Oral

2401.01989

https://aclanthology.org/2024.naacl-short.2.bib

https://aclanthology.org/2024.naacl-short.2/

@inproceedings{tang-etal-2024-struc, title = "Struc-Bench: Are Large Language Models Good at Generating Complex Structured Tabular Data?", author = "Tang, Xiangru and Zong, Yiming and Phang, Jason and Zhao, Yilun and Zhou, Wangchunshu and Cohan, Arman and Gerstein, Mark", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.2", doi = "10.18653/v1/2024.naacl-short.2", pages = "12--34", abstract = "Despite the remarkable capabilities of Large Language Models (LLMs) like GPT-4, producing complex, structured tabular data remains challenging. Our study assesses LLMs{'} proficiency in structuring tables and introduces a novel fine-tuning method, cognizant of data structures, to bolster their performance. We unveil Struc-Bench, a comprehensive benchmark featuring prominent LLMs (GPT-NeoX-20B, GPT-3.5, GPT-4, and Vicuna), which spans text tables, HTML, and LaTeX formats. Our proposed FormatCoT aids in crafting format-specific instructions from the intended outputs to populate this benchmark. Addressing the gap in task-centered evaluation, we propose two innovative metrics, P-Score (Prompting Score) and H-Score (Heuristical Score), to more accurately gauge LLM performance. Our experiments show that applying our structure-aware fine-tuning to LLaMA-7B leads to substantial performance gains, outshining its LLM counterparts across most measures. In-depth error analysis and creating an ability map across six dimensions, coverage, formatting, reasoning, comprehension, pragmatics, and hallucination, highlight areas for future enhancements and suggest forthcoming research trajectories. Our code and models can be found at https://github.com/gersteinlab/Struc-Bench.", }

Despite the remarkable capabilities of Large Language Models (LLMs) like GPT-4, producing complex, structured tabular data remains challenging. Our study assesses LLMs{'} proficiency in structuring tables and introduces a novel fine-tuning method, cognizant of data structures, to bolster their performance. We unveil Struc-Bench, a comprehensive benchmark featuring prominent LLMs (GPT-NeoX-20B, GPT-3.5, GPT-4, and Vicuna), which spans text tables, HTML, and LaTeX formats. Our proposed FormatCoT aids in crafting format-specific instructions from the intended outputs to populate this benchmark. Addressing the gap in task-centered evaluation, we propose two innovative metrics, P-Score (Prompting Score) and H-Score (Heuristical Score), to more accurately gauge LLM performance. Our experiments show that applying our structure-aware fine-tuning to LLaMA-7B leads to substantial performance gains, outshining its LLM counterparts across most measures. In-depth error analysis and creating an ability map across six dimensions, coverage, formatting, reasoning, comprehension, pragmatics, and hallucination, highlight areas for future enhancements and suggest forthcoming research trajectories. Our code and models can be found at https://github.com/gersteinlab/Struc-Bench.

Struc-Bench: Are Large Language Models Good at Generating Complex Structured Tabular Data?

naacl-short.2

Poster

https://aclanthology.org/2024.naacl-short.3.bib

https://aclanthology.org/2024.naacl-short.3/

@inproceedings{zhang-etal-2024-improving-toponym, title = "Improving Toponym Resolution by Predicting Attributes to Constrain Geographical Ontology Entries", author = "Zhang, Zeyu and Laparra, Egoitz and Bethard, Steven", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.3", doi = "10.18653/v1/2024.naacl-short.3", pages = "35--44", abstract = "Geocoding is the task of converting location mentions in text into structured geospatial data.We propose a new prompt-based paradigm for geocoding, where the machine learning algorithm encodes only the location mention and its context.We design a transformer network for predicting the country, state, and feature class of a location mention, and a deterministic algorithm that leverages the country, state, and feature class predictions as constraints in a search for compatible entries in the ontology.Our architecture, GeoPLACE, achieves new state-of-the-art performance on multiple datasets.Code and models are available at \url{https://github.com/clulab/geonorm}.", }

Geocoding is the task of converting location mentions in text into structured geospatial data.We propose a new prompt-based paradigm for geocoding, where the machine learning algorithm encodes only the location mention and its context.We design a transformer network for predicting the country, state, and feature class of a location mention, and a deterministic algorithm that leverages the country, state, and feature class predictions as constraints in a search for compatible entries in the ontology.Our architecture, GeoPLACE, achieves new state-of-the-art performance on multiple datasets.Code and models are available at \url{https://github.com/clulab/geonorm}.

Improving Toponym Resolution by Predicting Attributes to Constrain Geographical Ontology Entries

naacl-short.3

Poster

https://aclanthology.org/2024.naacl-short.4.bib

https://aclanthology.org/2024.naacl-short.4/

@inproceedings{fan-etal-2024-advancing, title = "Advancing Regular Language Reasoning in Linear Recurrent Neural Networks", author = "Fan, Ting-Han and Chi, Ta-Chung and Rudnicky, Alexander", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.4", doi = "10.18653/v1/2024.naacl-short.4", pages = "45--53", abstract = "In recent studies, linear recurrent neural networks (LRNNs) have achieved Transformer-level performance in natural language and long-range modeling, while offering rapid parallel training and constant inference cost. With the resurgence of interest in LRNNs, we study whether they can learn the hidden rules in training sequences, such as the grammatical structures of regular language. We theoretically analyze some existing LRNNs and discover their limitations in modeling regular language. Motivated by this analysis, we propose a new LRNN equipped with a block-diagonal and input-dependent transition matrix. Experiments suggest that the proposed model is the only LRNN capable of performing length extrapolation on regular language tasks such as Sum, Even Pair, and Modular Arithmetic. The code is released at \url{https://github.com/tinghanf/RegluarLRNN}.", }

In recent studies, linear recurrent neural networks (LRNNs) have achieved Transformer-level performance in natural language and long-range modeling, while offering rapid parallel training and constant inference cost. With the resurgence of interest in LRNNs, we study whether they can learn the hidden rules in training sequences, such as the grammatical structures of regular language. We theoretically analyze some existing LRNNs and discover their limitations in modeling regular language. Motivated by this analysis, we propose a new LRNN equipped with a block-diagonal and input-dependent transition matrix. Experiments suggest that the proposed model is the only LRNN capable of performing length extrapolation on regular language tasks such as Sum, Even Pair, and Modular Arithmetic. The code is released at \url{https://github.com/tinghanf/RegluarLRNN}.

Advancing Regular Language Reasoning in Linear Recurrent Neural Networks

naacl-short.4

Poster

2309.07412

https://aclanthology.org/2024.naacl-short.5.bib

https://aclanthology.org/2024.naacl-short.5/

@inproceedings{xie-etal-2024-extracting, title = "Extracting Lexical Features from Dialects via Interpretable Dialect Classifiers", author = "Xie, Roy and Ahia, Orevaoghene and Tsvetkov, Yulia and Anastasopoulos, Antonios", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.5", doi = "10.18653/v1/2024.naacl-short.5", pages = "54--69", abstract = "Identifying linguistic differences between dialects of a language often requires expert knowledge and meticulous human analysis. This is largely due to the complexity and nuance involved in studying various dialects. We present a novel approach to extract distinguishing lexical features of dialects by utilizing interpretable dialect classifiers, even in the absence of human experts. We explore both post-hoc and intrinsic approaches to interpretability, conduct experiments on Mandarin, Italian, and Low Saxon, and experimentally demonstrate that our method successfully identifies key language-specific lexical features that contribute to dialectal variations.", }

Identifying linguistic differences between dialects of a language often requires expert knowledge and meticulous human analysis. This is largely due to the complexity and nuance involved in studying various dialects. We present a novel approach to extract distinguishing lexical features of dialects by utilizing interpretable dialect classifiers, even in the absence of human experts. We explore both post-hoc and intrinsic approaches to interpretability, conduct experiments on Mandarin, Italian, and Low Saxon, and experimentally demonstrate that our method successfully identifies key language-specific lexical features that contribute to dialectal variations.

Extracting Lexical Features from Dialects via Interpretable Dialect Classifiers

naacl-short.5

Poster

2402.17914

https://aclanthology.org/2024.naacl-short.6.bib

https://aclanthology.org/2024.naacl-short.6/

@inproceedings{bai-etal-2024-clear, title = "Clear Up Confusion: Advancing Cross-Domain Few-Shot Relation Extraction through Relation-Aware Prompt Learning", author = "Bai, Ge and Lu, Chenji and Guo, Daichi and Li, Shilong and Liu, Ying and Zhang, Zhang and Dong, Guanting and Liu, Ruifang and Yong, Sun", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.6", doi = "10.18653/v1/2024.naacl-short.6", pages = "70--78", abstract = "Cross-domain few-shot Relation Extraction (RE) aims to transfer knowledge from a source domain to a different target domain to address low-resource problems.Previous work utilized label descriptions and entity information to leverage the knowledge of the source domain.However, these models are prone to confusion when directly applying this knowledge to a target domain with entirely new types of relations, which becomes particularly pronounced when facing similar relations.In this work, we propose a relation-aware prompt learning method with pre-training.Specifically, we empower the model to clear confusion by decomposing various relation types through an innovative label prompt, while a context prompt is employed to capture differences in different scenarios, enabling the model to further discern confusion. Two pre-training tasks are designed to leverage the prompt knowledge and paradigm.Experiments show that our method outperforms previous sota methods, yielding significantly better results on cross-domain few-shot RE tasks.", }

Cross-domain few-shot Relation Extraction (RE) aims to transfer knowledge from a source domain to a different target domain to address low-resource problems.Previous work utilized label descriptions and entity information to leverage the knowledge of the source domain.However, these models are prone to confusion when directly applying this knowledge to a target domain with entirely new types of relations, which becomes particularly pronounced when facing similar relations.In this work, we propose a relation-aware prompt learning method with pre-training.Specifically, we empower the model to clear confusion by decomposing various relation types through an innovative label prompt, while a context prompt is employed to capture differences in different scenarios, enabling the model to further discern confusion. Two pre-training tasks are designed to leverage the prompt knowledge and paradigm.Experiments show that our method outperforms previous sota methods, yielding significantly better results on cross-domain few-shot RE tasks.

Clear Up Confusion: Advancing Cross-Domain Few-Shot Relation Extraction through Relation-Aware Prompt Learning

naacl-short.6

Poster

https://aclanthology.org/2024.naacl-short.7.bib

https://aclanthology.org/2024.naacl-short.7/

@inproceedings{li-etal-2024-fusion, title = "Fusion Makes Perfection: An Efficient Multi-Grained Matching Approach for Zero-Shot Relation Extraction", author = "Li, Shilong and Bai, Ge and Zhang, Zhang and Liu, Ying and Lu, Chenji and Guo, Daichi and Liu, Ruifang and Yong, Sun", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.7", doi = "10.18653/v1/2024.naacl-short.7", pages = "79--85", abstract = "Predicting unseen relations that cannot be observed during the training phase is a challenging task in relation extraction. Previous works have made progress by matching the semantics between input instances and label descriptions. However, fine-grained matching often requires laborious manual annotation, and rich interactions between instances and label descriptions come with significant computational overhead. In this work, we propose an efficient multi-grained matching approach that uses virtual entity matching to reduce manual annotation cost, and fuses coarse-grained recall and fine-grained classification for rich interactions with guaranteed inference speed.Experimental results show that our approach outperforms the previous State Of The Art (SOTA) methods, and achieves a balance between inference efficiency and prediction accuracy in zero-shot relation extraction tasks.Our code is available at https://github.com/longls777/EMMA.", }

Predicting unseen relations that cannot be observed during the training phase is a challenging task in relation extraction. Previous works have made progress by matching the semantics between input instances and label descriptions. However, fine-grained matching often requires laborious manual annotation, and rich interactions between instances and label descriptions come with significant computational overhead. In this work, we propose an efficient multi-grained matching approach that uses virtual entity matching to reduce manual annotation cost, and fuses coarse-grained recall and fine-grained classification for rich interactions with guaranteed inference speed.Experimental results show that our approach outperforms the previous State Of The Art (SOTA) methods, and achieves a balance between inference efficiency and prediction accuracy in zero-shot relation extraction tasks.Our code is available at https://github.com/longls777/EMMA.

Fusion Makes Perfection: An Efficient Multi-Grained Matching Approach for Zero-Shot Relation Extraction

naacl-short.7

Poster

2406.11429

https://aclanthology.org/2024.naacl-short.8.bib

https://aclanthology.org/2024.naacl-short.8/

@inproceedings{xu-xu-2024-personalized, title = "Personalized Review Recommendation based on Implicit dimension mining", author = "Xu, Bei and Xu, Yifan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.8", doi = "10.18653/v1/2024.naacl-short.8", pages = "86--91", abstract = "Users usually browse product reviews before buying products from e-commerce websites. Lots of e-commerce websites can recommend reviews. However, existing research on review recommendation mainly focuses on the general usefulness of reviews and ignores personalized and implicit requirements. To address the issue, we propose a Large language model driven Personalized Review Recommendation model based on Implicit dimension mining (PRR-LI). The model mines implicit dimensions from reviews and requirements, and encodes them in the form of {``}text + dimension{''}. The experiments show that our model significantly outperforms other state-of-the-art textual models on the Amazon-MRHP dataset, with some of the metrics outperforming the state-of-the-art multimodal models. And we prove that encoding {``}text + dimension{''} is better than encoding {``}text{''} and {``}dimension{''} separately in review recommendation.", }

Users usually browse product reviews before buying products from e-commerce websites. Lots of e-commerce websites can recommend reviews. However, existing research on review recommendation mainly focuses on the general usefulness of reviews and ignores personalized and implicit requirements. To address the issue, we propose a Large language model driven Personalized Review Recommendation model based on Implicit dimension mining (PRR-LI). The model mines implicit dimensions from reviews and requirements, and encodes them in the form of {``}text + dimension{''}. The experiments show that our model significantly outperforms other state-of-the-art textual models on the Amazon-MRHP dataset, with some of the metrics outperforming the state-of-the-art multimodal models. And we prove that encoding {``}text + dimension{''} is better than encoding {``}text{''} and {``}dimension{''} separately in review recommendation.

Personalized Review Recommendation based on Implicit dimension mining

naacl-short.8

Poster

https://aclanthology.org/2024.naacl-short.9.bib

https://aclanthology.org/2024.naacl-short.9/

@inproceedings{liu-etal-2024-unlocking, title = "Unlocking Structure Measuring: Introducing {PDD}, an Automatic Metric for Positional Discourse Coherence", author = "Liu, Yinhong and Su, Yixuan and Shareghi, Ehsan and Collier, Nigel", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.9", doi = "10.18653/v1/2024.naacl-short.9", pages = "92--100", abstract = "Recent large language models (LLMs) have shown remarkable performance in aligning generated text with user intentions across various tasks. When it comes to long-form text generation, there has been a growing interest in generation from a discourse coherence perspective.However, existing lexical or semantic metrics such as BLEU, ROUGE, BertScore cannot effectively capture the discourse coherence.The development of discourse-specific automatic evaluation methods for assessing the output of LLMs warrants greater focus and exploration. In this paper, we present a novel automatic metric designed to quantify the discourse divergence between two long-form articles.Extensive experiments on three datasets from representative domains demonstrate that our metric aligns more closely with human preferences and GPT-4 coherence evaluation, outperforming existing evaluation methods.", }

Recent large language models (LLMs) have shown remarkable performance in aligning generated text with user intentions across various tasks. When it comes to long-form text generation, there has been a growing interest in generation from a discourse coherence perspective.However, existing lexical or semantic metrics such as BLEU, ROUGE, BertScore cannot effectively capture the discourse coherence.The development of discourse-specific automatic evaluation methods for assessing the output of LLMs warrants greater focus and exploration. In this paper, we present a novel automatic metric designed to quantify the discourse divergence between two long-form articles.Extensive experiments on three datasets from representative domains demonstrate that our metric aligns more closely with human preferences and GPT-4 coherence evaluation, outperforming existing evaluation methods.

Unlocking Structure Measuring: Introducing PDD, an Automatic Metric for Positional Discourse Coherence

naacl-short.9

Oral

2402.10175

https://aclanthology.org/2024.naacl-short.10.bib

https://aclanthology.org/2024.naacl-short.10/

@inproceedings{brei-etal-2024-returning, title = "Returning to the Start: Generating Narratives with Related Endpoints", author = "Brei, Anneliese and Zhao, Chao and Chaturvedi, Snigdha", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.10", doi = "10.18653/v1/2024.naacl-short.10", pages = "101--112", abstract = "Human writers often *bookend* their writing with ending sentences that relate back to the beginning sentences in order to compose a satisfying narrative that {``}closes the loop.{''} Motivated by this observation, we propose RENarGen, a controllable story-generation paradigm that generates narratives by ensuring the first and last sentences are related and then infilling the middle sentences. Our contributions include an initial exploration of how various methods of bookending from Narratology affect language modeling for stories. Automatic and human evaluations indicate RENarGen produces better stories with more narrative closure than current autoregressive models.", }

Human writers often *bookend* their writing with ending sentences that relate back to the beginning sentences in order to compose a satisfying narrative that {``}closes the loop.{''} Motivated by this observation, we propose RENarGen, a controllable story-generation paradigm that generates narratives by ensuring the first and last sentences are related and then infilling the middle sentences. Our contributions include an initial exploration of how various methods of bookending from Narratology affect language modeling for stories. Automatic and human evaluations indicate RENarGen produces better stories with more narrative closure than current autoregressive models.

Returning to the Start: Generating Narratives with Related Endpoints

naacl-short.10

Poster

2404.00829

https://aclanthology.org/2024.naacl-short.11.bib

https://aclanthology.org/2024.naacl-short.11/

@inproceedings{ong-etal-2024-unified, title = "Unified Examination of Entity Linking in Absence of Candidate Sets", author = "Ong, Nicolas and Shavarani, Hassan and Sarkar, Anoop", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.11", doi = "10.18653/v1/2024.naacl-short.11", pages = "113--123", abstract = "Despite remarkable strides made in the development of entity linking systems in recent years, a comprehensive comparative analysis of these systems using a unified framework is notably absent. This paper addresses this oversight by introducing a new black-box benchmark and conducting a comprehensive evaluation of all state-of-the-art entity linking methods. We use an ablation study to investigate the impact of candidate sets on the performance of entity linking. Our findings uncover exactly how much such entity linking systems depend on candidate sets, and how much this limits the general applicability of each system. We present an alternative approach to candidate sets, demonstrating that leveraging the entire in-domain candidate set can serve as a viable substitute for certain models. We show the trade-off between less restrictive candidate sets, increased inference time and memory footprint for some models.", }

Despite remarkable strides made in the development of entity linking systems in recent years, a comprehensive comparative analysis of these systems using a unified framework is notably absent. This paper addresses this oversight by introducing a new black-box benchmark and conducting a comprehensive evaluation of all state-of-the-art entity linking methods. We use an ablation study to investigate the impact of candidate sets on the performance of entity linking. Our findings uncover exactly how much such entity linking systems depend on candidate sets, and how much this limits the general applicability of each system. We present an alternative approach to candidate sets, demonstrating that leveraging the entire in-domain candidate set can serve as a viable substitute for certain models. We show the trade-off between less restrictive candidate sets, increased inference time and memory footprint for some models.

Unified Examination of Entity Linking in Absence of Candidate Sets

naacl-short.11

Poster

2404.11061

https://aclanthology.org/2024.naacl-short.12.bib

https://aclanthology.org/2024.naacl-short.12/

@inproceedings{dementieva-etal-2024-multiparadetox, title = "{M}ulti{P}ara{D}etox: Extending Text Detoxification with Parallel Data to New Languages", author = "Dementieva, Daryna and Babakov, Nikolay and Panchenko, Alexander", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.12", doi = "10.18653/v1/2024.naacl-short.12", pages = "124--140", abstract = "Text detoxification is a textual style transfer (TST) task where a text is paraphrased from a toxic surface form, e.g. featuring rude words, to the neutral register. Recently, text detoxification methods found their applications in various task such as detoxification of Large Language Models (LLMs) (Leong et al., 2023; He et al., 2024; Tang et al., 2023) and toxic speech combating in social networks (Deng et al., 2023; Mun et al., 2023; Agarwal et al., 2023). All these applications are extremely important to ensure safe communication in modern digital worlds. However, the previous approaches for parallel text detoxification corpora collection{---}ParaDetox (Logacheva et al., 2022) and APPADIA (Atwell et al., 2022){---}were explored only in monolingual setup. In this work, we aim to extend ParaDetox pipeline to multiple languages presenting MultiParaDetox to automate parallel detoxification corpus collection for potentially any language. Then, we experiment with different text detoxification models{---}from unsupervised baselines to LLMs and fine-tuned models on the presented parallel corpora{---}showing the great benefit of parallel corpus presence to obtain state-of-the-art text detoxification models for any language.", }

Text detoxification is a textual style transfer (TST) task where a text is paraphrased from a toxic surface form, e.g. featuring rude words, to the neutral register. Recently, text detoxification methods found their applications in various task such as detoxification of Large Language Models (LLMs) (Leong et al., 2023; He et al., 2024; Tang et al., 2023) and toxic speech combating in social networks (Deng et al., 2023; Mun et al., 2023; Agarwal et al., 2023). All these applications are extremely important to ensure safe communication in modern digital worlds. However, the previous approaches for parallel text detoxification corpora collection{---}ParaDetox (Logacheva et al., 2022) and APPADIA (Atwell et al., 2022){---}were explored only in monolingual setup. In this work, we aim to extend ParaDetox pipeline to multiple languages presenting MultiParaDetox to automate parallel detoxification corpus collection for potentially any language. Then, we experiment with different text detoxification models{---}from unsupervised baselines to LLMs and fine-tuned models on the presented parallel corpora{---}showing the great benefit of parallel corpus presence to obtain state-of-the-art text detoxification models for any language.

MultiParaDetox: Extending Text Detoxification with Parallel Data to New Languages

naacl-short.12

Poster

2404.02037

https://huggingface.co/papers/2404.02037

https://aclanthology.org/2024.naacl-short.13.bib

https://aclanthology.org/2024.naacl-short.13/

@inproceedings{ou-xu-2024-skicse, title = "{SKICSE}: Sentence Knowable Information Prompted by {LLM}s Improves Contrastive Sentence Embeddings", author = "Ou, Fangwei and Xu, Jinan", editor = "Duh, Kevin and Gomez, Helena and Bethard, Steven", booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.naacl-short.13", doi = "10.18653/v1/2024.naacl-short.13", pages = "141--146", abstract = "Contrastive learning, which utilizes positive pairs and in-batch negatives to optimize the loss objective, has been proven to be an effective method for learning sentence embeddings. However, we argue that the previous methods of constructing positive pairs only through dropout perturbation or entailment relation are limited. Since there is more sentence knowable information (SKI) to be mined, such as sentence external knowledge, semantic analysis, and grammatical description. In this work, we first hand-craft a simple and effective prompt template that is able to obtain the knowable information of input sentences from LLMs (e.g., LLaMA). Then we combine the original sentence and its knowable information to form a positive pair for contrastive learning. We evaluate our method on standard semantic textual similarity (STS) tasks. Experimental results show that our unsupervised and supervised models using $\text{BERT}_\text{base}$ achieve an average of 78.65{\%} and 82.45{\%} Spearman{'}s correlation respectively, a 2.40{\%} and 0.88{\%} improvement compared to SimCSE. Our model outperforms the previous state-of-the-art model PromptBERT in both unsupervised and supervised settings and specifically yields a new state-of-the-art performance in supervised setting.", }

Contrastive learning, which utilizes positive pairs and in-batch negatives to optimize the loss objective, has been proven to be an effective method for learning sentence embeddings. However, we argue that the previous methods of constructing positive pairs only through dropout perturbation or entailment relation are limited. Since there is more sentence knowable information (SKI) to be mined, such as sentence external knowledge, semantic analysis, and grammatical description. In this work, we first hand-craft a simple and effective prompt template that is able to obtain the knowable information of input sentences from LLMs (e.g., LLaMA). Then we combine the original sentence and its knowable information to form a positive pair for contrastive learning. We evaluate our method on standard semantic textual similarity (STS) tasks. Experimental results show that our unsupervised and supervised models using $\text{BERT}_\text{base}$ achieve an average of 78.65{\%} and 82.45{\%} Spearman{'}s correlation respectively, a 2.40{\%} and 0.88{\%} improvement compared to SimCSE. Our model outperforms the previous state-of-the-art model PromptBERT in both unsupervised and supervised settings and specifically yields a new state-of-the-art performance in supervised setting.

SKICSE: Sentence Knowable Information Prompted by LLMs Improves Contrastive Sentence Embeddings

naacl-short.13

Poster