Datasets:

COLING

/

COLING2024-papers

like 0

Follow

International Conference on Computational Linguistics 1

https://aclanthology.org/2024.lrec-main.601.bib

https://aclanthology.org/2024.lrec-main.601/

@inproceedings{yadav-etal-2024-explicit, title = "Explicit over Implict: Explicit Diversity Conditions for Effective Question Answer Generation", author = "Yadav, Vikas and Kwon, Hyuk joon and Srinivasan, Vijay and Jin, Hongxia", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.601", pages = "6876--6882", abstract = "Question Answer Generation (QAG) is an effective data augmentation technique to improve the accuracy of question answering systems, especially in low-resource domains. While recent pretrained and large language model-based QAG methods have made substantial progress, they face the critical issue of redundant QA pair generation, affecting downstream QA systems. Implicit diversity techniques such as sampling and diverse beam search are proven effective solutions but often yield smaller diversity. We present explicit diversity conditions for QAG, focusing on spatial aspects, question types, and entities, substantially increasing diversity in QA generation. Our work emphasizes the need of explicit diversity conditions for generating diverse question-answer synthetic data by showing significant improvements in downstream QA task over existing implicit diversity techniques. In particular, generated QA pairs from explicit diversity conditions result in an average 4.1{\%} exact match and 4.5{\%} F1 improvement over implicit sampling techniques on SQuAD-DU. Our work emphasizes the need for explicit diversity conditions even more in low-resource datasets (SubjQA), where average QA performance improvements are {\textasciitilde}12{\%} EM.", }

Question Answer Generation (QAG) is an effective data augmentation technique to improve the accuracy of question answering systems, especially in low-resource domains. While recent pretrained and large language model-based QAG methods have made substantial progress, they face the critical issue of redundant QA pair generation, affecting downstream QA systems. Implicit diversity techniques such as sampling and diverse beam search are proven effective solutions but often yield smaller diversity. We present explicit diversity conditions for QAG, focusing on spatial aspects, question types, and entities, substantially increasing diversity in QA generation. Our work emphasizes the need of explicit diversity conditions for generating diverse question-answer synthetic data by showing significant improvements in downstream QA task over existing implicit diversity techniques. In particular, generated QA pairs from explicit diversity conditions result in an average 4.1{\%} exact match and 4.5{\%} F1 improvement over implicit sampling techniques on SQuAD-DU. Our work emphasizes the need for explicit diversity conditions even more in low-resource datasets (SubjQA), where average QA performance improvements are {\textasciitilde}12{\%} EM.

Explicit over Implict: Explicit Diversity Conditions for Effective Question Answer Generation

lrec-main.601

Poster

https://aclanthology.org/2024.lrec-main.602.bib

https://aclanthology.org/2024.lrec-main.602/

@inproceedings{sun-etal-2024-exploring, title = "Exploring and Mitigating Shortcut Learning for Generative Large Language Models", author = "Sun, Zechen and Xiao, Yisheng and Li, Juntao and Ji, Yixin and Chen, Wenliang and Zhang, Min", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.602", pages = "6883--6893", abstract = "Recent generative large language models (LLMs) have exhibited incredible instruction-following capabilities while keeping strong task completion ability, even without task-specific fine-tuning. Some works attribute this to the bonus of the new scaling law, in which the continuous improvement of model capacity yields emergent capabilities, e.g., reasoning and universal generalization. However, we point out that recent LLMs still show shortcut learning behavior, where the models tend to exploit spurious correlations between non-robust features and labels for prediction, which might lead to overestimating model capabilities. LLMs memorize more complex spurious correlations (i.e., task $\leftrightarrow$ feature $\leftrightarrow$ label) compared with that learned from previous pre-training and task-specific fine-tuning paradigm (i.e., feature $\leftrightarrow$ label). Based on our findings, we propose FSLI, a framework for encouraging LLMs to \textbf{F}orget \textbf{S}purious correlations and \textbf{L}earn from \textbf{I}n-context information. Experiments on three tasks show that FSFI can effectively mitigate shortcut learning. Besides, we argue not to overestimate the capabilities of LLMs and conduct evaluations in more challenging and complete test scenarios.", }

Recent generative large language models (LLMs) have exhibited incredible instruction-following capabilities while keeping strong task completion ability, even without task-specific fine-tuning. Some works attribute this to the bonus of the new scaling law, in which the continuous improvement of model capacity yields emergent capabilities, e.g., reasoning and universal generalization. However, we point out that recent LLMs still show shortcut learning behavior, where the models tend to exploit spurious correlations between non-robust features and labels for prediction, which might lead to overestimating model capabilities. LLMs memorize more complex spurious correlations (i.e., task $\leftrightarrow$ feature $\leftrightarrow$ label) compared with that learned from previous pre-training and task-specific fine-tuning paradigm (i.e., feature $\leftrightarrow$ label). Based on our findings, we propose FSLI, a framework for encouraging LLMs to \textbf{F}orget \textbf{S}purious correlations and \textbf{L}earn from \textbf{I}n-context information. Experiments on three tasks show that FSFI can effectively mitigate shortcut learning. Besides, we argue not to overestimate the capabilities of LLMs and conduct evaluations in more challenging and complete test scenarios.

Exploring and Mitigating Shortcut Learning for Generative Large Language Models

lrec-main.602

Poster

https://aclanthology.org/2024.lrec-main.603.bib

https://aclanthology.org/2024.lrec-main.603/

@inproceedings{das-etal-2024-exploring, title = "Exploring {BERT}-Based Classification Models for Detecting Phobia Subtypes: A Novel Tweet Dataset and Comparative Analysis", author = "Das, Anik and King, Milton and Hughes, James Alexander", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.603", pages = "6894--6908", abstract = "Phobias, characterized by irrational fears of specific objects or situations, can profoundly affect an individual{'}s quality of life. This research presents a comprehensive investigation into phobia classification, where we propose a novel dataset of 811,569 English tweets from user timelines spanning 102 phobia subtypes over six months, including 47,614 self-diagnosed phobia users. BERT models were leveraged to differentiate non-phobia from phobia users and classify them into 65 specific phobia subtypes. The study produced promising results, with the highest f1-score of 78.44{\%} in binary classification (phobic user or not phobic user) and 24.01{\%} in a multi-class classification (detecting the specific phobia subtype of a user). This research provides insights into people with phobias on social media and emphasizes the capacity of natural language processing and machine learning to automate the evaluation and support of mental health.", }

Phobias, characterized by irrational fears of specific objects or situations, can profoundly affect an individual{'}s quality of life. This research presents a comprehensive investigation into phobia classification, where we propose a novel dataset of 811,569 English tweets from user timelines spanning 102 phobia subtypes over six months, including 47,614 self-diagnosed phobia users. BERT models were leveraged to differentiate non-phobia from phobia users and classify them into 65 specific phobia subtypes. The study produced promising results, with the highest f1-score of 78.44{\%} in binary classification (phobic user or not phobic user) and 24.01{\%} in a multi-class classification (detecting the specific phobia subtype of a user). This research provides insights into people with phobias on social media and emphasizes the capacity of natural language processing and machine learning to automate the evaluation and support of mental health.

Exploring BERT-Based Classification Models for Detecting Phobia Subtypes: A Novel Tweet Dataset and Comparative Analysis

lrec-main.603

Poster

https://aclanthology.org/2024.lrec-main.604.bib

https://aclanthology.org/2024.lrec-main.604/

@inproceedings{verma-etal-2024-exploring, title = "Exploring Geometric Representational Disparities between Multilingual and Bilingual Translation Models", author = "Verma, Neha and Murray, Kenton and Duh, Kevin", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.604", pages = "6909--6921", abstract = "Multilingual machine translation has proven immensely useful for both parameter efficiency and overall performance across many language pairs via complete multilingual parameter sharing. However, some language pairs in multilingual models can see worse performance than in bilingual models, especially in the one-to-many translation setting. Motivated by their empirical differences, we examine the geometric differences in representations from bilingual models versus those from one-to-many multilingual models. Specifically, we compute the isotropy of these representations using intrinsic dimensionality and IsoScore, in order to measure how the representations utilize the dimensions in their underlying vector space. Using the same evaluation data in both models, we find that for a given language pair, its multilingual model decoder representations are consistently less isotropic and occupy fewer dimensions than comparable bilingual model decoder representations. Additionally, we show that much of the anisotropy in multilingual decoder representations can be attributed to modeling language-specific information, therefore limiting remaining representational capacity.", }

Multilingual machine translation has proven immensely useful for both parameter efficiency and overall performance across many language pairs via complete multilingual parameter sharing. However, some language pairs in multilingual models can see worse performance than in bilingual models, especially in the one-to-many translation setting. Motivated by their empirical differences, we examine the geometric differences in representations from bilingual models versus those from one-to-many multilingual models. Specifically, we compute the isotropy of these representations using intrinsic dimensionality and IsoScore, in order to measure how the representations utilize the dimensions in their underlying vector space. Using the same evaluation data in both models, we find that for a given language pair, its multilingual model decoder representations are consistently less isotropic and occupy fewer dimensions than comparable bilingual model decoder representations. Additionally, we show that much of the anisotropy in multilingual decoder representations can be attributed to modeling language-specific information, therefore limiting remaining representational capacity.

Exploring Geometric Representational Disparities between Multilingual and Bilingual Translation Models

lrec-main.604

Poster

https://aclanthology.org/2024.lrec-main.605.bib

https://aclanthology.org/2024.lrec-main.605/

@inproceedings{musil-marecek-2024-exploring, title = "Exploring Interpretability of Independent Components of Word Embeddings with Automated Word Intruder Test", author = "Musil, Tom{\'a}{\v{s}} and Mare{\v{c}}ek, David", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.605", pages = "6922--6928", abstract = "Independent Component Analysis (ICA) is an algorithm originally developed for finding separate sources in a mixed signal, such as a recording of multiple people in the same room speaking at the same time. Unlike Principal Component Analysis (PCA), ICA permits the representation of a word as an unstructured set of features, without any particular feature being deemed more significant than the others. In this paper, we used ICA to analyze word embeddings. We have found that ICA can be used to find semantic features of the words and these features can easily be combined to search for words that satisfy the combination. We show that most of the independent components represent such features. To quantify the interpretability of the components, we use the word intruder test, performed both by humans and by large language models. We propose to use the automated version of the word intruder test as a fast and inexpensive way of quantifying vector interpretability without the need for human effort.", }

Independent Component Analysis (ICA) is an algorithm originally developed for finding separate sources in a mixed signal, such as a recording of multiple people in the same room speaking at the same time. Unlike Principal Component Analysis (PCA), ICA permits the representation of a word as an unstructured set of features, without any particular feature being deemed more significant than the others. In this paper, we used ICA to analyze word embeddings. We have found that ICA can be used to find semantic features of the words and these features can easily be combined to search for words that satisfy the combination. We show that most of the independent components represent such features. To quantify the interpretability of the components, we use the word intruder test, performed both by humans and by large language models. We propose to use the automated version of the word intruder test as a fast and inexpensive way of quantifying vector interpretability without the need for human effort.

Exploring Interpretability of Independent Components of Word Embeddings with Automated Word Intruder Test

lrec-main.605

Poster

2212.09580

https://aclanthology.org/2024.lrec-main.606.bib

https://aclanthology.org/2024.lrec-main.606/

@inproceedings{martinelli-etal-2024-exploring, title = "Exploring Neural Topic Modeling on a Classical {L}atin Corpus", author = "Martinelli, Ginevra and Impiccich{\'e}, Paola and Fersini, Elisabetta and Mambrini, Francesco and Passarotti, Marco", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.606", pages = "6929--6934", abstract = "The large availability of processable textual resources for Classical Latin has made it possible to study Latin literature through methods and tools that support distant reading. This paper describes a number of experiments carried out to test the possibility of investigating the thematic distribution of the Classical Latin corpus Opera Latina by means of topic modeling. For this purpose, we train, optimize and compare two neural models, Product-of-Experts LDA (ProdLDA) and Embedded Topic Model (ETM), opportunely revised to deal with the textual data from a Classical Latin corpus, to evaluate which one performs better both on the basis of topic diversity and topic coherence metrics, and from a human judgment point of view. Our results show that the topics extracted by neural models are coherent and interpretable and that they are significant from the perspective of a Latin scholar. The source code of the proposed model is available at https://github.com/MIND-Lab/LatinProdLDA.", }

The large availability of processable textual resources for Classical Latin has made it possible to study Latin literature through methods and tools that support distant reading. This paper describes a number of experiments carried out to test the possibility of investigating the thematic distribution of the Classical Latin corpus Opera Latina by means of topic modeling. For this purpose, we train, optimize and compare two neural models, Product-of-Experts LDA (ProdLDA) and Embedded Topic Model (ETM), opportunely revised to deal with the textual data from a Classical Latin corpus, to evaluate which one performs better both on the basis of topic diversity and topic coherence metrics, and from a human judgment point of view. Our results show that the topics extracted by neural models are coherent and interpretable and that they are significant from the perspective of a Latin scholar. The source code of the proposed model is available at https://github.com/MIND-Lab/LatinProdLDA.

Exploring Neural Topic Modeling on a Classical Latin Corpus

lrec-main.606

Poster

https://aclanthology.org/2024.lrec-main.607.bib

https://aclanthology.org/2024.lrec-main.607/

@inproceedings{nguyen-etal-2024-exploring, title = "Exploring Pathological Speech Quality Assessment with {ASR}-Powered {W}av2{V}ec2 in Data-Scarce Context", author = "Nguyen, Tuan and Fredouille, Corinne and Ghio, Alain and Balaguer, Mathieu and Woisard, Virginie", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.607", pages = "6935--6944", abstract = "Automatic speech quality assessment has raised more attention as an alternative or support to traditional perceptual clinical evaluation. However, most research so far only gains good results on simple tasks such as binary classification, largely due to data scarcity. To deal with this challenge, current works tend to segment patients{'} audio files into many samples to augment the datasets. Nevertheless, this approach has limitations, as it indirectly relates overall audio scores to individual segments. This paper introduces a novel approach where the system learns at the audio level instead of segments despite data scarcity. This paper proposes to use the pre-trained Wav2Vec2 architecture for both SSL, and ASR as feature extractor in speech assessment. Carried out on the HNC dataset, our ASR-driven approach established a new baseline compared with other approaches, obtaining average MSE = 0.73 and MSE = 1.15 for the prediction of intelligibility and severity scores respectively, using only 95 training samples. It shows that the ASR based Wav2Vec2 model brings the best results and may indicate a strong correlation between ASR and speech quality assessment. We also measure its ability on variable segment durations and speech content, exploring factors influencing its decision.", }

Automatic speech quality assessment has raised more attention as an alternative or support to traditional perceptual clinical evaluation. However, most research so far only gains good results on simple tasks such as binary classification, largely due to data scarcity. To deal with this challenge, current works tend to segment patients{'} audio files into many samples to augment the datasets. Nevertheless, this approach has limitations, as it indirectly relates overall audio scores to individual segments. This paper introduces a novel approach where the system learns at the audio level instead of segments despite data scarcity. This paper proposes to use the pre-trained Wav2Vec2 architecture for both SSL, and ASR as feature extractor in speech assessment. Carried out on the HNC dataset, our ASR-driven approach established a new baseline compared with other approaches, obtaining average MSE = 0.73 and MSE = 1.15 for the prediction of intelligibility and severity scores respectively, using only 95 training samples. It shows that the ASR based Wav2Vec2 model brings the best results and may indicate a strong correlation between ASR and speech quality assessment. We also measure its ability on variable segment durations and speech content, exploring factors influencing its decision.

Exploring Pathological Speech Quality Assessment with ASR-Powered Wav2Vec2 in Data-Scarce Context

lrec-main.607

Poster

2403.20184

https://aclanthology.org/2024.lrec-main.608.bib

https://aclanthology.org/2024.lrec-main.608/

@inproceedings{dragos-etal-2024-exploring, title = "Exploring the Emotional Dimension of {F}rench Online Toxic Content", author = "Dragos, Valentina and Battistelli, Delphine and Sow, Fatou and Etienne, Aline", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.608", pages = "6945--6954", abstract = "One of the biggest hurdles for the effective analysis of data collected on social platforms is the need for deeper insights on the content and meaning of this data. Emotion annotation can bring new perspectives on this issue and can enable the identification of content{--}specific features. This study aims at investigating the ways in which variation in online content can be explored through emotion annotation and corpus-based analysis. The paper describes the emotion annotation of three data sets in French composed of extremist, sexist and hateful messages respectively. To this end, first a fine-grained, corpus annotation scheme was used to annotate the data sets and then several empirical studies were carried out to characterize the content in the light of emotional categories. Results suggest that emotion annotations can provide new insights for online content analysis and stronger empirical background for automatic content detection.", }

One of the biggest hurdles for the effective analysis of data collected on social platforms is the need for deeper insights on the content and meaning of this data. Emotion annotation can bring new perspectives on this issue and can enable the identification of content{--}specific features. This study aims at investigating the ways in which variation in online content can be explored through emotion annotation and corpus-based analysis. The paper describes the emotion annotation of three data sets in French composed of extremist, sexist and hateful messages respectively. To this end, first a fine-grained, corpus annotation scheme was used to annotate the data sets and then several empirical studies were carried out to characterize the content in the light of emotional categories. Results suggest that emotion annotations can provide new insights for online content analysis and stronger empirical background for automatic content detection.

Exploring the Emotional Dimension of French Online Toxic Content

lrec-main.608

Poster

https://aclanthology.org/2024.lrec-main.609.bib

https://aclanthology.org/2024.lrec-main.609/

@inproceedings{yang-2024-exploring, title = "Exploring the Generalization of Cancer Clinical Trial Eligibility Classifiers across Diseases", author = "Yang, Yumeng", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.609", pages = "6955--6965", abstract = "Clinical trials are pivotal in medical research, and NLP can enhance their success, with application in recruitment. This study aims to evaluate the generalizability of eligibility classification across a broad spectrum of clinical trials. Starting with phase 3 cancer trials, annotated with seven eligibility exclusions, then to determine how well models can generalize to non-cancer and non-phase 3 trials. To assess this, we have compiled eligibility criteria data for five types of trials: (1) additional phase 3 cancer trials, (2) phase 1 and 2 cancer trials, (3) heart disease trials, (4) type 2 diabetes trials, and (5) observational trials for any disease, comprising 2,490 annotated eligibility criteria across seven exclusion types. Our results show that models trained on the extensive cancer dataset can effectively handle criteria commonly found in non-cancer trials, such as autoimmune diseases. However, they struggle with criteria disproportionately prevalent in cancer trials, like prior malignancy. We also experiment with few-shot learning, demonstrating that a limited number of disease-specific examples can partially overcome this performance gap. We are releasing this new dataset of annotated eligibility statements to promote the development of cross-disease generalization in clinical trial classification.", }

Clinical trials are pivotal in medical research, and NLP can enhance their success, with application in recruitment. This study aims to evaluate the generalizability of eligibility classification across a broad spectrum of clinical trials. Starting with phase 3 cancer trials, annotated with seven eligibility exclusions, then to determine how well models can generalize to non-cancer and non-phase 3 trials. To assess this, we have compiled eligibility criteria data for five types of trials: (1) additional phase 3 cancer trials, (2) phase 1 and 2 cancer trials, (3) heart disease trials, (4) type 2 diabetes trials, and (5) observational trials for any disease, comprising 2,490 annotated eligibility criteria across seven exclusion types. Our results show that models trained on the extensive cancer dataset can effectively handle criteria commonly found in non-cancer trials, such as autoimmune diseases. However, they struggle with criteria disproportionately prevalent in cancer trials, like prior malignancy. We also experiment with few-shot learning, demonstrating that a limited number of disease-specific examples can partially overcome this performance gap. We are releasing this new dataset of annotated eligibility statements to promote the development of cross-disease generalization in clinical trial classification.

Exploring the Generalization of Cancer Clinical Trial Eligibility Classifiers across Diseases

lrec-main.609

Poster

2403.17135

https://aclanthology.org/2024.lrec-main.610.bib

https://aclanthology.org/2024.lrec-main.610/

@inproceedings{finch-etal-2024-exploring, title = "Exploring the Impact of Human Evaluator Group on Chat-Oriented Dialogue Evaluation", author = "Finch, Sarah E. and Finch, James D. and Choi, Jinho D.", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.610", pages = "6966--6973", abstract = "Human evaluation has been widely accepted as the standard for evaluating chat-oriented dialogue systems. However, there is a significant variation in previous work regarding who gets recruited as evaluators. Evaluator groups such as domain experts, university students, and crowdworkers have been used to assess and compare dialogue systems, although it is unclear to what extent the choice of an evaluator group can affect results. This paper analyzes the evaluator group impact on dialogue system evaluation by testing 4 state-of-the-art dialogue systems using 4 distinct evaluator groups. Our analysis reveals a robustness towards evaluator groups for Likert evaluations that is not seen for Pairwise, with only minor differences observed when changing evaluator groups. Furthermore, two notable limitations to this robustness are observed, which reveal discrepancies between evaluators with different levels of chatbot expertise and indicate that evaluator objectivity is beneficial for certain dialogue metrics.", }

Human evaluation has been widely accepted as the standard for evaluating chat-oriented dialogue systems. However, there is a significant variation in previous work regarding who gets recruited as evaluators. Evaluator groups such as domain experts, university students, and crowdworkers have been used to assess and compare dialogue systems, although it is unclear to what extent the choice of an evaluator group can affect results. This paper analyzes the evaluator group impact on dialogue system evaluation by testing 4 state-of-the-art dialogue systems using 4 distinct evaluator groups. Our analysis reveals a robustness towards evaluator groups for Likert evaluations that is not seen for Pairwise, with only minor differences observed when changing evaluator groups. Furthermore, two notable limitations to this robustness are observed, which reveal discrepancies between evaluators with different levels of chatbot expertise and indicate that evaluator objectivity is beneficial for certain dialogue metrics.

Exploring the Impact of Human Evaluator Group on Chat-Oriented Dialogue Evaluation

lrec-main.610

Poster

2309.07998

https://aclanthology.org/2024.lrec-main.611.bib

https://aclanthology.org/2024.lrec-main.611/

@inproceedings{subedi-etal-2024-exploring, title = "Exploring the Potential of Large Language Models ({LLM}s) for Low-resource Languages: A Study on Named-Entity Recognition ({NER}) and Part-Of-Speech ({POS}) Tagging for {N}epali Language", author = "Subedi, Bipesh and Regmi, Sunil and Bal, Bal Krishna and Acharya, Praveen", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.611", pages = "6974--6979", abstract = "Large Language Models (LLMs) have made significant advancements in Natural Language Processing (NLP) by excelling in various NLP tasks. This study specifically focuses on evaluating the performance of LLMs for Named Entity Recognition (NER) and Part-of-Speech (POS) tagging for a low-resource language, Nepali. The aim is to study the effectiveness of these models for languages with limited resources by conducting experiments involving various parameters and fine-tuning and evaluating two datasets namely, ILPRL and EBIQUITY. In this work, we have experimented with eight LLMs for Nepali NER and POS tagging. While some prior works utilized larger datasets than ours, our contribution lies in presenting a comprehensive analysis of multiple LLMs in a unified setting. The findings indicate that NepBERTa, trained solely in the Nepali language, demonstrated the highest performance with F1-scores of 0.76 and 0.90 in ILPRL dataset. Similarly, it achieved 0.79 and 0.97 in EBIQUITY dataset for NER and POS respectively. This study not only highlights the potential of LLMs in performing classification tasks for low-resource languages but also compares their performance with that of alternative approaches deployed for the tasks.", }

Large Language Models (LLMs) have made significant advancements in Natural Language Processing (NLP) by excelling in various NLP tasks. This study specifically focuses on evaluating the performance of LLMs for Named Entity Recognition (NER) and Part-of-Speech (POS) tagging for a low-resource language, Nepali. The aim is to study the effectiveness of these models for languages with limited resources by conducting experiments involving various parameters and fine-tuning and evaluating two datasets namely, ILPRL and EBIQUITY. In this work, we have experimented with eight LLMs for Nepali NER and POS tagging. While some prior works utilized larger datasets than ours, our contribution lies in presenting a comprehensive analysis of multiple LLMs in a unified setting. The findings indicate that NepBERTa, trained solely in the Nepali language, demonstrated the highest performance with F1-scores of 0.76 and 0.90 in ILPRL dataset. Similarly, it achieved 0.79 and 0.97 in EBIQUITY dataset for NER and POS respectively. This study not only highlights the potential of LLMs in performing classification tasks for low-resource languages but also compares their performance with that of alternative approaches deployed for the tasks.

Exploring the Potential of Large Language Models (LLMs) for Low-resource Languages: A Study on Named-Entity Recognition (NER) and Part-Of-Speech (POS) Tagging for Nepali Language

lrec-main.611

Poster

https://aclanthology.org/2024.lrec-main.612.bib

https://aclanthology.org/2024.lrec-main.612/

@inproceedings{zhao-etal-2024-exploring, title = "Exploring the Synergy of Dual-path Encoder and Alignment Module for Better Graph-to-Text Generation", author = "Zhao, Tianxin and Liu, Yingxin and Su, Xiangdong and Li, Jiang and Gao, Guanglai", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.612", pages = "6980--6991", abstract = "The mainstream approaches view the knowledge graph-to-text (KG-to-text) generation as a sequence-to-sequence task and fine-tune the pre-trained model (PLM) to generate the target text from the linearized knowledge graph. However, the linearization of knowledge graphs and the structure of PLMs lead to the loss of a large amount of graph structure information. Moreover, PLMs lack an explicit graph-text alignment strategy because of the discrepancy between structural and textual information. To solve these two problems, we propose a synergetic KG-to-text model with a dual-path encoder, an alignment module, and a guidance module. The dual-path encoder consists of a graph structure encoder and a text encoder, which can better encode the structure and text information of the knowledge graph. The alignment module contains a two-layer Transformer block and an MLP block, which aligns and integrates the information from the dual encoder. The guidance module combines an improved pointer network and an MLP block to avoid error-generated entities and ensures the fluency and accuracy of the generated text. Our approach obtains very competitive performance on three benchmark datasets. Our code is available from https://github.com/IMu-MachineLearningsxD/G2T.", }

The mainstream approaches view the knowledge graph-to-text (KG-to-text) generation as a sequence-to-sequence task and fine-tune the pre-trained model (PLM) to generate the target text from the linearized knowledge graph. However, the linearization of knowledge graphs and the structure of PLMs lead to the loss of a large amount of graph structure information. Moreover, PLMs lack an explicit graph-text alignment strategy because of the discrepancy between structural and textual information. To solve these two problems, we propose a synergetic KG-to-text model with a dual-path encoder, an alignment module, and a guidance module. The dual-path encoder consists of a graph structure encoder and a text encoder, which can better encode the structure and text information of the knowledge graph. The alignment module contains a two-layer Transformer block and an MLP block, which aligns and integrates the information from the dual encoder. The guidance module combines an improved pointer network and an MLP block to avoid error-generated entities and ensures the fluency and accuracy of the generated text. Our approach obtains very competitive performance on three benchmark datasets. Our code is available from https://github.com/IMu-MachineLearningsxD/G2T.

Exploring the Synergy of Dual-path Encoder and Alignment Module for Better Graph-to-Text Generation

lrec-main.612

Poster

https://aclanthology.org/2024.lrec-main.613.bib

https://aclanthology.org/2024.lrec-main.613/

@inproceedings{nikolaidis-etal-2024-exploring, title = "Exploring the Usability of Persuasion Techniques for Downstream Misinformation-related Classification Tasks", author = "Nikolaidis, Nikolaos and Piskorski, Jakub and Stefanovitch, Nicolas", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.613", pages = "6992--7006", abstract = "We systematically explore the predictive power of features derived from Persuasion Techniques detected in texts, for solving different tasks of interest for media analysis; notably: detecting mis/disinformation, fake news, propaganda, partisan news and conspiracy theories. Firstly, we propose a set of meaningful features, aiming to capture the persuasiveness of a text. Secondly, we assess the discriminatory power of these features in different text classification tasks on 8 selected datasets from the literature using two metrics. We also evaluate the per-task discriminatory power of each Persuasion Technique and report on different insights. We find out that most of these features have a noticeable potential to distinguish conspiracy theories, hyperpartisan news and propaganda, while we observed mixed results in the context of fake news detection.", }

We systematically explore the predictive power of features derived from Persuasion Techniques detected in texts, for solving different tasks of interest for media analysis; notably: detecting mis/disinformation, fake news, propaganda, partisan news and conspiracy theories. Firstly, we propose a set of meaningful features, aiming to capture the persuasiveness of a text. Secondly, we assess the discriminatory power of these features in different text classification tasks on 8 selected datasets from the literature using two metrics. We also evaluate the per-task discriminatory power of each Persuasion Technique and report on different insights. We find out that most of these features have a noticeable potential to distinguish conspiracy theories, hyperpartisan news and propaganda, while we observed mixed results in the context of fake news detection.

Exploring the Usability of Persuasion Techniques for Downstream Misinformation-related Classification Tasks

lrec-main.613

Poster

https://aclanthology.org/2024.lrec-main.614.bib

https://aclanthology.org/2024.lrec-main.614/

@inproceedings{challant-filhol-2024-extending, title = "Extending {AZ}ee with Non-manual Gesture Rules for {F}rench {S}ign {L}anguage", author = "Challant, Camille and Filhol, Michael", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.614", pages = "7007--7016", abstract = "This paper presents a study on non-manual gestures, using a formal model named AZee. This is an approach which allows to formally represent Sign Language (SL) discourses, but also to animate them with a virtual signer. As non-manual gestures are essential in SL and therefore necessary for a quality synthesis, we wanted to extend AZee with them, by adding some production rules to the AZee production set. For this purpose, we applied a methodology which allows to find new production rules on a corpus representing one hour of French Sign Language, the 40 br{\`e}ves (Challant and Filhol, 2022). 23 production rules for non-manual gestures in LSF have thus been determined. We took advantage of this study to directly insert these new rules in the first corpus of AZee discourses expressions, which describe with AZee the productions in SL of the 40 br{\`e}ves corpus. 533 non-manual rules were inserted in the corpus, and some updates were made. This article proposes a new version of this AZee expressions corpus.", }

This paper presents a study on non-manual gestures, using a formal model named AZee. This is an approach which allows to formally represent Sign Language (SL) discourses, but also to animate them with a virtual signer. As non-manual gestures are essential in SL and therefore necessary for a quality synthesis, we wanted to extend AZee with them, by adding some production rules to the AZee production set. For this purpose, we applied a methodology which allows to find new production rules on a corpus representing one hour of French Sign Language, the 40 br{\`e}ves (Challant and Filhol, 2022). 23 production rules for non-manual gestures in LSF have thus been determined. We took advantage of this study to directly insert these new rules in the first corpus of AZee discourses expressions, which describe with AZee the productions in SL of the 40 br{\`e}ves corpus. 533 non-manual rules were inserted in the corpus, and some updates were made. This article proposes a new version of this AZee expressions corpus.

Extending AZee with Non-manual Gesture Rules for French Sign Language

lrec-main.614

Poster

https://aclanthology.org/2024.lrec-main.615.bib

https://aclanthology.org/2024.lrec-main.615/

@inproceedings{fischer-etal-2024-extending, title = "Extending the Discourse Analysis Tool Suite with Whiteboards for Visual Qualitative Analysis", author = "Fischer, Tim and Schneider, Florian and Petersen-Frey, Fynn and Haque, Anja Silvia Mollah and Eiser, Isabel and Koch, Gertraud and Biemann, Chris", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.615", pages = "7017--7022", abstract = "In this system demonstration paper, we describe the Whiteboards extension for an existing web-based platform for digital qualitative discourse analysis. Whiteboards comprise interactive graph-based interfaces to organize and manipulate objects, which can be qualitative research data, such as documents, images, etc., and analyses of these research data, such as annotations, tags, and code structures. The proposed extension offers a customizable view of the material and a wide range of actions that enable new ways of interacting and working with such resources. We show that the visualizations facilitate various use cases of qualitative data analysis, including reflection of the research process through sampling maps, creation of actor networks, and refining code taxonomies.", }

In this system demonstration paper, we describe the Whiteboards extension for an existing web-based platform for digital qualitative discourse analysis. Whiteboards comprise interactive graph-based interfaces to organize and manipulate objects, which can be qualitative research data, such as documents, images, etc., and analyses of these research data, such as annotations, tags, and code structures. The proposed extension offers a customizable view of the material and a wide range of actions that enable new ways of interacting and working with such resources. We show that the visualizations facilitate various use cases of qualitative data analysis, including reflection of the research process through sampling maps, creation of actor networks, and refining code taxonomies.

Extending the Discourse Analysis Tool Suite with Whiteboards for Visual Qualitative Analysis

lrec-main.615

Poster

https://aclanthology.org/2024.lrec-main.616.bib

https://aclanthology.org/2024.lrec-main.616/

@inproceedings{ebadi-etal-2024-extracting, title = "Extracting Biomedical Entities from Noisy Audio Transcripts", author = "Ebadi, Nima and Morgan, Kellen and Tan, Adrian and Linares, Billy and Osborn, Sheri and Majors, Emma and Davis, Jeremy and Rios, Anthony", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.616", pages = "7023--7034", abstract = "Automatic Speech Recognition (ASR) technology is fundamental in transcribing spoken language into text, with considerable applications in the clinical realm, including streamlining medical transcription and integrating with Electronic Health Record (EHR) systems. Nevertheless, challenges persist, especially when transcriptions contain noise, leading to significant drops in performance when Natural Language Processing (NLP) models are applied. Named Entity Recognition (NER), an essential clinical task, is particularly affected by such noise, often termed the ASR-NLP gap. Prior works have primarily studied ASR{'}s efficiency in clean recordings, leaving a research gap concerning the performance in noisy environments. This paper introduces a novel dataset, BioASR-NER, designed to bridge the ASR-NLP gap in the biomedical domain, focusing on extracting adverse drug reactions and mentions of entities from the Brief Test of Adult Cognition by Telephone (BTACT) exam. Our dataset offers a comprehensive collection of almost 2,000 clean and noisy recordings. In addressing the noise challenge, we present an innovative transcript-cleaning method using GPT-4, investigating both zero-shot and few-shot methodologies. Our study further delves into an error analysis, shedding light on the types of errors in transcription software, corrections by GPT-4, and the challenges GPT-4 faces. This paper aims to foster improved understanding and potential solutions for the ASR-NLP gap, ultimately supporting enhanced healthcare documentation practices.", }

Automatic Speech Recognition (ASR) technology is fundamental in transcribing spoken language into text, with considerable applications in the clinical realm, including streamlining medical transcription and integrating with Electronic Health Record (EHR) systems. Nevertheless, challenges persist, especially when transcriptions contain noise, leading to significant drops in performance when Natural Language Processing (NLP) models are applied. Named Entity Recognition (NER), an essential clinical task, is particularly affected by such noise, often termed the ASR-NLP gap. Prior works have primarily studied ASR{'}s efficiency in clean recordings, leaving a research gap concerning the performance in noisy environments. This paper introduces a novel dataset, BioASR-NER, designed to bridge the ASR-NLP gap in the biomedical domain, focusing on extracting adverse drug reactions and mentions of entities from the Brief Test of Adult Cognition by Telephone (BTACT) exam. Our dataset offers a comprehensive collection of almost 2,000 clean and noisy recordings. In addressing the noise challenge, we present an innovative transcript-cleaning method using GPT-4, investigating both zero-shot and few-shot methodologies. Our study further delves into an error analysis, shedding light on the types of errors in transcription software, corrections by GPT-4, and the challenges GPT-4 faces. This paper aims to foster improved understanding and potential solutions for the ASR-NLP gap, ultimately supporting enhanced healthcare documentation practices.

Extracting Biomedical Entities from Noisy Audio Transcripts

lrec-main.616

Poster

2403.17363

https://aclanthology.org/2024.lrec-main.617.bib

https://aclanthology.org/2024.lrec-main.617/

@inproceedings{huang-etal-2024-extracting, title = "Extracting Financial Events from Raw Texts via Matrix Chunking", author = "Huang, Yusheng and Hu, Ning and Li, Kunping and Wang, Nan and Lin, Zhouhan", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.617", pages = "7035--7044", abstract = "Event Extraction (EE) is widely used in the Chinese financial field to provide valuable structured information. However, there are two key challenges for Chinese financial EE in application scenarios. First, events need to be extracted from raw texts, which sets it apart from previous works like the Automatic Content Extraction (ACE) EE task, where EE is treated as a classification problem given the entity spans. Second, recognizing financial entities can be laborious, as they may involve multiple elements. In this paper, we introduce CFTE, a novel task for Chinese Financial Text-to-Event extraction, which directly extracts financial events from raw texts. We further present FINEED, a Chinese FINancial Event Extraction Dataset, and an efficient MAtrix-ChunKing method called MACK, designed for the extraction of financial events from raw texts. Specifically, FINEED is manually annotated with rich linguistic features. We propose a novel two-dimensional annotation method for FINEED, which can visualize the interactions among text components. Our MACK method is fault-tolerant by preserving the tag frequency distribution when identifying financial entities. We conduct extensive experiments and the results verify the effectiveness of our MACK method.", }

Event Extraction (EE) is widely used in the Chinese financial field to provide valuable structured information. However, there are two key challenges for Chinese financial EE in application scenarios. First, events need to be extracted from raw texts, which sets it apart from previous works like the Automatic Content Extraction (ACE) EE task, where EE is treated as a classification problem given the entity spans. Second, recognizing financial entities can be laborious, as they may involve multiple elements. In this paper, we introduce CFTE, a novel task for Chinese Financial Text-to-Event extraction, which directly extracts financial events from raw texts. We further present FINEED, a Chinese FINancial Event Extraction Dataset, and an efficient MAtrix-ChunKing method called MACK, designed for the extraction of financial events from raw texts. Specifically, FINEED is manually annotated with rich linguistic features. We propose a novel two-dimensional annotation method for FINEED, which can visualize the interactions among text components. Our MACK method is fault-tolerant by preserving the tag frequency distribution when identifying financial entities. We conduct extensive experiments and the results verify the effectiveness of our MACK method.

Extracting Financial Events from Raw Texts via Matrix Chunking

lrec-main.617

Poster

https://aclanthology.org/2024.lrec-main.618.bib

https://aclanthology.org/2024.lrec-main.618/

@inproceedings{fu-etal-2024-extracting, title = "Extracting Social Determinants of Health from Pediatric Patient Notes Using Large Language Models: Novel Corpus and Methods", author = {Fu, Yujuan and Ramachandran, Giridhar Kaushik and Dobbins, Nicholas J. and Park, Namu and Leu, Michael and Rosenberg, Abby R. and Lybarger, Kevin and Xia, Fei and Uzuner, {\"O}zlem and Yetisgen, Meliha}, editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.618", pages = "7045--7056", abstract = "Social determinants of health (SDoH) play a critical role in shaping health outcomes, particularly in pediatric populations where interventions can have long-term implications. SDoH are frequently studied in the Electronic Health Record (EHR), which provides a rich repository for diverse patient data. In this work, we present a novel annotated corpus, the Pediatric Social History Annotation Corpus (PedSHAC), and evaluate the automatic extraction of detailed SDoH representations using fine-tuned and in-context learning methods with Large Language Models (LLMs). PedSHAC comprises annotated social history sections from 1,260 clinical notes obtained from pediatric patients within the University of Washington (UW) hospital system. Employing an event-based annotation scheme, PedSHAC captures ten distinct health determinants to encompass living and economic stability, prior trauma, education access, substance use history, and mental health with an overall annotator agreement of 81.9 F1. Our proposed fine-tuning LLM-based extractors achieve high performance at 78.4 F1 for event arguments. In-context learning approaches with GPT-4 demonstrate promise for reliable SDoH extraction with limited annotated examples, with extraction performance at 82.3 F1 for event triggers.", }

Social determinants of health (SDoH) play a critical role in shaping health outcomes, particularly in pediatric populations where interventions can have long-term implications. SDoH are frequently studied in the Electronic Health Record (EHR), which provides a rich repository for diverse patient data. In this work, we present a novel annotated corpus, the Pediatric Social History Annotation Corpus (PedSHAC), and evaluate the automatic extraction of detailed SDoH representations using fine-tuned and in-context learning methods with Large Language Models (LLMs). PedSHAC comprises annotated social history sections from 1,260 clinical notes obtained from pediatric patients within the University of Washington (UW) hospital system. Employing an event-based annotation scheme, PedSHAC captures ten distinct health determinants to encompass living and economic stability, prior trauma, education access, substance use history, and mental health with an overall annotator agreement of 81.9 F1. Our proposed fine-tuning LLM-based extractors achieve high performance at 78.4 F1 for event arguments. In-context learning approaches with GPT-4 demonstrate promise for reliable SDoH extraction with limited annotated examples, with extraction performance at 82.3 F1 for event triggers.

Extracting Social Determinants of Health from Pediatric Patient Notes Using Large Language Models: Novel Corpus and Methods

lrec-main.618

Poster

2404.00826

https://aclanthology.org/2024.lrec-main.619.bib

https://aclanthology.org/2024.lrec-main.619/

@inproceedings{zhang-hollenstein-2024-eye, title = "Eye-Tracking Features Masking Transformer Attention in Question-Answering Tasks", author = "Zhang, Leran and Hollenstein, Nora", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.619", pages = "7057--7070", abstract = "Eye movement features are considered to be direct signals reflecting human attention distribution with a low cost to obtain, inspiring researchers to augment language models with eye-tracking (ET) data. In this study, we select first fixation duration (FFD) and total reading time (TRT) as the cognitive signals to guide Transformer attention in question-answering (QA) tasks. We design three different ET attention masks based on the two features, either collected from human reading events or generated by a gaze-predicting model. We augment BERT and ALBERT models with attention masks structured based on the ET data. We find that augmenting a model with ET data carries linguistic features complementing the information captured by the model. It improves the models{'} performance but compromises the stability. Different Transformer models benefit from different types of ET attention masks, while ALBERT performs better than BERT. Moreover, ET data collected from real-life reading events has better model augmenting ability than the model-predicted data.", }

Eye movement features are considered to be direct signals reflecting human attention distribution with a low cost to obtain, inspiring researchers to augment language models with eye-tracking (ET) data. In this study, we select first fixation duration (FFD) and total reading time (TRT) as the cognitive signals to guide Transformer attention in question-answering (QA) tasks. We design three different ET attention masks based on the two features, either collected from human reading events or generated by a gaze-predicting model. We augment BERT and ALBERT models with attention masks structured based on the ET data. We find that augmenting a model with ET data carries linguistic features complementing the information captured by the model. It improves the models{'} performance but compromises the stability. Different Transformer models benefit from different types of ET attention masks, while ALBERT performs better than BERT. Moreover, ET data collected from real-life reading events has better model augmenting ability than the model-predicted data.

Eye-Tracking Features Masking Transformer Attention in Question-Answering Tasks

lrec-main.619

Poster

https://aclanthology.org/2024.lrec-main.620.bib

https://aclanthology.org/2024.lrec-main.620/

@inproceedings{chen-etal-2024-factorized, title = "Factorized Learning Assisted with Large Language Model for Gloss-free Sign Language Translation", author = "Chen, Zhigang and Zhou, Benjia and Li, Jun and Wan, Jun and Lei, Zhen and Jiang, Ning and Lu, Quan and Zhao, Guoqing", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.620", pages = "7071--7081", abstract = "Previous Sign Language Translation (SLT) methods achieve superior performance by relying on gloss annotations. However, labeling high-quality glosses is a labor-intensive task, which limits the further development of SLT. Although some approaches work towards gloss-free SLT through jointly training the visual encoder and translation network, these efforts still suffer from poor performance and inefficient use of the powerful Large Language Model (LLM). Most seriously, we find that directly introducing LLM into SLT will lead to insufficient learning of visual representations as LLM dominates the learning curve. To address these problems, we propose Factorized Learning assisted with Large Language Model (FLa-LLM) for gloss-free SLT. Concretely, we factorize the training process into two stages. In the visual initialing stage, we employ a lightweight translation model after the visual encoder to pre-train the visual encoder. In the LLM fine-tuning stage, we freeze the acquired knowledge in the visual encoder and integrate it with a pre-trained LLM to inspire the LLM{'}s translation potential. This factorized training strategy proves to be highly effective as evidenced by significant improvements achieved across three SLT datasets which are all conducted under the gloss-free setting.", }

Previous Sign Language Translation (SLT) methods achieve superior performance by relying on gloss annotations. However, labeling high-quality glosses is a labor-intensive task, which limits the further development of SLT. Although some approaches work towards gloss-free SLT through jointly training the visual encoder and translation network, these efforts still suffer from poor performance and inefficient use of the powerful Large Language Model (LLM). Most seriously, we find that directly introducing LLM into SLT will lead to insufficient learning of visual representations as LLM dominates the learning curve. To address these problems, we propose Factorized Learning assisted with Large Language Model (FLa-LLM) for gloss-free SLT. Concretely, we factorize the training process into two stages. In the visual initialing stage, we employ a lightweight translation model after the visual encoder to pre-train the visual encoder. In the LLM fine-tuning stage, we freeze the acquired knowledge in the visual encoder and integrate it with a pre-trained LLM to inspire the LLM{'}s translation potential. This factorized training strategy proves to be highly effective as evidenced by significant improvements achieved across three SLT datasets which are all conducted under the gloss-free setting.

Factorized Learning Assisted with Large Language Model for Gloss-free Sign Language Translation

lrec-main.620

Poster

2403.12556

https://aclanthology.org/2024.lrec-main.621.bib

https://aclanthology.org/2024.lrec-main.621/

@inproceedings{luo-etal-2024-faganet, title = "{F}a{GAN}et: An Evidence-Based Fact-Checking Model with Integrated Encoder Leveraging Contextual Information", author = "Luo, Weiyao and Ran, Junfeng and Tian, Zailong and Li, Sujian and Sui, Zhifang", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.621", pages = "7082--7088", abstract = "In the face of the rapidly growing spread of false and misleading information in the real world, manual evidence-based fact-checking efforts become increasingly challenging and time-consuming. In order to tackle this issue, we propose FaGANet, an automated and accurate fact-checking model that leverages the power of sentence-level attention and graph attention network to enhance performance. This model adeptly integrates encoder-only models with graph attention network, effectively fusing claims and evidence information for accurate identification of even well-disguised data. Experiment results showcase the significant improvement in accuracy achieved by our FaGANet model, as well as its state-of-the-art performance in the evidence-based fact-checking task. We release our code and data in https://github.com/WeiyaoLuo/FaGANet.", }

In the face of the rapidly growing spread of false and misleading information in the real world, manual evidence-based fact-checking efforts become increasingly challenging and time-consuming. In order to tackle this issue, we propose FaGANet, an automated and accurate fact-checking model that leverages the power of sentence-level attention and graph attention network to enhance performance. This model adeptly integrates encoder-only models with graph attention network, effectively fusing claims and evidence information for accurate identification of even well-disguised data. Experiment results showcase the significant improvement in accuracy achieved by our FaGANet model, as well as its state-of-the-art performance in the evidence-based fact-checking task. We release our code and data in https://github.com/WeiyaoLuo/FaGANet.

FaGANet: An Evidence-Based Fact-Checking Model with Integrated Encoder Leveraging Contextual Information

lrec-main.621

Poster

https://aclanthology.org/2024.lrec-main.622.bib

https://aclanthology.org/2024.lrec-main.622/

@inproceedings{yang-etal-2024-faima, title = "{F}ai{MA}: Feature-aware In-context Learning for Multi-domain Aspect-based Sentiment Analysis", author = "Yang, Songhua and Jiang, Xinke and Zhao, Hanjie and Zeng, Wenxuan and Liu, Hongde and Jia, Yuxiang", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.622", pages = "7089--7100", abstract = "Multi-domain aspect-based sentiment analysis (ABSA) seeks to capture fine-grained sentiment across diverse domains. While existing research narrowly focuses on single-domain applications constrained by methodological limitations and data scarcity, the reality is that sentiment naturally traverses multiple domains. Although large language models (LLMs) offer a promising solution for ABSA, it is difficult to integrate effectively with established techniques, including graph-based models and linguistics, because modifying their internal architecture is not easy. To alleviate this problem, we propose a novel framework, Feature-aware In-context Learning for Multi-domain ABSA (FaiMA). The core insight of FaiMA is to utilize in-context learning (ICL) as a feature-aware mechanism that facilitates adaptive learning in multi-domain ABSA tasks. Specifically, we employ a multi-head graph attention network as a text encoder optimized by heuristic rules for linguistic, domain, and sentiment features. Through contrastive learning, we optimize sentence representations by focusing on these diverse features. Additionally, we construct an efficient indexing mechanism, allowing FaiMA to stably retrieve highly relevant examples across multiple dimensions for any given input. To evaluate the efficacy of FaiMA, we build the first multi-domain ABSA benchmark dataset. Extensive experimental results demonstrate that FaiMA achieves significant performance improvements in multiple domains compared to baselines, increasing F1 by 2.07{\%} on average. Source code and data sets are available at https://github.com/SupritYoung/FaiMA.", }

Multi-domain aspect-based sentiment analysis (ABSA) seeks to capture fine-grained sentiment across diverse domains. While existing research narrowly focuses on single-domain applications constrained by methodological limitations and data scarcity, the reality is that sentiment naturally traverses multiple domains. Although large language models (LLMs) offer a promising solution for ABSA, it is difficult to integrate effectively with established techniques, including graph-based models and linguistics, because modifying their internal architecture is not easy. To alleviate this problem, we propose a novel framework, Feature-aware In-context Learning for Multi-domain ABSA (FaiMA). The core insight of FaiMA is to utilize in-context learning (ICL) as a feature-aware mechanism that facilitates adaptive learning in multi-domain ABSA tasks. Specifically, we employ a multi-head graph attention network as a text encoder optimized by heuristic rules for linguistic, domain, and sentiment features. Through contrastive learning, we optimize sentence representations by focusing on these diverse features. Additionally, we construct an efficient indexing mechanism, allowing FaiMA to stably retrieve highly relevant examples across multiple dimensions for any given input. To evaluate the efficacy of FaiMA, we build the first multi-domain ABSA benchmark dataset. Extensive experimental results demonstrate that FaiMA achieves significant performance improvements in multiple domains compared to baselines, increasing F1 by 2.07{\%} on average. Source code and data sets are available at https://github.com/SupritYoung/FaiMA.

FaiMA: Feature-aware In-context Learning for Multi-domain Aspect-based Sentiment Analysis

lrec-main.622

Poster

2403.01063

https://aclanthology.org/2024.lrec-main.623.bib

https://aclanthology.org/2024.lrec-main.623/

@inproceedings{sanders-etal-2024-fairification, title = "{FAIR}ification of {L}ei{L}an{D}", author = "Sanders, Eric and Petrollino, Sara and Scheifer, Gilles R. and van den Heuvel, Henk and Handy, Christopher", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.623", pages = "7101--7106", abstract = "LeiLanD (Leiden Language Data) is a searchable catalogue initiated by the Leiden University Centre for Linguistics (LUCL) with the support of CLARIAH. The catalogue contains metadata about language datasets collected at LUCL and other institutes of Leiden University. This paper describes a project to FAIRify the datasets increasing their findability and accessibility through a standardised metadata format CMDI so as to obtain a rich metadata description for all resources and to make them findable through CLARIN{'}s Virtual Language Observatory. The paper describes the creation of the catalogue and the steps that led from unstructured metadata to CMDI standards. This FAIRifi- cation of LeiLanD has enhanced the findability and accessibility of incredibly diverse collection of language datasets.", }

LeiLanD (Leiden Language Data) is a searchable catalogue initiated by the Leiden University Centre for Linguistics (LUCL) with the support of CLARIAH. The catalogue contains metadata about language datasets collected at LUCL and other institutes of Leiden University. This paper describes a project to FAIRify the datasets increasing their findability and accessibility through a standardised metadata format CMDI so as to obtain a rich metadata description for all resources and to make them findable through CLARIN{'}s Virtual Language Observatory. The paper describes the creation of the catalogue and the steps that led from unstructured metadata to CMDI standards. This FAIRifi- cation of LeiLanD has enhanced the findability and accessibility of incredibly diverse collection of language datasets.

FAIRification of LeiLanD

lrec-main.623

Poster

https://aclanthology.org/2024.lrec-main.624.bib

https://aclanthology.org/2024.lrec-main.624/

@inproceedings{pineiro-martin-etal-2024-falai, title = "{F}al{AI}: A Dataset for End-to-end Spoken Language Understanding in a Low-Resource Scenario", author = "Pineiro-Martin, Andres and Garcia-Mateo, Carmen and Docio-Fernandez, Laura and Lopez-Perez, Maria del Carmen and Gandarela-Rodriguez, Jose", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.624", pages = "7107--7116", abstract = "End-to-end (E2E) Spoken Language Understanding (SLU) systems infer structured information directly from the speech signal using a single model. Due to the success of virtual assistants and the increasing demand for speech interfaces, these architectures are being actively researched for their potential to improve system performance by exploiting acoustic information and avoiding the cascading errors of traditional architectures. However, these systems require large amounts of specific, well-labelled speech data for training, which is expensive to obtain even in English, where the number of public audio datasets for SLU is limited. In this paper, we release the FalAI dataset, the largest public SLU dataset in terms of hours (250 hours), recordings (260,000) and participants (over 10,000), which is also the first SLU dataset in Galician and the first to be obtained in a low-resource scenario. Furthermore, we present new measures of complexity for the text corpora, the strategies followed for the design, collection and validation of the dataset, and we define splits for noisy audio, hesitant audio and audio where the sentence has changed but the structured information is preserved. These novel splits provide a unique resource for testing SLU systems in challenging, real-world scenarios.", }

End-to-end (E2E) Spoken Language Understanding (SLU) systems infer structured information directly from the speech signal using a single model. Due to the success of virtual assistants and the increasing demand for speech interfaces, these architectures are being actively researched for their potential to improve system performance by exploiting acoustic information and avoiding the cascading errors of traditional architectures. However, these systems require large amounts of specific, well-labelled speech data for training, which is expensive to obtain even in English, where the number of public audio datasets for SLU is limited. In this paper, we release the FalAI dataset, the largest public SLU dataset in terms of hours (250 hours), recordings (260,000) and participants (over 10,000), which is also the first SLU dataset in Galician and the first to be obtained in a low-resource scenario. Furthermore, we present new measures of complexity for the text corpora, the strategies followed for the design, collection and validation of the dataset, and we define splits for noisy audio, hesitant audio and audio where the sentence has changed but the structured information is preserved. These novel splits provide a unique resource for testing SLU systems in challenging, real-world scenarios.

FalAI: A Dataset for End-to-end Spoken Language Understanding in a Low-Resource Scenario

lrec-main.624

Poster

https://aclanthology.org/2024.lrec-main.625.bib

https://aclanthology.org/2024.lrec-main.625/

@inproceedings{zhang-etal-2024-fast, title = "Fast Adaptation via Prompted Data: An Efficient Cross-Domain Fine-tuning Method for Large Language Models", author = "Zhang, Yiming and Yang, Hantao and Wang, Haobo and Zhao, Jake", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.625", pages = "7117--7132", abstract = "Large language models (LLMs) have achieved great success in a variety of natural language understanding tasks. However, domain discrepancies between the downstream task and the pre-training corpora may have hurdled LLMs to excel further in the vertical applications. Contrary to prior computational-heavy methods, we propose a lightweight solution to further bridge the gap in applying LLMs to diverse downstream tasks {---} a Fast Adaptation method for LLMs via Prompted Data, in short FAvPD. Notably, with FAvPD, we establish an additional adaptive tuning procedure, wherein we integrate downstream text corpora, gold labels as well as external knowledge sources and then envelop them into a form of highly controllable prompt. As a simple, easy-to-use, and versatile solution, FAvPD lies in the intersection of regimes like knowledge-augmented LLMs, fine-tuning, and adaptation techniques. With extensive experiments, we prove that FAvPD excels in both performance efficacy and training efficiency over related prior works. FAvPD is publicly available at https://github.com/Hyatio/FAvPD.", }

Large language models (LLMs) have achieved great success in a variety of natural language understanding tasks. However, domain discrepancies between the downstream task and the pre-training corpora may have hurdled LLMs to excel further in the vertical applications. Contrary to prior computational-heavy methods, we propose a lightweight solution to further bridge the gap in applying LLMs to diverse downstream tasks {---} a Fast Adaptation method for LLMs via Prompted Data, in short FAvPD. Notably, with FAvPD, we establish an additional adaptive tuning procedure, wherein we integrate downstream text corpora, gold labels as well as external knowledge sources and then envelop them into a form of highly controllable prompt. As a simple, easy-to-use, and versatile solution, FAvPD lies in the intersection of regimes like knowledge-augmented LLMs, fine-tuning, and adaptation techniques. With extensive experiments, we prove that FAvPD excels in both performance efficacy and training efficiency over related prior works. FAvPD is publicly available at https://github.com/Hyatio/FAvPD.

Fast Adaptation via Prompted Data: An Efficient Cross-Domain Fine-tuning Method for Large Language Models

lrec-main.625

Poster

https://aclanthology.org/2024.lrec-main.626.bib

https://aclanthology.org/2024.lrec-main.626/

@inproceedings{banon-etal-2024-fastspell, title = "{F}ast{S}pell: The {L}ang{I}d Magic Spell", author = "Ba{\~n}{\'o}n, Marta and Ram{\'\i}rez-S{\'a}nchez, Gema and Zaragoza-Bernabeu, Jaume and Ortiz Rojas, Sergio", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.626", pages = "7133--7140", abstract = "Language identification is a crucial component in the automated production of language resources, particularly in multilingual and big data contexts. However, commonly used language identifiers struggle to differentiate between similar or closely-related languages. This paper introduces FastSpell, a language identifier that combines fastText (a pre-trained language identifier tool) and Hunspell (a spell checker) with the aim of having a refined second-opinion before deciding which language should be assigned to a text. We provide a description of the FastSpell algorithm along with an explanation on how to use and configure it. To that end, we motivate the need of such a tool and present a benchmark including some popular language identifiers evaluated during the development of FastSpell. We show how FastSpell is useful not only to improve identification of similar languages, but also to identify new ones ignored by other tools.", }

Language identification is a crucial component in the automated production of language resources, particularly in multilingual and big data contexts. However, commonly used language identifiers struggle to differentiate between similar or closely-related languages. This paper introduces FastSpell, a language identifier that combines fastText (a pre-trained language identifier tool) and Hunspell (a spell checker) with the aim of having a refined second-opinion before deciding which language should be assigned to a text. We provide a description of the FastSpell algorithm along with an explanation on how to use and configure it. To that end, we motivate the need of such a tool and present a benchmark including some popular language identifiers evaluated during the development of FastSpell. We show how FastSpell is useful not only to improve identification of similar languages, but also to identify new ones ignored by other tools.

FastSpell: The LangId Magic Spell

lrec-main.626

Poster

2404.08345

https://aclanthology.org/2024.lrec-main.627.bib

https://aclanthology.org/2024.lrec-main.627/

@inproceedings{zhu-etal-2024-fcds, title = "{FCDS}: Fusing Constituency and Dependency Syntax into Document-Level Relation Extraction", author = "Zhu, Xudong and Kang, Zhao and Hui, Bei", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.627", pages = "7141--7152", abstract = "Document-level Relation Extraction (DocRE) aims to identify relation labels between entities within a single document. It requires handling several sentences and reasoning over them. State-of-the-art DocRE methods use a graph structure to connect entities across the document to capture dependency syntax information. However, this is insufficient to fully exploit the rich syntax information in the document. In this work, we propose to fuse constituency and dependency syntax into DocRE. It uses constituency syntax to aggregate the whole sentence information and select the instructive sentences for the pairs of targets. It exploits dependency syntax in a graph structure with constituency syntax enhancement and chooses the path between entity pairs based on the dependency graph. The experimental results on datasets from various domains demonstrate the effectiveness of the proposed method.", }

Document-level Relation Extraction (DocRE) aims to identify relation labels between entities within a single document. It requires handling several sentences and reasoning over them. State-of-the-art DocRE methods use a graph structure to connect entities across the document to capture dependency syntax information. However, this is insufficient to fully exploit the rich syntax information in the document. In this work, we propose to fuse constituency and dependency syntax into DocRE. It uses constituency syntax to aggregate the whole sentence information and select the instructive sentences for the pairs of targets. It exploits dependency syntax in a graph structure with constituency syntax enhancement and chooses the path between entity pairs based on the dependency graph. The experimental results on datasets from various domains demonstrate the effectiveness of the proposed method.

FCDS: Fusing Constituency and Dependency Syntax into Document-Level Relation Extraction

lrec-main.627

Poster

2403.01886

https://aclanthology.org/2024.lrec-main.628.bib

https://aclanthology.org/2024.lrec-main.628/

@inproceedings{li-etal-2024-feature, title = "Feature Structure Matching for Multi-source Sentiment Analysis with Efficient Adaptive Tuning", author = "Li, Rui and Liu, Cheng and Tong, Yu and Dazhi, Jiang", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.628", pages = "7153--7162", abstract = "Recently, fine-tuning the large pre-trained language models on the labeled sentiment dataset achieves appealing performance. However, the obtained model may not generalize well to the other domains due to the domain shift, and it is expensive to update the entire parameters within the large models. Although some existing domain matching methods are proposed to alleviate the above issues, there are multiple relevant source domains in practice which makes the whole training more costly and complicated. To this end, we focus on the efficient unsupervised multi-source sentiment adaptation task which is more challenging and beneficial for real-world applications. Specifically, we propose to extract multi-layer features from the large pre-trained model, and design a dynamic parameters fusion module to exploit these features for both efficient and adaptive tuning. Furthermore, we propose a novel feature structure matching constraint, which enforces similar feature-wise correlations across different domains. Compared with the traditional domain matching methods which tend to pull all feature instances close, we show that the proposed feature structure matching is more robust and generalizable in the multi-source scenario. Extensive experiments on several multi-source sentiment analysis benchmarks demonstrate the effectiveness and superiority of our proposed framework.", }

Recently, fine-tuning the large pre-trained language models on the labeled sentiment dataset achieves appealing performance. However, the obtained model may not generalize well to the other domains due to the domain shift, and it is expensive to update the entire parameters within the large models. Although some existing domain matching methods are proposed to alleviate the above issues, there are multiple relevant source domains in practice which makes the whole training more costly and complicated. To this end, we focus on the efficient unsupervised multi-source sentiment adaptation task which is more challenging and beneficial for real-world applications. Specifically, we propose to extract multi-layer features from the large pre-trained model, and design a dynamic parameters fusion module to exploit these features for both efficient and adaptive tuning. Furthermore, we propose a novel feature structure matching constraint, which enforces similar feature-wise correlations across different domains. Compared with the traditional domain matching methods which tend to pull all feature instances close, we show that the proposed feature structure matching is more robust and generalizable in the multi-source scenario. Extensive experiments on several multi-source sentiment analysis benchmarks demonstrate the effectiveness and superiority of our proposed framework.

Feature Structure Matching for Multi-source Sentiment Analysis with Efficient Adaptive Tuning

lrec-main.628

Poster

https://aclanthology.org/2024.lrec-main.629.bib

https://aclanthology.org/2024.lrec-main.629/

@inproceedings{xiao-etal-2024-federated, title = "Federated Document-Level Biomedical Relation Extraction with Localized Context Contrast", author = "Xiao, Yan and Jin, Yaochu and Hao, Kuangrong", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.629", pages = "7163--7173", abstract = "Existing studies on relation extraction focus at the document level in a centralized training environment, requiring the collection of documents from various sources. However, this raises concerns about privacy protection, especially in sensitive domains such as finance and healthcare. For the first time, this work extends document-level relation extraction to a federated environment. The proposed federated framework, called FedLCC, is tailored for biomedical relation extraction that enables collaborative training without sharing raw medical texts. To fully exploit the models of all participating clients and improve the local training on individual clients, we propose a novel concept of localized context contrast on the basis of contrastive learning. By comparing and rectifying the similarity of localized context in documents between clients and the central server, the global model can better represent the documents on individual clients. Due to the lack of a widely accepted measure of non-IID text data, we introduce a novel non-IID scenario based on graph structural entropy. Experimental results on three document-level biomedical relation extraction datasets demonstrate the effectiveness of our method. Our code is available at https://github.com/xxxxyan/FedLCC.", }

Existing studies on relation extraction focus at the document level in a centralized training environment, requiring the collection of documents from various sources. However, this raises concerns about privacy protection, especially in sensitive domains such as finance and healthcare. For the first time, this work extends document-level relation extraction to a federated environment. The proposed federated framework, called FedLCC, is tailored for biomedical relation extraction that enables collaborative training without sharing raw medical texts. To fully exploit the models of all participating clients and improve the local training on individual clients, we propose a novel concept of localized context contrast on the basis of contrastive learning. By comparing and rectifying the similarity of localized context in documents between clients and the central server, the global model can better represent the documents on individual clients. Due to the lack of a widely accepted measure of non-IID text data, we introduce a novel non-IID scenario based on graph structural entropy. Experimental results on three document-level biomedical relation extraction datasets demonstrate the effectiveness of our method. Our code is available at https://github.com/xxxxyan/FedLCC.

Federated Document-Level Biomedical Relation Extraction with Localized Context Contrast

lrec-main.629

Poster

https://aclanthology.org/2024.lrec-main.630.bib

https://aclanthology.org/2024.lrec-main.630/

@inproceedings{yu-etal-2024-federated, title = "Federated Foundation Models: Privacy-Preserving and Collaborative Learning for Large Models", author = "Yu, Sixing and Munoz, Juan Pablo and Jannesari, Ali", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.630", pages = "7174--7184", abstract = "Foundation Models (FMs), such as LLaMA, BERT, GPT, ViT, and CLIP, have demonstrated remarkable success in a wide range of applications, driven by their ability to leverage vast amounts of data for pre-training. However, optimizing FMs often requires access to sensitive data, raising privacy concerns and limiting their applicability in many domains. In this paper, we propose the Federated Foundation Models (FFMs) paradigm, which combines the benefits of FMs and Federated Learning (FL) to enable privacy-preserving and collaborative learning across multiple end-users. We discuss the potential benefits and challenges of integrating FL into the lifespan of FMs, covering pre-training, fine-tuning, and application. We further outline potential future research avenues in FFM, including FFM pre-training, FFM fine-tuning, and federated prompt tuning, which allow the development of more personalized and context-aware models while ensuring data privacy. Moreover, we explore the possibility of continual/lifelong learning in FFMs, as increased computational power at the edge may unlock the potential for optimizing FMs using newly generated private data close to the data source. The proposed FFM concepts offer a flexible and scalable framework for training large language models in a privacy-preserving manner, setting the stage for subsequent advancements in both FM training and federated learning.", }

Foundation Models (FMs), such as LLaMA, BERT, GPT, ViT, and CLIP, have demonstrated remarkable success in a wide range of applications, driven by their ability to leverage vast amounts of data for pre-training. However, optimizing FMs often requires access to sensitive data, raising privacy concerns and limiting their applicability in many domains. In this paper, we propose the Federated Foundation Models (FFMs) paradigm, which combines the benefits of FMs and Federated Learning (FL) to enable privacy-preserving and collaborative learning across multiple end-users. We discuss the potential benefits and challenges of integrating FL into the lifespan of FMs, covering pre-training, fine-tuning, and application. We further outline potential future research avenues in FFM, including FFM pre-training, FFM fine-tuning, and federated prompt tuning, which allow the development of more personalized and context-aware models while ensuring data privacy. Moreover, we explore the possibility of continual/lifelong learning in FFMs, as increased computational power at the edge may unlock the potential for optimizing FMs using newly generated private data close to the data source. The proposed FFM concepts offer a flexible and scalable framework for training large language models in a privacy-preserving manner, setting the stage for subsequent advancements in both FM training and federated learning.

Federated Foundation Models: Privacy-Preserving and Collaborative Learning for Large Models

lrec-main.630

Poster

2305.11414

https://aclanthology.org/2024.lrec-main.631.bib

https://aclanthology.org/2024.lrec-main.631/

@inproceedings{li-etal-2024-shot, title = "Few-Shot Learning for Cold-Start Recommendation", author = "Li, Mingming and Hu, Songlin and Zhu, Fuqing and Zhu, Qiannan", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.631", pages = "7185--7195", abstract = "Cold-start is a significant problem in recommender systems. Recently, with the development of few-shot learning and meta-learning techniques, many researchers have devoted themselves to adopting meta-learning into recommendation as the natural scenario of few-shots. Nevertheless, we argue that recent work has a huge gap between few-shot learning and recommendations. In particular, users are locally dependent, not globally independent in recommendation. Therefore, it is necessary to formulate the local relationships between users. To accomplish this, we present a novel Few-shot learning method for Cold-Start (FCS) recommendation that consists of three hierarchical structures. More concretely, this first hierarchy is the global-meta parameters for learning the global information of all users; the second hierarchy is the local-meta parameters whose goal is to learn the adaptive cluster of local users; the third hierarchy is the specific parameters of the target user. Both the global and local information are formulated, addressing the new user{'}s problem in accordance with the few-shot records rapidly. Experimental results on two public real-world datasets show that the FCS method could produce stable improvements compared with the state-of-the-art.", }

Cold-start is a significant problem in recommender systems. Recently, with the development of few-shot learning and meta-learning techniques, many researchers have devoted themselves to adopting meta-learning into recommendation as the natural scenario of few-shots. Nevertheless, we argue that recent work has a huge gap between few-shot learning and recommendations. In particular, users are locally dependent, not globally independent in recommendation. Therefore, it is necessary to formulate the local relationships between users. To accomplish this, we present a novel Few-shot learning method for Cold-Start (FCS) recommendation that consists of three hierarchical structures. More concretely, this first hierarchy is the global-meta parameters for learning the global information of all users; the second hierarchy is the local-meta parameters whose goal is to learn the adaptive cluster of local users; the third hierarchy is the specific parameters of the target user. Both the global and local information are formulated, addressing the new user{'}s problem in accordance with the few-shot records rapidly. Experimental results on two public real-world datasets show that the FCS method could produce stable improvements compared with the state-of-the-art.

Few-Shot Learning for Cold-Start Recommendation

lrec-main.631

Poster

https://aclanthology.org/2024.lrec-main.632.bib

https://aclanthology.org/2024.lrec-main.632/

@inproceedings{wei-etal-2024-shot, title = "Few-shot Link Prediction on Hyper-relational Facts", author = "Wei, Jiyao and Guan, Saiping and Jin, Xiaolong and Guo, Jiafeng and Cheng, Xueqi", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.632", pages = "7196--7207", abstract = "Hyper-relational facts, which consist of a primary triple (head entity, relation, tail entity) and auxiliary attribute-value pairs, are widely present in real-world Knowledge Graphs (KGs). Link Prediction on Hyper-relational Facts (LPHFs) is to predict a missing element in a hyper-relational fact, which helps populate and enrich KGs. However, existing LPHFs studies usually require an amount of high-quality data. They overlook few-shot relations, which have limited instances, yet are common in real-world scenarios. Thus, we introduce a new task, Few-Shot Link Prediction on Hyper-relational Facts (FSLPHFs). It aims to predict a missing entity in a hyper-relational fact with limited support instances. To tackle FSLPHFs, we propose MetaRH, a model that learns Meta Relational information in Hyper-relational facts. MetaRH comprises three modules: relation learning, support-specific adjustment, and query inference. By capturing meta relational information from limited support instances, MetaRH can accurately predict the missing entity in a query. As there is no existing dataset available for this new task, we construct three datasets to validate the effectiveness of MetaRH. Experimental results on these datasets demonstrate that MetaRH significantly outperforms existing representative models.", }

Hyper-relational facts, which consist of a primary triple (head entity, relation, tail entity) and auxiliary attribute-value pairs, are widely present in real-world Knowledge Graphs (KGs). Link Prediction on Hyper-relational Facts (LPHFs) is to predict a missing element in a hyper-relational fact, which helps populate and enrich KGs. However, existing LPHFs studies usually require an amount of high-quality data. They overlook few-shot relations, which have limited instances, yet are common in real-world scenarios. Thus, we introduce a new task, Few-Shot Link Prediction on Hyper-relational Facts (FSLPHFs). It aims to predict a missing entity in a hyper-relational fact with limited support instances. To tackle FSLPHFs, we propose MetaRH, a model that learns Meta Relational information in Hyper-relational facts. MetaRH comprises three modules: relation learning, support-specific adjustment, and query inference. By capturing meta relational information from limited support instances, MetaRH can accurately predict the missing entity in a query. As there is no existing dataset available for this new task, we construct three datasets to validate the effectiveness of MetaRH. Experimental results on these datasets demonstrate that MetaRH significantly outperforms existing representative models.

Few-shot Link Prediction on Hyper-relational Facts

lrec-main.632

Poster

https://aclanthology.org/2024.lrec-main.633.bib

https://aclanthology.org/2024.lrec-main.633/

@inproceedings{lu-etal-2024-shot, title = "Few-Shot Multimodal Named Entity Recognition Based on Mutlimodal Causal Intervention Graph", author = "Lu, Feihong and Yang, Xiaocui and Li, Qian and Sun, Qingyun and Jiang, Ke and Ji, Cheng and Li, Jianxin", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.633", pages = "7208--7219", abstract = "Multimodal Named Entity Recognition (MNER) models typically require a significant volume of labeled data for effective training to extract relations between entities. In real-world scenarios, we frequently encounter unseen relation types. Nevertheless, existing methods are predominantly tailored for complete datasets and are not equipped to handle these new relation types. In this paper, we introduce the Few-shot Multimodal Named Entity Recognition (FMNER) task to address these novel relation types. FMNER trains in the source domain (seen types) and tests in the target domain (unseen types) with different distributions. Due to limited available resources for sampling, each sampling instance yields different content, resulting in data bias and alignment problems of multimodal units (image patches and words). To alleviate the above challenge, we propose a novel Multimodal causal Intervention graphs (MOUSING) model for FMNER. Specifically, we begin by constructing a multimodal graph that incorporates fine-grained information from multiple modalities. Subsequently, we introduce the Multimodal Causal Intervention Strategy to update the multimodal graph. It aims to decrease spurious correlations and emphasize accurate correlations between multimodal units, resulting in effectively aligned multimodal representations. Extensive experiments on two multimodal named entity recognition datasets demonstrate the superior performance of our model in the few-shot setting.", }

Multimodal Named Entity Recognition (MNER) models typically require a significant volume of labeled data for effective training to extract relations between entities. In real-world scenarios, we frequently encounter unseen relation types. Nevertheless, existing methods are predominantly tailored for complete datasets and are not equipped to handle these new relation types. In this paper, we introduce the Few-shot Multimodal Named Entity Recognition (FMNER) task to address these novel relation types. FMNER trains in the source domain (seen types) and tests in the target domain (unseen types) with different distributions. Due to limited available resources for sampling, each sampling instance yields different content, resulting in data bias and alignment problems of multimodal units (image patches and words). To alleviate the above challenge, we propose a novel Multimodal causal Intervention graphs (MOUSING) model for FMNER. Specifically, we begin by constructing a multimodal graph that incorporates fine-grained information from multiple modalities. Subsequently, we introduce the Multimodal Causal Intervention Strategy to update the multimodal graph. It aims to decrease spurious correlations and emphasize accurate correlations between multimodal units, resulting in effectively aligned multimodal representations. Extensive experiments on two multimodal named entity recognition datasets demonstrate the superior performance of our model in the few-shot setting.

Few-Shot Multimodal Named Entity Recognition Based on Mutlimodal Causal Intervention Graph

lrec-main.633

Poster

https://aclanthology.org/2024.lrec-main.634.bib

https://aclanthology.org/2024.lrec-main.634/

@inproceedings{chen-etal-2024-shot-named, title = "Few-shot Named Entity Recognition via Superposition Concept Discrimination", author = "Chen, Jiawei and Lin, Hongyu and Han, Xianpei and Lu, Yaojie and Jiang, Shanshan and Dong, Bin and Sun, Le", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.634", pages = "7220--7231", abstract = "Few-shot NER aims to identify entities of target types with only limited number of illustrative instances. Unfortunately, few-shot NER is severely challenged by the intrinsic precise generalization problem, i.e., it is hard to accurately determine the desired target type due to the ambiguity stemming from information deficiency. In this paper, we propose Superposition Concept Discriminator (SuperCD), which resolves the above challenge via an active learning paradigm. Specifically, a concept extractor is first introduced to identify superposition concepts from illustrative instances, with each concept corresponding to a possible generalization boundary. Then a superposition instance retriever is applied to retrieve corresponding instances of these superposition concepts from large-scale text corpus. Finally, annotators are asked to annotate the retrieved instances and these annotated instances together with original illustrative instances are used to learn FS-NER models. To this end, we learn a universal concept extractor and superposition instance retriever using a large-scale openly available knowledge bases. Experiments show that SuperCD can effectively identify superposition concepts from illustrative instances, retrieve superposition instances from large-scale corpus, and significantly improve the few-shot NER performance with minimal additional efforts.", }

Few-shot NER aims to identify entities of target types with only limited number of illustrative instances. Unfortunately, few-shot NER is severely challenged by the intrinsic precise generalization problem, i.e., it is hard to accurately determine the desired target type due to the ambiguity stemming from information deficiency. In this paper, we propose Superposition Concept Discriminator (SuperCD), which resolves the above challenge via an active learning paradigm. Specifically, a concept extractor is first introduced to identify superposition concepts from illustrative instances, with each concept corresponding to a possible generalization boundary. Then a superposition instance retriever is applied to retrieve corresponding instances of these superposition concepts from large-scale text corpus. Finally, annotators are asked to annotate the retrieved instances and these annotated instances together with original illustrative instances are used to learn FS-NER models. To this end, we learn a universal concept extractor and superposition instance retriever using a large-scale openly available knowledge bases. Experiments show that SuperCD can effectively identify superposition concepts from illustrative instances, retrieve superposition instances from large-scale corpus, and significantly improve the few-shot NER performance with minimal additional efforts.

Few-shot Named Entity Recognition via Superposition Concept Discrimination

lrec-main.634

Poster

2403.16463

https://aclanthology.org/2024.lrec-main.635.bib

https://aclanthology.org/2024.lrec-main.635/

@inproceedings{gong-eldardiry-2024-shot, title = "Few-Shot Relation Extraction with Hybrid Visual Evidence", author = "Gong, Jiaying and Eldardiry, Hoda", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.635", pages = "7232--7247", abstract = "The goal of few-shot relation extraction is to predict relations between name entities in a sentence when only a few labeled instances are available for training. Existing few-shot relation extraction methods focus on uni-modal information such as text only. This reduces performance when there is no clear contexts between the name entities described in text. We propose a multi-modal few-shot relation extraction model (MFS-HVE) that leverages both textual and visual semantic information to learn a multi-modal representation jointly. The MFS-HVE includes semantic feature extractors and multi-modal fusion components. The MFS-HVE semantic feature extractors are developed to extract both textual and visual features. The visual features include global image features and local object features within the image. The MFS-HVE multi-modal fusion unit integrates information from various modalities using image-guided attention, object-guided attention, and hybrid feature attention to fully capture the semantic interaction between visual regions of images and relevant texts. Extensive experiments conducted on two public datasets demonstrate that semantic visual information significantly improves performance of few-shot relation prediction.", }

The goal of few-shot relation extraction is to predict relations between name entities in a sentence when only a few labeled instances are available for training. Existing few-shot relation extraction methods focus on uni-modal information such as text only. This reduces performance when there is no clear contexts between the name entities described in text. We propose a multi-modal few-shot relation extraction model (MFS-HVE) that leverages both textual and visual semantic information to learn a multi-modal representation jointly. The MFS-HVE includes semantic feature extractors and multi-modal fusion components. The MFS-HVE semantic feature extractors are developed to extract both textual and visual features. The visual features include global image features and local object features within the image. The MFS-HVE multi-modal fusion unit integrates information from various modalities using image-guided attention, object-guided attention, and hybrid feature attention to fully capture the semantic interaction between visual regions of images and relevant texts. Extensive experiments conducted on two public datasets demonstrate that semantic visual information significantly improves performance of few-shot relation prediction.

Few-Shot Relation Extraction with Hybrid Visual Evidence

lrec-main.635

Poster

2403.00724

https://aclanthology.org/2024.lrec-main.636.bib

https://aclanthology.org/2024.lrec-main.636/

@inproceedings{li-etal-2024-shot-semantic, title = "Few-Shot Semantic Dependency Parsing via Graph Contrastive Learning", author = "Li, Bin and Fan, Yunlong and Sataer, Yikemaiti and Shi, Chuanqi and Gao, Miao and Gao, Zhiqiang", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.636", pages = "7248--7258", abstract = "Graph neural networks (GNNs) have achieved promising performance on semantic dependency parsing (SDP), owing to their powerful graph representation learning ability. However, training a high-performing GNN-based model requires a large amount of labeled data and it is prone to over-fitting in the absence of sufficient labeled data. To address this drawback, we propose a syntax-guided graph contrastive learning framework to pre-train GNNs with plenty of unlabeled data and fine-tune pre-trained GNNs with few-shot labeled SDP data. Through extensive experiments conducted on the SemEval-2015 Task 18 English dataset in three formalisms (DM, PAS, and PSD), we demonstrate that our framework achieves promising results when few-shot training samples are available. Furthermore, benefiting from the pre-training process, our framework exhibits notable advantages in the out-of-domain test sets.", }

Graph neural networks (GNNs) have achieved promising performance on semantic dependency parsing (SDP), owing to their powerful graph representation learning ability. However, training a high-performing GNN-based model requires a large amount of labeled data and it is prone to over-fitting in the absence of sufficient labeled data. To address this drawback, we propose a syntax-guided graph contrastive learning framework to pre-train GNNs with plenty of unlabeled data and fine-tune pre-trained GNNs with few-shot labeled SDP data. Through extensive experiments conducted on the SemEval-2015 Task 18 English dataset in three formalisms (DM, PAS, and PSD), we demonstrate that our framework achieves promising results when few-shot training samples are available. Furthermore, benefiting from the pre-training process, our framework exhibits notable advantages in the out-of-domain test sets.

Few-Shot Semantic Dependency Parsing via Graph Contrastive Learning

lrec-main.636

Poster

https://aclanthology.org/2024.lrec-main.637.bib

https://aclanthology.org/2024.lrec-main.637/

@inproceedings{li-etal-2024-shot-temporal, title = "Few-shot Temporal Pruning Accelerates Diffusion Models for Text Generation", author = "Li, Bocheng and Gao, Zhujin and Zhu, Yongxin and Yin, Kun and Cao, Haoyu and Jiang, Deqiang and Xu, Linli", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.637", pages = "7259--7269", abstract = "Diffusion models have achieved significant success in computer vision and shown immense potential in natural language processing applications, particularly for text generation tasks. However, generating high-quality text using these models often necessitates thousands of iterations, leading to slow sampling rates. Existing acceleration methods either neglect the importance of the distribution of sampling steps, resulting in compromised performance with smaller number of iterations, or require additional training, introducing considerable computational overheads. In this paper, we present Few-shot Temporal Pruning, a novel technique designed to accelerate diffusion models for text generation without supplementary training while effectively leveraging limited data. Employing a Bayesian optimization approach, our method effectively eliminates redundant sampling steps during the sampling process, thereby enhancing the generation speed. A comprehensive evaluation of discrete and continuous diffusion models across various tasks, including machine translation, question generation, and paraphrasing, reveals that our approach achieves competitive performance even with minimal sampling steps after down to less than 1 minute of optimization, yielding a significant acceleration of up to 400x in text generation tasks.", }

Diffusion models have achieved significant success in computer vision and shown immense potential in natural language processing applications, particularly for text generation tasks. However, generating high-quality text using these models often necessitates thousands of iterations, leading to slow sampling rates. Existing acceleration methods either neglect the importance of the distribution of sampling steps, resulting in compromised performance with smaller number of iterations, or require additional training, introducing considerable computational overheads. In this paper, we present Few-shot Temporal Pruning, a novel technique designed to accelerate diffusion models for text generation without supplementary training while effectively leveraging limited data. Employing a Bayesian optimization approach, our method effectively eliminates redundant sampling steps during the sampling process, thereby enhancing the generation speed. A comprehensive evaluation of discrete and continuous diffusion models across various tasks, including machine translation, question generation, and paraphrasing, reveals that our approach achieves competitive performance even with minimal sampling steps after down to less than 1 minute of optimization, yielding a significant acceleration of up to 400x in text generation tasks.

Few-shot Temporal Pruning Accelerates Diffusion Models for Text Generation

lrec-main.637

Poster

https://aclanthology.org/2024.lrec-main.638.bib

https://aclanthology.org/2024.lrec-main.638/

@inproceedings{kponou-etal-2024-ffstc, title = "{FFSTC}: Fongbe to {F}rench Speech Translation Corpus", author = "Kponou, D. Fortun{\'e} and Laleye, Fr{\'e}jus A. A. and Ezin, Eug{\`e}ne Cokou", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.638", pages = "7270--7276", abstract = "In this paper, we introduce the Fongbe to French Speech Translation Corpus (FFSTC). This corpus encompasses approximately 31 hours of collected Fongbe language content, featuring both French transcriptions and corresponding Fongbe voice recordings. FFSTC represents a comprehensive dataset compiled through various collection methods and the efforts of dedicated individuals. Furthermore, we conduct baseline experiments using Fairseq{'}s transformer{\_}s and conformer models to evaluate data quality and validity. Our results indicate a score BLEU of 8.96 for the transformer{\_}s model and 8.14 for the conformer model, establishing a baseline for the FFSTC corpus.", }

In this paper, we introduce the Fongbe to French Speech Translation Corpus (FFSTC). This corpus encompasses approximately 31 hours of collected Fongbe language content, featuring both French transcriptions and corresponding Fongbe voice recordings. FFSTC represents a comprehensive dataset compiled through various collection methods and the efforts of dedicated individuals. Furthermore, we conduct baseline experiments using Fairseq{'}s transformer{\_}s and conformer models to evaluate data quality and validity. Our results indicate a score BLEU of 8.96 for the transformer{\_}s model and 8.14 for the conformer model, establishing a baseline for the FFSTC corpus.

FFSTC: Fongbe to French Speech Translation Corpus

lrec-main.638

Poster

2403.05488

https://aclanthology.org/2024.lrec-main.639.bib

https://aclanthology.org/2024.lrec-main.639/

@inproceedings{hamotskyi-etal-2024-fincorpus, title = "{F}in{C}orpus-{DE}10k: A Corpus for the {G}erman Financial Domain", author = {Hamotskyi, Serhii and Kozaeva, Nata and H{\"a}nig, Christian}, editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.639", pages = "7277--7285", abstract = "We introduce a predominantly German corpus comprising 12.5k PDF documents sourced from the financial domain. The corresponding extracted textual data encompasses more than 165 million tokens derived predominantly from German, and to a lesser extent, bilingual documents. We provide detailed information about the document types included in the corpus, such as final terms, base prospectuses, annual reports, information materials, law documents, international financial reporting standards, and monthly reports from the Bundesbank, accompanied by comprehensive statistical analysis. To our knowledge, it is the first non-email German financial corpus available, and we hope it will fill this gap and foster further research in the financial domain both in the German language and in multilingual contexts.", }

We introduce a predominantly German corpus comprising 12.5k PDF documents sourced from the financial domain. The corresponding extracted textual data encompasses more than 165 million tokens derived predominantly from German, and to a lesser extent, bilingual documents. We provide detailed information about the document types included in the corpus, such as final terms, base prospectuses, annual reports, information materials, law documents, international financial reporting standards, and monthly reports from the Bundesbank, accompanied by comprehensive statistical analysis. To our knowledge, it is the first non-email German financial corpus available, and we hope it will fill this gap and foster further research in the financial domain both in the German language and in multilingual contexts.

FinCorpus-DE10k: A Corpus for the German Financial Domain

lrec-main.639

Poster

https://aclanthology.org/2024.lrec-main.640.bib

https://aclanthology.org/2024.lrec-main.640/

@inproceedings{nam-etal-2024-finding, title = "Finding Educationally Supportive Contexts for Vocabulary Learning with Attention-Based Models", author = "Nam, Sungjin and Collins-Thompson, Kevyn and Jurgens, David and Tong, Xin", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.640", pages = "7286--7295", abstract = "When learning new vocabulary, both humans and machines acquire critical information about the meaning of an unfamiliar word through contextual information in a sentence or passage. However, not all contexts are equally helpful for learning an unfamiliar {`}target{'} word. Some contexts provide a rich set of semantic clues to the target word{'}s meaning, while others are less supportive. We explore the task of finding educationally supportive contexts with respect to a given target word for vocabulary learning scenarios, particularly for improving student literacy skills. Because of their inherent context-based nature, attention-based deep learning methods provide an ideal starting point. We evaluate attention-based approaches for predicting the amount of educational support from contexts, ranging from a simple custom model using pre-trained embeddings with an additional attention layer, to a commercial Large Language Model (LLM). Using an existing major benchmark dataset for educational context support prediction, we found that a sophisticated but generic LLM had poor performance, while a simpler model using a custom attention-based approach achieved the best-known performance to date on this dataset.", }

When learning new vocabulary, both humans and machines acquire critical information about the meaning of an unfamiliar word through contextual information in a sentence or passage. However, not all contexts are equally helpful for learning an unfamiliar {`}target{'} word. Some contexts provide a rich set of semantic clues to the target word{'}s meaning, while others are less supportive. We explore the task of finding educationally supportive contexts with respect to a given target word for vocabulary learning scenarios, particularly for improving student literacy skills. Because of their inherent context-based nature, attention-based deep learning methods provide an ideal starting point. We evaluate attention-based approaches for predicting the amount of educational support from contexts, ranging from a simple custom model using pre-trained embeddings with an additional attention layer, to a commercial Large Language Model (LLM). Using an existing major benchmark dataset for educational context support prediction, we found that a sophisticated but generic LLM had poor performance, while a simpler model using a custom attention-based approach achieved the best-known performance to date on this dataset.

Finding Educationally Supportive Contexts for Vocabulary Learning with Attention-Based Models

lrec-main.640

Poster

https://aclanthology.org/2024.lrec-main.641.bib

https://aclanthology.org/2024.lrec-main.641/

@inproceedings{jahan-etal-2024-finding, title = "Finding Spoken Identifications: Using {GPT}-4 Annotation for an Efficient and Fast Dataset Creation Pipeline", author = "Jahan, Maliha and Wang, Helin and Thebaud, Thomas and Sun, Yinglun and Le, Giang Ha and Fagyal, Zsuzsanna and Scharenborg, Odette and Hasegawa-Johnson, Mark and Moro Velazquez, Laureano and Dehak, Najim", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.641", pages = "7296--7306", abstract = "The growing emphasis on fairness in speech-processing tasks requires datasets with speakers from diverse subgroups that allow training and evaluating fair speech technology systems. However, creating such datasets through manual annotation can be costly. To address this challenge, we present a semi-automated dataset creation pipeline that leverages large language models. We use this pipeline to generate a dataset of speakers identifying themself or another speaker as belonging to a particular race, ethnicity, or national origin group. We use OpenaAI{'}s GPT-4 to perform two complex annotation tasks- separating files relevant to our intended dataset from the irrelevant ones (filtering) and finding and extracting information on identifications within a transcript (tagging). By evaluating GPT-4{'}s performance using human annotations as ground truths, we show that it can reduce resources required by dataset annotation while barely losing any important information. For the filtering task, GPT-4 had a very low miss rate of 6.93{\%}. GPT-4{'}s tagging performance showed a trade-off between precision and recall, where the latter got as high as 97{\%}, but precision never exceeded 45{\%}. Our approach reduces the time required for the filtering and tagging tasks by 95{\%} and 80{\%}, respectively. We also present an in-depth error analysis of GPT-4{'}s performance.", }

The growing emphasis on fairness in speech-processing tasks requires datasets with speakers from diverse subgroups that allow training and evaluating fair speech technology systems. However, creating such datasets through manual annotation can be costly. To address this challenge, we present a semi-automated dataset creation pipeline that leverages large language models. We use this pipeline to generate a dataset of speakers identifying themself or another speaker as belonging to a particular race, ethnicity, or national origin group. We use OpenaAI{'}s GPT-4 to perform two complex annotation tasks- separating files relevant to our intended dataset from the irrelevant ones (filtering) and finding and extracting information on identifications within a transcript (tagging). By evaluating GPT-4{'}s performance using human annotations as ground truths, we show that it can reduce resources required by dataset annotation while barely losing any important information. For the filtering task, GPT-4 had a very low miss rate of 6.93{\%}. GPT-4{'}s tagging performance showed a trade-off between precision and recall, where the latter got as high as 97{\%}, but precision never exceeded 45{\%}. Our approach reduces the time required for the filtering and tagging tasks by 95{\%} and 80{\%}, respectively. We also present an in-depth error analysis of GPT-4{'}s performance.

Finding Spoken Identifications: Using GPT-4 Annotation for an Efficient and Fast Dataset Creation Pipeline

lrec-main.641

Poster

https://aclanthology.org/2024.lrec-main.642.bib

https://aclanthology.org/2024.lrec-main.642/

@inproceedings{shi-etal-2024-find, title = "Find-the-Common: A Benchmark for Explaining Visual Patterns from Images", author = "Shi, Yuting and Inoue, Naoya and Wei, Houjing and Zhao, Yufeng and Jin, Tao", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.642", pages = "7307--7313", abstract = "Recent advances in Instruction-fine-tuned Vision and Language Models (IVLMs), such as GPT-4V and InstructBLIP, have prompted some studies have started an in-depth analysis of the reasoning capabilities of IVLMs. However, Inductive Visual Reasoning, a vital skill for text-image understanding, remains underexplored due to the absence of benchmarks. In this paper, we introduce Find-the-Common (FTC): a new vision and language task for Inductive Visual Reasoning. In this task, models are required to identify an answer that explains the common attributes across visual scenes. We create a new dataset for the FTC and assess the performance of several contemporary approaches including Image-Based Reasoning, Text-Based Reasoning, and Image-Text-Based Reasoning with various models. Extensive experiments show that even state-of-the-art models like GPT-4V can only archive with 48{\%} accuracy on the FTC, for which, the FTC is a new challenge for the visual reasoning research community. Our dataset has been released and is available online: https://github.com/SSSSSeki/Find-the-common.", }

Recent advances in Instruction-fine-tuned Vision and Language Models (IVLMs), such as GPT-4V and InstructBLIP, have prompted some studies have started an in-depth analysis of the reasoning capabilities of IVLMs. However, Inductive Visual Reasoning, a vital skill for text-image understanding, remains underexplored due to the absence of benchmarks. In this paper, we introduce Find-the-Common (FTC): a new vision and language task for Inductive Visual Reasoning. In this task, models are required to identify an answer that explains the common attributes across visual scenes. We create a new dataset for the FTC and assess the performance of several contemporary approaches including Image-Based Reasoning, Text-Based Reasoning, and Image-Text-Based Reasoning with various models. Extensive experiments show that even state-of-the-art models like GPT-4V can only archive with 48{\%} accuracy on the FTC, for which, the FTC is a new challenge for the visual reasoning research community. Our dataset has been released and is available online: https://github.com/SSSSSeki/Find-the-common.

Find-the-Common: A Benchmark for Explaining Visual Patterns from Images

lrec-main.642

Poster

https://aclanthology.org/2024.lrec-main.643.bib

https://aclanthology.org/2024.lrec-main.643/

@inproceedings{mikulova-2024-fine, title = "Fine-grained Classification of Circumstantial Meanings within the {P}rague Dependency Treebank Annotation Scheme", author = "Mikulova, Marie", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.643", pages = "7314--7323", abstract = "In the contribution, we propose a formally and semantically based fine-grained classification of circumstantial meanings based on the analysis of a large number of valuable examples from the Prague Dependency Treebanks. The methodology and principles of the presented approach are elaborated in detail and demonstrated on two case studies. The classification of circumstantial meanings is carried out for the Czech language, but the methodology and principles used are language independent. The contribution also addresses the question of language universality and specificity through a comparison with English. The aim of this work is to enrich the annotation in the Prague Dependency Treebanks with detailed information on circumstantial meanings but it may also be useful for other semantically oriented projects. To the best of our knowledge, a similar corpus-based and corpus-verified elaborate classification of circumstantial meanings has not yet been proposed in any annotation project. The contribution presents the results of an ongoing work.", }

In the contribution, we propose a formally and semantically based fine-grained classification of circumstantial meanings based on the analysis of a large number of valuable examples from the Prague Dependency Treebanks. The methodology and principles of the presented approach are elaborated in detail and demonstrated on two case studies. The classification of circumstantial meanings is carried out for the Czech language, but the methodology and principles used are language independent. The contribution also addresses the question of language universality and specificity through a comparison with English. The aim of this work is to enrich the annotation in the Prague Dependency Treebanks with detailed information on circumstantial meanings but it may also be useful for other semantically oriented projects. To the best of our knowledge, a similar corpus-based and corpus-verified elaborate classification of circumstantial meanings has not yet been proposed in any annotation project. The contribution presents the results of an ongoing work.

Fine-grained Classification of Circumstantial Meanings within the Prague Dependency Treebank Annotation Scheme

lrec-main.643

Poster

https://aclanthology.org/2024.lrec-main.644.bib

https://aclanthology.org/2024.lrec-main.644/

@inproceedings{xiao-etal-2024-fine, title = "Fine-Grained Legal Argument-Pair Extraction via Coarse-Grained Pre-training", author = "Xiao, Chaojun and Sun, Yutao and Yao, Yuan and Han, Xu and Zhang, Wenbin and Liu, Zhiyuan and Sun, Maosong", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.644", pages = "7324--7335", abstract = "Legal Argument-Pair Extraction (LAE) is dedicated to the identification of interactive arguments targeting the same subject matter within legal complaints and corresponding defenses. This process serves as a foundation for automatically recognizing the focal points of disputes. Current methodologies predominantly conceptualize LAE as a supervised sentence-pair classification problem and usually necessitate extensive manual annotations, thereby constraining their scalability and general applicability. To this end, we present an innovative approach to LAE that focuses on fine-grained alignment of argument pairs, building upon coarse-grained complaint-defense pairs. This strategy stems from two key observations: 1) In general, every argument presented in a legal complaint is likely to be addressed by at least one corresponding argument in the defense. 2) It{'}s rare for multiple complaint arguments to be addressed by a single defense argument; rather, each complaint argument usually corresponds to a unique defense argument. Motivated by these insights, we develop a specialized pre-training framework. Our model employs pre-training objectives designed to exploit the coarse-grained supervision signals. This enables expressive representations of legal arguments for LAE, even when working with a limited amount of labeled data. To verify the effectiveness of our model, we construct the largest LAE datasets from two representative causes, private lending, and contract dispute. The experimental results demonstrate that our model can effectively capture informative argument knowledge from unlabeled complaint-defense pairs and outperform the unsupervised and supervised baselines by 3.7 and 2.4 points on average respectively. Besides, our model can reach superior accuracy with only half manually annotated data. The datasets and code can be found in https://github.com/thunlp/LAE.", }

Legal Argument-Pair Extraction (LAE) is dedicated to the identification of interactive arguments targeting the same subject matter within legal complaints and corresponding defenses. This process serves as a foundation for automatically recognizing the focal points of disputes. Current methodologies predominantly conceptualize LAE as a supervised sentence-pair classification problem and usually necessitate extensive manual annotations, thereby constraining their scalability and general applicability. To this end, we present an innovative approach to LAE that focuses on fine-grained alignment of argument pairs, building upon coarse-grained complaint-defense pairs. This strategy stems from two key observations: 1) In general, every argument presented in a legal complaint is likely to be addressed by at least one corresponding argument in the defense. 2) It{'}s rare for multiple complaint arguments to be addressed by a single defense argument; rather, each complaint argument usually corresponds to a unique defense argument. Motivated by these insights, we develop a specialized pre-training framework. Our model employs pre-training objectives designed to exploit the coarse-grained supervision signals. This enables expressive representations of legal arguments for LAE, even when working with a limited amount of labeled data. To verify the effectiveness of our model, we construct the largest LAE datasets from two representative causes, private lending, and contract dispute. The experimental results demonstrate that our model can effectively capture informative argument knowledge from unlabeled complaint-defense pairs and outperform the unsupervised and supervised baselines by 3.7 and 2.4 points on average respectively. Besides, our model can reach superior accuracy with only half manually annotated data. The datasets and code can be found in https://github.com/thunlp/LAE.

Fine-Grained Legal Argument-Pair Extraction via Coarse-Grained Pre-training

lrec-main.644

Poster

https://aclanthology.org/2024.lrec-main.645.bib

https://aclanthology.org/2024.lrec-main.645/

@inproceedings{gulli-etal-2024-fine, title = "Fine-Tuning a Pre-Trained {W}av2{V}ec2 Model for Automatic Speech Recognition- Experiments with De Zahrar Sproche", author = "Gulli, Andrea and Costantini, Francesco and Sidraschi, Diego and Li Destri, Emanuela", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.645", pages = "7336--7342", abstract = "We present the results of an Automatic Speech Recognition system developed to support linguistic documentation efforts. The test case is the zahrar sproche language, a Southern Bavarian variety spoken in the language island of Sauris/Zahre in Italy. We collected a dataset of 9,000 words and approximately 80 minutes of speech. The goal is to reduce the transcription workload of field linguists. The method used is a deep learning approach based on the language-specific tuning of a generic pre-trained representation model, XLS-R. The transcription quality of the experiments on the collected dataset is promising. We test the model{'}s performance on some fieldwork historical recordings, report the results, and evaluate them qualitatively. Finally, we indicate possibilities for improvement in this challenging task.", }

We present the results of an Automatic Speech Recognition system developed to support linguistic documentation efforts. The test case is the zahrar sproche language, a Southern Bavarian variety spoken in the language island of Sauris/Zahre in Italy. We collected a dataset of 9,000 words and approximately 80 minutes of speech. The goal is to reduce the transcription workload of field linguists. The method used is a deep learning approach based on the language-specific tuning of a generic pre-trained representation model, XLS-R. The transcription quality of the experiments on the collected dataset is promising. We test the model{'}s performance on some fieldwork historical recordings, report the results, and evaluate them qualitatively. Finally, we indicate possibilities for improvement in this challenging task.

Fine-Tuning a Pre-Trained Wav2Vec2 Model for Automatic Speech Recognition- Experiments with De Zahrar Sproche

lrec-main.645

Poster

https://aclanthology.org/2024.lrec-main.646.bib

https://aclanthology.org/2024.lrec-main.646/

@inproceedings{pulini-list-2024-first, title = "First Steps Towards the Integration of Resources on Historical Glossing Traditions in the History of {C}hinese: A Collection of Standardized F{\v{a}}nqi{\`e} Spellings from the Gu{\v{a}}ngy{\`u}n", author = "Pulini, Michele and List, Johann-Mattis", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.646", pages = "7343--7348", abstract = "Due to the peculiar nature of the Chinese writing system, it is difficult to assess the pronunciation of historical varieties of Chinese. In order to reconstruct ancient pronunciations, historical glossing practices play a crucial role. However, although studied thoroughly by numerous scholars, most research has been carried out in a qualitative manner, and no attempt at providing integrated resources of historical glossing practices has been made so far. Here, we present a first step towards the integration of resources on historical glossing traditions in the history of Chinese. Our starting point are so-called f{\v{a}}nqi{\`e} spellings in the Gu{\v{a}}ngy{\`u}n, one of the early rhyme books in the history of Chinese, providing pronunciations for more than 20000 Chinese characters. By standardizing digital versions of the resource using tools from computational historical linguistics, we show that we can predict historical spellings with high precision and at the same time shed light on the precision of ancient glossing practices. Although a considerably small first step, our resource could be the starting point for an integrated, standardized collection that could ultimately shed new light on the history of Chinese.", }

Due to the peculiar nature of the Chinese writing system, it is difficult to assess the pronunciation of historical varieties of Chinese. In order to reconstruct ancient pronunciations, historical glossing practices play a crucial role. However, although studied thoroughly by numerous scholars, most research has been carried out in a qualitative manner, and no attempt at providing integrated resources of historical glossing practices has been made so far. Here, we present a first step towards the integration of resources on historical glossing traditions in the history of Chinese. Our starting point are so-called f{\v{a}}nqi{\`e} spellings in the Gu{\v{a}}ngy{\`u}n, one of the early rhyme books in the history of Chinese, providing pronunciations for more than 20000 Chinese characters. By standardizing digital versions of the resource using tools from computational historical linguistics, we show that we can predict historical spellings with high precision and at the same time shed light on the precision of ancient glossing practices. Although a considerably small first step, our resource could be the starting point for an integrated, standardized collection that could ultimately shed new light on the history of Chinese.

First Steps Towards the Integration of Resources on Historical Glossing Traditions in the History of Chinese: A Collection of Standardized Fǎnqiè Spellings from the Guǎngyùn

lrec-main.646

Poster

https://aclanthology.org/2024.lrec-main.647.bib

https://aclanthology.org/2024.lrec-main.647/

@inproceedings{d-k-etal-2024-fisher, title = "Fisher Mask Nodes for Language Model Merging", author = "D K, Thennal and Nathan, Ganesh and M S, Suchithra", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.647", pages = "7349--7355", abstract = "Fine-tuning pre-trained models provides significant advantages in downstream performance. The ubiquitous nature of pre-trained models such as BERT and its derivatives in natural language processing has also led to a proliferation of task-specific fine-tuned models. As these models typically only perform one task well, additional training or ensembling is required in multi-task scenarios. The growing field of model merging provides a solution, dealing with the challenge of combining multiple task-specific models into a single multi-task model. In this study, we introduce a novel model merging method for Transformers, combining insights from previous work in Fisher-weighted averaging and the use of Fisher information in model pruning. Utilizing the Fisher information of mask nodes within the Transformer architecture, we devise a computationally efficient weighted-averaging scheme. Our method exhibits a regular and significant performance increase across various models in the BERT family, outperforming full-scale Fisher-weighted averaging in a fraction of the computational cost, with baseline performance improvements of up to +6.5 and a speedup between 57.4x and 321.7x across models. Our results prove the potential of our method in current multi-task learning environments and suggest its scalability and adaptability to new model architectures and learning scenarios.", }

Fine-tuning pre-trained models provides significant advantages in downstream performance. The ubiquitous nature of pre-trained models such as BERT and its derivatives in natural language processing has also led to a proliferation of task-specific fine-tuned models. As these models typically only perform one task well, additional training or ensembling is required in multi-task scenarios. The growing field of model merging provides a solution, dealing with the challenge of combining multiple task-specific models into a single multi-task model. In this study, we introduce a novel model merging method for Transformers, combining insights from previous work in Fisher-weighted averaging and the use of Fisher information in model pruning. Utilizing the Fisher information of mask nodes within the Transformer architecture, we devise a computationally efficient weighted-averaging scheme. Our method exhibits a regular and significant performance increase across various models in the BERT family, outperforming full-scale Fisher-weighted averaging in a fraction of the computational cost, with baseline performance improvements of up to +6.5 and a speedup between 57.4x and 321.7x across models. Our results prove the potential of our method in current multi-task learning environments and suggest its scalability and adaptability to new model architectures and learning scenarios.

Fisher Mask Nodes for Language Model Merging

lrec-main.647

Poster

2403.09891

https://aclanthology.org/2024.lrec-main.648.bib

https://aclanthology.org/2024.lrec-main.648/

@inproceedings{zhang-etal-2024-flattenquant, title = "{F}latten{Q}uant: Breaking through the Inference Compute-bound for Large Language Models with Per-tensor Quantization", author = "Zhang, Yi and Yang, Fei and Peng, Shuang and Wang, Fangyu and Pan, Aimin", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.648", pages = "7356--7365", abstract = "Large language models (LLMs) have demonstrated state-of-the-art accuracies across various tasks. However, the latency of inference and the large GPU memory consumption of LLMs restrict their deployment performance. Recently, there have been some efficient attempts to quantize LLMs, yet inference with large batch size or long sequence still has the issue of being compute-bound. Fine-grained quantization methods have showcased their proficiency in achieving low-bit quantization for LLMs, while requiring FP16 data type for linear layer computations, which is time-consuming when dealing with large batch size or long sequence. In this paper, we introduce a method called FlattenQuant, which significantly reduces the maximum value of the tensor by flattening the larger channels in the tensor, to achieve low bit per-tensor quantization with minimal accuracy loss. Our experiments show that FlattenQuant can directly use 4 bits to achieve 48.29{\%} of the linear layer calculation in LLMs, with the remaining layer using 8 bits. The 4-bit matrix multiplication introduced in the FlattenQuant method can effectively address the compute-bound caused by large matrix calculation. Our work achieves up to 2{\mbox{$\times$}} speedup and 2.3{\mbox{$\times$}} memory reduction for LLMs with negligible loss in accuracy.", }

Large language models (LLMs) have demonstrated state-of-the-art accuracies across various tasks. However, the latency of inference and the large GPU memory consumption of LLMs restrict their deployment performance. Recently, there have been some efficient attempts to quantize LLMs, yet inference with large batch size or long sequence still has the issue of being compute-bound. Fine-grained quantization methods have showcased their proficiency in achieving low-bit quantization for LLMs, while requiring FP16 data type for linear layer computations, which is time-consuming when dealing with large batch size or long sequence. In this paper, we introduce a method called FlattenQuant, which significantly reduces the maximum value of the tensor by flattening the larger channels in the tensor, to achieve low bit per-tensor quantization with minimal accuracy loss. Our experiments show that FlattenQuant can directly use 4 bits to achieve 48.29{\%} of the linear layer calculation in LLMs, with the remaining layer using 8 bits. The 4-bit matrix multiplication introduced in the FlattenQuant method can effectively address the compute-bound caused by large matrix calculation. Our work achieves up to 2{\mbox{$\times$}} speedup and 2.3{\mbox{$\times$}} memory reduction for LLMs with negligible loss in accuracy.

FlattenQuant: Breaking through the Inference Compute-bound for Large Language Models with Per-tensor Quantization

lrec-main.648

Poster

2402.17985

https://aclanthology.org/2024.lrec-main.649.bib

https://aclanthology.org/2024.lrec-main.649/

@inproceedings{gazeau-lareau-2024-flexible, title = "Flexible Lexicalization in Rule-based Text Realization", author = "Gazeau, Avril and Lareau, Francois", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.649", pages = "7366--7376", abstract = "GenDR is a text realizer that takes as input a graph-based semantic representation and outputs the corresponding syntactic dependency trees. One of the tasks in this transduction is lexicalization, i.e., choosing the right lexical units to express a given semanteme. To do so, GenDR uses a semantic dictionary that maps semantemes to corresponding lexical units in a given language. This study aims to develop a flexible lexicalization module to automatically build a rich semantic dictionary for French. To achieve this, we tried two methods. The first one consisted in extracting information from the French Lexical Network, a large-scale French lexical resource, and adapting it to GenDR. The second one was to test a contextual neural language model{'}s ability to generate potential additional lexicalizations. The first method significantly broadened the coverage of GenDR, while the additional lexicalizations produced by the language model turned out to be of limited use, which brings us to the conclusion that it is not suited to perform the task we{'}ve asked from it.", }

GenDR is a text realizer that takes as input a graph-based semantic representation and outputs the corresponding syntactic dependency trees. One of the tasks in this transduction is lexicalization, i.e., choosing the right lexical units to express a given semanteme. To do so, GenDR uses a semantic dictionary that maps semantemes to corresponding lexical units in a given language. This study aims to develop a flexible lexicalization module to automatically build a rich semantic dictionary for French. To achieve this, we tried two methods. The first one consisted in extracting information from the French Lexical Network, a large-scale French lexical resource, and adapting it to GenDR. The second one was to test a contextual neural language model{'}s ability to generate potential additional lexicalizations. The first method significantly broadened the coverage of GenDR, while the additional lexicalizations produced by the language model turned out to be of limited use, which brings us to the conclusion that it is not suited to perform the task we{'}ve asked from it.

Flexible Lexicalization in Rule-based Text Realization

lrec-main.649

Poster

https://aclanthology.org/2024.lrec-main.650.bib

https://aclanthology.org/2024.lrec-main.650/

@inproceedings{da-dalt-etal-2024-flor, title = "{FLOR}: On the Effectiveness of Language Adaptation", author = "Da Dalt, Severino and Llop, Joan and Baucells, Irene and Pamies, Marc and Xu, Yishi and Gonzalez-Agirre, Aitor and Villegas, Marta", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.650", pages = "7377--7388", abstract = "Large language models have amply proven their great capabilities, both in downstream tasks and real-life settings. However, low- and mid-resource languages do not have access to the necessary means to train such models from scratch, and often have to rely on multilingual models despite being underrepresented in the training data. For the particular case of the Catalan language, we prove that continued pre-training with vocabulary adaptation is a better alternative to take the most out of already pre-trained models, even if these have not seen any Catalan data during their pre-training phase. We curate a 26B tokens corpus and use it to further pre-train BLOOM, giving rise to the FLOR models. We perform an extensive evaluation to assess the effectiveness of our method, obtaining consistent gains across Catalan and Spanish tasks. The models, training data, and evaluation framework are made freely available under permissive licenses.", }

Large language models have amply proven their great capabilities, both in downstream tasks and real-life settings. However, low- and mid-resource languages do not have access to the necessary means to train such models from scratch, and often have to rely on multilingual models despite being underrepresented in the training data. For the particular case of the Catalan language, we prove that continued pre-training with vocabulary adaptation is a better alternative to take the most out of already pre-trained models, even if these have not seen any Catalan data during their pre-training phase. We curate a 26B tokens corpus and use it to further pre-train BLOOM, giving rise to the FLOR models. We perform an extensive evaluation to assess the effectiveness of our method, obtaining consistent gains across Catalan and Spanish tasks. The models, training data, and evaluation framework are made freely available under permissive licenses.

FLOR: On the Effectiveness of Language Adaptation

lrec-main.650

Poster

https://aclanthology.org/2024.lrec-main.651.bib

https://aclanthology.org/2024.lrec-main.651/

@inproceedings{ahmad-etal-2024-forc4cl, title = "{F}o{RC}4{CL}: A Fine-grained Field of Research Classification and Annotated Dataset of {NLP} Articles", author = "Ahmad, Raia Abu and Borisova, Ekaterina and Rehm, Georg", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.651", pages = "7389--7394", abstract = "The steep increase in the number of scholarly publications has given rise to various digital repositories, libraries and knowledge graphs aimed to capture, manage, and preserve scientific data. Efficiently navigating such databases requires a system able to classify scholarly documents according to the respective research (sub-)field. However, not every digital repository possesses a relevant classification schema for categorising publications. For instance, one of the largest digital archives in Computational Linguistics (CL) and Natural Language Processing (NLP), the ACL Anthology, lacks a system for classifying papers into topics and sub-topics. This paper addresses this gap by constructing a corpus of 1,500 ACL Anthology publications annotated with their main contributions using a novel hierarchical taxonomy of core CL/NLP topics and sub-topics. The corpus is used in a shared task with the goal of classifying CL/NLP papers into their respective sub-topics.", }

The steep increase in the number of scholarly publications has given rise to various digital repositories, libraries and knowledge graphs aimed to capture, manage, and preserve scientific data. Efficiently navigating such databases requires a system able to classify scholarly documents according to the respective research (sub-)field. However, not every digital repository possesses a relevant classification schema for categorising publications. For instance, one of the largest digital archives in Computational Linguistics (CL) and Natural Language Processing (NLP), the ACL Anthology, lacks a system for classifying papers into topics and sub-topics. This paper addresses this gap by constructing a corpus of 1,500 ACL Anthology publications annotated with their main contributions using a novel hierarchical taxonomy of core CL/NLP topics and sub-topics. The corpus is used in a shared task with the goal of classifying CL/NLP papers into their respective sub-topics.

FoRC4CL: A Fine-grained Field of Research Classification and Annotated Dataset of NLP Articles

lrec-main.651

Poster

https://aclanthology.org/2024.lrec-main.652.bib

https://aclanthology.org/2024.lrec-main.652/

@inproceedings{gorska-etal-2024-forecast2023, title = "{FORECAST}2023: A Forecast and Reasoning Corpus of Argumentation Structures", author = "G{\'o}rska, Kamila and Lawrence, John and Reed, Chris", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.652", pages = "7395--7405", abstract = "It is known from large-scale crowd experimentation that some people are innately better at analysing complex situations and making justified predictions {--} the so-called {`}superforecasters{'}. Surprisingly, however, there has to date been no work exploring the role played by the reasoning in those justifications. Bag-of-words analyses might tell us something, but the real value lies in understanding what features of reasoning and argumentation lead to better forecasts {--} both in providing an objective measure for argument quality, and even more importantly, in providing guidance on how to improve forecasting performance. The work presented here covers the creation of a unique dataset of such prediction rationales, the structure of which naturally lends itself to partially automated annotation which in turn is used as the basis for subsequent manual enhancement that provides a uniquely fine-grained and close characterisation of the structure of argumentation, with potential impact on forecasting domains from intelligence analysis to investment decision-making.", }

It is known from large-scale crowd experimentation that some people are innately better at analysing complex situations and making justified predictions {--} the so-called {`}superforecasters{'}. Surprisingly, however, there has to date been no work exploring the role played by the reasoning in those justifications. Bag-of-words analyses might tell us something, but the real value lies in understanding what features of reasoning and argumentation lead to better forecasts {--} both in providing an objective measure for argument quality, and even more importantly, in providing guidance on how to improve forecasting performance. The work presented here covers the creation of a unique dataset of such prediction rationales, the structure of which naturally lends itself to partially automated annotation which in turn is used as the basis for subsequent manual enhancement that provides a uniquely fine-grained and close characterisation of the structure of argumentation, with potential impact on forecasting domains from intelligence analysis to investment decision-making.

FORECAST2023: A Forecast and Reasoning Corpus of Argumentation Structures

lrec-main.652

Poster

https://aclanthology.org/2024.lrec-main.653.bib

https://aclanthology.org/2024.lrec-main.653/

@inproceedings{kumar-le-2024-foto, title = "{F}o{T}o: Targeted Visual Topic Modeling for Focused Analysis of Short Texts", author = "Kumar, Sanuj and Le, Tuan", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.653", pages = "7406--7416", abstract = "Given a corpus of documents, focused analysis aims to find topics relevant to aspects that a user is interested in. The aspects are often expressed by a set of keywords provided by the user. Short texts such as microblogs and tweets pose several challenges to this task because the sparsity of word co-occurrences may hinder the extraction of meaningful and relevant topics. Moreover, most of the existing topic models perform a full corpus analysis that treats all topics equally, which may make the learned topics not be on target. In this paper, we propose a novel targeted topic model for semantic short-text embedding which aims to learn all topics and low-dimensional visual representations of documents, while preserving relevant topics for focused analysis of short texts. To preserve the relevant topics in the visualization space, we propose jointly modeling topics and the pairwise document ranking based on document-keyword distances in the visualization space. The extensive experiments on several real-world datasets demonstrate the effectiveness of our proposed model in terms of targeted topic modeling and visualization.", }

Given a corpus of documents, focused analysis aims to find topics relevant to aspects that a user is interested in. The aspects are often expressed by a set of keywords provided by the user. Short texts such as microblogs and tweets pose several challenges to this task because the sparsity of word co-occurrences may hinder the extraction of meaningful and relevant topics. Moreover, most of the existing topic models perform a full corpus analysis that treats all topics equally, which may make the learned topics not be on target. In this paper, we propose a novel targeted topic model for semantic short-text embedding which aims to learn all topics and low-dimensional visual representations of documents, while preserving relevant topics for focused analysis of short texts. To preserve the relevant topics in the visualization space, we propose jointly modeling topics and the pairwise document ranking based on document-keyword distances in the visualization space. The extensive experiments on several real-world datasets demonstrate the effectiveness of our proposed model in terms of targeted topic modeling and visualization.

FoTo: Targeted Visual Topic Modeling for Focused Analysis of Short Texts

lrec-main.653

Poster

https://aclanthology.org/2024.lrec-main.654.bib

https://aclanthology.org/2024.lrec-main.654/

@inproceedings{richard-etal-2024-fracas, title = "{FRACAS}: a {FR}ench Annotated Corpus of Attribution relations in new{S}", author = "Richard, Ange and Alonzo Canul, Laura Cristina and Portet, Fran{\c{c}}ois", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.654", pages = "7417--7428", abstract = "Quotation extraction is a widely useful task both from a sociological and from a Natural Language Processing perspective. However, very little data is available to study this task in languages other than English. In this paper, we present FRACAS, a manually annotated corpus of 1,676 newswire texts in French for quotation extraction and source attribution. We first describe the composition of our corpus and the choices that were made in selecting the data. We then detail the annotation guidelines, the annotation process and give relevant statistics about our corpus. We give results for the inter-annotator agreement, which is substantially high for such a difficult linguistic phenomenon. We use this new resource to test the ability of a neural state-of-the-art relation extraction system to extract quotes and their source and we compare this model to the latest available system for quotation extraction for the French language, which is rule-based. Experiments using our dataset on the state-of-the-art system show very promising results considering the difficulty of the task at hand.", }

Quotation extraction is a widely useful task both from a sociological and from a Natural Language Processing perspective. However, very little data is available to study this task in languages other than English. In this paper, we present FRACAS, a manually annotated corpus of 1,676 newswire texts in French for quotation extraction and source attribution. We first describe the composition of our corpus and the choices that were made in selecting the data. We then detail the annotation guidelines, the annotation process and give relevant statistics about our corpus. We give results for the inter-annotator agreement, which is substantially high for such a difficult linguistic phenomenon. We use this new resource to test the ability of a neural state-of-the-art relation extraction system to extract quotes and their source and we compare this model to the latest available system for quotation extraction for the French language, which is rule-based. Experiments using our dataset on the state-of-the-art system show very promising results considering the difficulty of the task at hand.

FRACAS: a FRench Annotated Corpus of Attribution relations in newS

lrec-main.654

Poster

2309.10604

https://aclanthology.org/2024.lrec-main.655.bib

https://aclanthology.org/2024.lrec-main.655/

@inproceedings{belcavello-etal-2024-frame2, title = "Frame2: A {F}rame{N}et-based Multimodal Dataset for Tackling Text-image Interactions in Video", author = "Belcavello, Frederico and Timponi Torrent, Tiago and Matos, Ely E. and Pagano, Adriana S. and Gamonal, Maucha and Sigiliano, Natalia and Dutra, L{\'\i}via Vicente and de Andrade Abreu, Helen and Samagaio, Mairon and Carvalho, Mariane and Campos, Franciany and Azalim, Gabrielly and Mazzei, Bruna and de Oliveira, Mateus Fonseca and Lo{\c{c}}asso Luz, Ana Carolina and P{\'a}dua Ruiz, L{\'\i}via and Bellei, J{\'u}lia and Pestana, Amanda and Costa, Josiane and Rabelo, Iasmin and Silva, Anna Beatriz and Roza, Raquel and Souza, Mariana and Oliveira, Igor", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.655", pages = "7429--7437", abstract = "This paper presents the Frame2 dataset, a multimodal dataset built from a corpus of a Brazilian travel TV show annotated for FrameNet categories for both the text and image communicative modes. Frame2 comprises 230 minutes of video, which are correlated with 2,915 sentences either transcribing the audio spoken during the episodes or the subtitling segments of the show where the host conducts interviews in English. For this first release of the dataset, a total of 11,796 annotation sets for the sentences and 6,841 for the video are included. Each of the former includes a target lexical unit evoking a frame or one or more frame elements. For each video annotation, a bounding box in the image is correlated with a frame, a frame element and lexical unit evoking a frame in FrameNet.", }

This paper presents the Frame2 dataset, a multimodal dataset built from a corpus of a Brazilian travel TV show annotated for FrameNet categories for both the text and image communicative modes. Frame2 comprises 230 minutes of video, which are correlated with 2,915 sentences either transcribing the audio spoken during the episodes or the subtitling segments of the show where the host conducts interviews in English. For this first release of the dataset, a total of 11,796 annotation sets for the sentences and 6,841 for the video are included. Each of the former includes a target lexical unit evoking a frame or one or more frame elements. For each video annotation, a bounding box in the image is correlated with a frame, a frame element and lexical unit evoking a frame in FrameNet.

Frame2: A FrameNet-based Multimodal Dataset for Tackling Text-image Interactions in Video

lrec-main.655

Poster

https://aclanthology.org/2024.lrec-main.656.bib

https://aclanthology.org/2024.lrec-main.656/

@inproceedings{viridiano-etal-2024-framed, title = "Framed {M}ulti30{K}: A Frame-Based Multimodal-Multilingual Dataset", author = "Viridiano, Marcelo and Lorenzi, Arthur and Timponi Torrent, Tiago and Matos, Ely E. and Pagano, Adriana S. and Sathler Sigiliano, Nat{\'a}lia and Gamonal, Maucha and de Andrade Abreu, Helen and Vicente Dutra, L{\'\i}via and Samagaio, Mairon and Carvalho, Mariane and Campos, Franciany and Azalim, Gabrielly and Mazzei, Bruna and Fonseca de Oliveira, Mateus and Luz, Ana Carolina and Padua Ruiz, Livia and Bellei, J{\'u}lia and Pestana, Amanda and Costa, Josiane and Rabelo, Iasmin and Silva, Anna Beatriz and Roza, Raquel and Souza Mota, Mariana and Oliveira, Igor and Pelegrino de Freitas, M{\'a}rcio Henrique", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.656", pages = "7438--7449", abstract = "This paper presents Framed Multi30K (FM30K), a novel frame-based Brazilian Portuguese multimodal-multilingual dataset which i) extends the Multi30K dataset (Elliot et al., 2016) with 158,915 original Brazilian Portuguese descriptions, and 30,104 Brazilian Portuguese translations from original English descriptions; and ii) adds 2,677,613 frame evocation labels to the 158,915 English descriptions and to the ones created for Brazilian Portuguese; (iii) extends the Flickr30k Entities dataset (Plummer et al., 2015) with 190,608 frames and Frame Elements correlations with the existing phrase-to-region correlations.", }

This paper presents Framed Multi30K (FM30K), a novel frame-based Brazilian Portuguese multimodal-multilingual dataset which i) extends the Multi30K dataset (Elliot et al., 2016) with 158,915 original Brazilian Portuguese descriptions, and 30,104 Brazilian Portuguese translations from original English descriptions; and ii) adds 2,677,613 frame evocation labels to the 158,915 English descriptions and to the ones created for Brazilian Portuguese; (iii) extends the Flickr30k Entities dataset (Plummer et al., 2015) with 190,608 frames and Frame Elements correlations with the existing phrase-to-region correlations.

Framed Multi30K: A Frame-Based Multimodal-Multilingual Dataset

lrec-main.656

Poster

https://aclanthology.org/2024.lrec-main.657.bib

https://aclanthology.org/2024.lrec-main.657/

@inproceedings{zaghir-etal-2024-frasimed, title = "{FRASIMED}: A Clinical {F}rench Annotated Resource Produced through Crosslingual {BERT}-Based Annotation Projection", author = {Zaghir, Jamil and Bjelogrlic, Mina and Goldman, Jean-Philippe and Aananou, Souka{\"\i}na and Gaudet-Blavignac, Christophe and Lovis, Christian}, editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.657", pages = "7450--7460", abstract = "Natural language processing (NLP) applications such as named entity recognition (NER) for low-resource corpora do not benefit from recent advances in the development of large language models (LLMs) where there is still a need for larger annotated datasets. This research article introduces a methodology for generating translated versions of annotated datasets through crosslingual annotation projection and is freely available on GitHub (link: https://github.com/JamilProg/crosslingual{\_}bert{\_}annotation{\_}projection). Leveraging a language agnostic BERT-based approach, it is an efficient solution to increase low-resource corpora with few human efforts and by only using already available open data resources. Quantitative and qualitative evaluations are often lacking when it comes to evaluating the quality and effectiveness of semi-automatic data generation strategies. The evaluation of our crosslingual annotation projection approach showed both effectiveness and high accuracy in the resulting dataset. As a practical application of this methodology, we present the creation of French Annotated Resource with Semantic Information for Medical Entities Detection (FRASIMED), an annotated corpus comprising 2{'}051 synthetic clinical cases in French. The corpus is now available for researchers and practitioners to develop and refine French natural language processing (NLP) applications in the clinical field (https://zenodo.org/record/8355629), making it the largest open annotated corpus with linked medical concepts in French.", }

Natural language processing (NLP) applications such as named entity recognition (NER) for low-resource corpora do not benefit from recent advances in the development of large language models (LLMs) where there is still a need for larger annotated datasets. This research article introduces a methodology for generating translated versions of annotated datasets through crosslingual annotation projection and is freely available on GitHub (link: https://github.com/JamilProg/crosslingual{\_}bert{\_}annotation{\_}projection). Leveraging a language agnostic BERT-based approach, it is an efficient solution to increase low-resource corpora with few human efforts and by only using already available open data resources. Quantitative and qualitative evaluations are often lacking when it comes to evaluating the quality and effectiveness of semi-automatic data generation strategies. The evaluation of our crosslingual annotation projection approach showed both effectiveness and high accuracy in the resulting dataset. As a practical application of this methodology, we present the creation of French Annotated Resource with Semantic Information for Medical Entities Detection (FRASIMED), an annotated corpus comprising 2{'}051 synthetic clinical cases in French. The corpus is now available for researchers and practitioners to develop and refine French natural language processing (NLP) applications in the clinical field (https://zenodo.org/record/8355629), making it the largest open annotated corpus with linked medical concepts in French.

FRASIMED: A Clinical French Annotated Resource Produced through Crosslingual BERT-Based Annotation Projection

lrec-main.657

Poster

2309.10770

https://aclanthology.org/2024.lrec-main.658.bib

https://aclanthology.org/2024.lrec-main.658/

@inproceedings{le-cloirec-ait-yahya-etal-2024-frend, title = "{FR}e{ND}: A {F}rench Resource of Negation Data", author = "Le Cloirec - Ait Yahya, Hafida and Seminck, Olga and Amsili, Pascal", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.658", pages = "7461--7468", abstract = "FReND is a freely available corpus of French language in which negations are hand-annotated. Negations are annotated by their cues and scopes. Comprising 590K tokens and over 8.9K negations, it is the largest dataset available for French. A variety of types of textual genres are covered: literature, blog posts, Wikipedia articles, political debates, clinical reports and newspaper articles. As the understanding of negation is not yet mastered by current state of the art AI-models, FReND is not only a valuable resource for linguistic research into negation, but also as training data for AI tasks such as negation detection.", }

FReND is a freely available corpus of French language in which negations are hand-annotated. Negations are annotated by their cues and scopes. Comprising 590K tokens and over 8.9K negations, it is the largest dataset available for French. A variety of types of textual genres are covered: literature, blog posts, Wikipedia articles, political debates, clinical reports and newspaper articles. As the understanding of negation is not yet mastered by current state of the art AI-models, FReND is not only a valuable resource for linguistic research into negation, but also as training data for AI tasks such as negation detection.

FReND: A French Resource of Negation Data

lrec-main.658

Poster

https://aclanthology.org/2024.lrec-main.659.bib

https://aclanthology.org/2024.lrec-main.659/

@inproceedings{li-etal-2024-graph, title = "From Graph to Word Bag: Introducing Domain Knowledge to Confusing Charge Prediction", author = "Li, Ang and Chen, Qiangchao and Wu, Yiquan and Zhou, Xiang and Kuang, Kun and Wu, Fei and Cai, Ming", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.659", pages = "7469--7479", abstract = "Confusing charge prediction is a challenging task in legal AI, which involves predicting confusing charges based on fact descriptions. While existing charge prediction methods have shown impressive performance, they face significant challenges when dealing with confusing charges, such as Snatch and Robbery. In the legal domain, constituent elements play a pivotal role in distinguishing confusing charges. Constituent elements are fundamental behaviors underlying criminal punishment and have subtle distinctions among charges. In this paper, we introduce a novel From Graph to Word Bag (FWGB) approach, which introduces domain knowledge regarding constituent elements to guide the model in making judgments on confusing charges, much like a judge{'}s reasoning process. Specifically, we first construct a legal knowledge graph containing constituent elements to help select keywords for each charge, forming a word bag. Subsequently, to guide the model{'}s attention towards the differentiating information for each charge within the context, we expand the attention mechanism and introduce a new loss function with attention supervision through words in the word bag. We construct the confusing charges dataset from real-world judicial documents. Experiments demonstrate the effectiveness of our method, especially in maintaining exceptional performance in imbalanced label distributions.", }

Confusing charge prediction is a challenging task in legal AI, which involves predicting confusing charges based on fact descriptions. While existing charge prediction methods have shown impressive performance, they face significant challenges when dealing with confusing charges, such as Snatch and Robbery. In the legal domain, constituent elements play a pivotal role in distinguishing confusing charges. Constituent elements are fundamental behaviors underlying criminal punishment and have subtle distinctions among charges. In this paper, we introduce a novel From Graph to Word Bag (FWGB) approach, which introduces domain knowledge regarding constituent elements to guide the model in making judgments on confusing charges, much like a judge{'}s reasoning process. Specifically, we first construct a legal knowledge graph containing constituent elements to help select keywords for each charge, forming a word bag. Subsequently, to guide the model{'}s attention towards the differentiating information for each charge within the context, we expand the attention mechanism and introduce a new loss function with attention supervision through words in the word bag. We construct the confusing charges dataset from real-world judicial documents. Experiments demonstrate the effectiveness of our method, especially in maintaining exceptional performance in imbalanced label distributions.

From Graph to Word Bag: Introducing Domain Knowledge to Confusing Charge Prediction

lrec-main.659

Poster

2403.04369

https://aclanthology.org/2024.lrec-main.660.bib

https://aclanthology.org/2024.lrec-main.660/

@inproceedings{ponnusamy-etal-2024-laughter, title = "From Laughter to Inequality: Annotated Dataset for Misogyny Detection in {T}amil and {M}alayalam Memes", author = "Ponnusamy, Rahul and Pannerselvam, Kathiravan and R, Saranya and Kumaresan, Prasanna Kumar and Thavareesan, Sajeetha and S, Bhuvaneswari and K.a, Anshid and Kumar, Susminu S and Buitelaar, Paul and Chakravarthi, Bharathi Raja", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.660", pages = "7480--7488", abstract = "In this digital era, memes have become a prevalent online expression, humor, sarcasm, and social commentary. However, beneath their surface lies concerning issues such as the propagation of misogyny, gender-based bias, and harmful stereotypes. To overcome these issues, we introduced MDMD (Misogyny Detection Meme Dataset) in this paper. This article focuses on creating an annotated dataset with detailed annotation guidelines to delve into online misogyny within the Tamil and Malayalam-speaking communities. Through analyzing memes, we uncover the intricate world of gender bias and stereotypes in these communities, shedding light on their manifestations and impact. This dataset, along with its comprehensive annotation guidelines, is a valuable resource for understanding the prevalence, origins, and manifestations of misogyny in various contexts, aiding researchers, policymakers, and organizations in developing effective strategies to combat gender-based discrimination and promote equality and inclusivity. It enables a deeper understanding of the issue and provides insights that can inform strategies for cultivating a more equitable and secure online environment. This work represents a crucial step in raising awareness and addressing gender-based discrimination in the digital space.", }

In this digital era, memes have become a prevalent online expression, humor, sarcasm, and social commentary. However, beneath their surface lies concerning issues such as the propagation of misogyny, gender-based bias, and harmful stereotypes. To overcome these issues, we introduced MDMD (Misogyny Detection Meme Dataset) in this paper. This article focuses on creating an annotated dataset with detailed annotation guidelines to delve into online misogyny within the Tamil and Malayalam-speaking communities. Through analyzing memes, we uncover the intricate world of gender bias and stereotypes in these communities, shedding light on their manifestations and impact. This dataset, along with its comprehensive annotation guidelines, is a valuable resource for understanding the prevalence, origins, and manifestations of misogyny in various contexts, aiding researchers, policymakers, and organizations in developing effective strategies to combat gender-based discrimination and promote equality and inclusivity. It enables a deeper understanding of the issue and provides insights that can inform strategies for cultivating a more equitable and secure online environment. This work represents a crucial step in raising awareness and addressing gender-based discrimination in the digital space.

From Laughter to Inequality: Annotated Dataset for Misogyny Detection in Tamil and Malayalam Memes

lrec-main.660

Poster

https://aclanthology.org/2024.lrec-main.661.bib

https://aclanthology.org/2024.lrec-main.661/

@inproceedings{trajanov-etal-2024-linguistic, title = "From Linguistic Linked Data to Big Data", author = "Trajanov, Dimitar and Apostol, Elena and Garabik, Radovan and Gkirtzou, Katerina and Gromann, Dagmar and Liebeskind, Chaya and Palma, Cosimo and Rosner, Michael and Sampri, Alexia and S{\'e}rasset, Gilles and Spahiu, Blerina and Truic{\u{a}}, Ciprian-Octavian and Valunaite Oleskeviciene, Giedre", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.661", pages = "7489--7502", abstract = "With advances in the field of Linked (Open) Data (LOD), language data on the LOD cloud has grown in number, size, and variety. With an increased volume and variety of language data, optimizations of methods for distributing, storing, and querying these data become more central. To this end, this position paper investigates use cases at the intersection of LLOD and Big Data, existing approaches to utilizing Big Data techniques within the context of linked data, and discusses the challenges and benefits of this union.", }

With advances in the field of Linked (Open) Data (LOD), language data on the LOD cloud has grown in number, size, and variety. With an increased volume and variety of language data, optimizations of methods for distributing, storing, and querying these data become more central. To this end, this position paper investigates use cases at the intersection of LLOD and Big Data, existing approaches to utilizing Big Data techniques within the context of linked data, and discusses the challenges and benefits of this union.

From Linguistic Linked Data to Big Data

lrec-main.661

Poster

https://aclanthology.org/2024.lrec-main.662.bib

https://aclanthology.org/2024.lrec-main.662/

@inproceedings{barta-etal-2024-news, title = "From News to Summaries: Building a {H}ungarian Corpus for Extractive and Abstractive Summarization", author = "Barta, Botond and Lakatos, Dorina and Nagy, Attila and Nyist, Mil{\'a}n Konor and {\'A}cs, Judit", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.662", pages = "7503--7509", abstract = "Training summarization models requires substantial amounts of training data. However for less resourceful languages like Hungarian, openly available models and datasets are notably scarce. To address this gap our paper introduces an open-source Hungarian corpus suitable for training abstractive and extractive summarization models. The dataset is assembled from segments of the Common Crawl corpus undergoing thorough cleaning, preprocessing and deduplication. In addition to abstractive summarization we generate sentence-level labels for extractive summarization using sentence similarity. We train baseline models for both extractive and abstractive summarization using the collected dataset. To demonstrate the effectiveness of the trained models, we perform both quantitative and qualitative evaluation. Our models and dataset will be made publicly available, encouraging replication, further research, and real-world applications across various domains.", }

Training summarization models requires substantial amounts of training data. However for less resourceful languages like Hungarian, openly available models and datasets are notably scarce. To address this gap our paper introduces an open-source Hungarian corpus suitable for training abstractive and extractive summarization models. The dataset is assembled from segments of the Common Crawl corpus undergoing thorough cleaning, preprocessing and deduplication. In addition to abstractive summarization we generate sentence-level labels for extractive summarization using sentence similarity. We train baseline models for both extractive and abstractive summarization using the collected dataset. To demonstrate the effectiveness of the trained models, we perform both quantitative and qualitative evaluation. Our models and dataset will be made publicly available, encouraging replication, further research, and real-world applications across various domains.

From News to Summaries: Building a Hungarian Corpus for Extractive and Abstractive Summarization

lrec-main.662

Poster

2404.03555

https://huggingface.co/papers/2404.03555

https://aclanthology.org/2024.lrec-main.663.bib

https://aclanthology.org/2024.lrec-main.663/

@inproceedings{hazem-etal-2024-technology, title = "From Technology to Market. Bilingual Corpus on the Evaluation of Technology Opportunity Discovery", author = "Hazem, Amir and Motohashi, Kazuyuki and Zhu, Chen", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.663", pages = "7510--7520", abstract = "As companies aim to enhance and expand their product portfolios, Technology Opportunity Discovery (TOD) has gained increasing interest. To comprehend the role of emerging technologies in innovation, we introduce a novel technology-market corpus in English and Japanese languages, and conduct a comprehensive empirical evaluation of the linkage between technology and the market. Our dataset comprises English patents extracted from the USPTO database and Japanese patents from the Japanese Patent Office (JPO), along with their associated products for each stock market company. We compare several static and contextualized word embedding methods to construct a technology-market space and propose an effective methodology based on a fine-tuned BERT model for linking technology to the market.", }

As companies aim to enhance and expand their product portfolios, Technology Opportunity Discovery (TOD) has gained increasing interest. To comprehend the role of emerging technologies in innovation, we introduce a novel technology-market corpus in English and Japanese languages, and conduct a comprehensive empirical evaluation of the linkage between technology and the market. Our dataset comprises English patents extracted from the USPTO database and Japanese patents from the Japanese Patent Office (JPO), along with their associated products for each stock market company. We compare several static and contextualized word embedding methods to construct a technology-market space and propose an effective methodology based on a fine-tuned BERT model for linking technology to the market.

From Technology to Market. Bilingual Corpus on the Evaluation of Technology Opportunity Discovery

lrec-main.663

Poster

https://aclanthology.org/2024.lrec-main.664.bib

https://aclanthology.org/2024.lrec-main.664/

@inproceedings{liang-etal-2024-text, title = "From Text to Historical Ecological Knowledge: The Construction and Application of the {S}han Jing Knowledge Base", author = "Liang, Ke and Huang, Chu-Ren and Jiang, Xin-Lan", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.664", pages = "7521--7530", abstract = "Traditional Ecological Knowledge (TEK) has been recognized as a shared cultural heritage and a crucial instrument to tackle today{'}s environmental challenges. In this paper, we deal with historical ecological knowledge, a special type of TEK that is based on ancient language texts. In particular, we aim to build a language resource based on Shanhai Jing (The Classic of Mountains and Seas). Written 2000 years ago, Shanhai Jing is a record of flora and fauna in ancient China, anchored by mountains (shan) and seas (hai). This study focuses on the entities in the Shan Jing part and builds a knowledge base for them. We adopt a pattern-driven and bottom-up strategy to accommodate two features of the source: highly stylized narrative and juxtaposition of knowledge from multiple domains. The PRF values of both entity and relationship extraction are above 96{\%}. Quality assurance measures like entity disambiguation and resolution were done by domain experts. Neo4j graph database is used to visualize the result. We think the knowledge base, containing 1432 systematically classified entities and 3294 relationships, can provide the foundation for the construction of a historical ecological knowledge base of China. Additionally, the ruled-based text-matching method can be helpful in ancient language processing.", }

Traditional Ecological Knowledge (TEK) has been recognized as a shared cultural heritage and a crucial instrument to tackle today{'}s environmental challenges. In this paper, we deal with historical ecological knowledge, a special type of TEK that is based on ancient language texts. In particular, we aim to build a language resource based on Shanhai Jing (The Classic of Mountains and Seas). Written 2000 years ago, Shanhai Jing is a record of flora and fauna in ancient China, anchored by mountains (shan) and seas (hai). This study focuses on the entities in the Shan Jing part and builds a knowledge base for them. We adopt a pattern-driven and bottom-up strategy to accommodate two features of the source: highly stylized narrative and juxtaposition of knowledge from multiple domains. The PRF values of both entity and relationship extraction are above 96{\%}. Quality assurance measures like entity disambiguation and resolution were done by domain experts. Neo4j graph database is used to visualize the result. We think the knowledge base, containing 1432 systematically classified entities and 3294 relationships, can provide the foundation for the construction of a historical ecological knowledge base of China. Additionally, the ruled-based text-matching method can be helpful in ancient language processing.

From Text to Historical Ecological Knowledge: The Construction and Application of the Shan Jing Knowledge Base

lrec-main.664

Poster

https://aclanthology.org/2024.lrec-main.665.bib

https://aclanthology.org/2024.lrec-main.665/

@inproceedings{antoun-etal-2024-text, title = "From Text to Source: Results in Detecting Large Language Model-Generated Content", author = "Antoun, Wissam and Sagot, Beno{\^\i}t and Seddah, Djam{\'e}", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.665", pages = "7531--7543", abstract = "The widespread use of Large Language Models (LLMs), celebrated for their ability to generate human-like text, has raised concerns about misinformation and ethical implications. Addressing these concerns necessitates the development of robust methods to detect and attribute text generated by LLMs. This paper investigates {``}Cross-Model Detection,{''} by evaluating whether a classifier trained to distinguish between source LLM-generated and human-written text can also detect text from a target LLM without further training. The study comprehensively explores various LLM sizes and families and assesses the impact of conversational fine-tuning techniques, quantization, and watermarking on classifier generalization. The research also explores Model Attribution, encompassing source model identification, model family, and model size classification, in addition to quantization and watermarking detection. Our results reveal several key findings: a clear inverse relationship between classifier effectiveness and model size, with larger LLMs being more challenging to detect, especially when the classifier is trained on data from smaller models. Training on data from similarly sized LLMs can improve detection performance from larger models but may lead to decreased performance when dealing with smaller models. Additionally, model attribution experiments show promising results in identifying source models and model families, highlighting detectable signatures in LLM-generated text, with particularly remarkable outcomes in watermarking detection, while no detectable signatures of quantization were observed. Overall, our study contributes valuable insights into the interplay of model size, family, and training data in LLM detection and attribution.", }

The widespread use of Large Language Models (LLMs), celebrated for their ability to generate human-like text, has raised concerns about misinformation and ethical implications. Addressing these concerns necessitates the development of robust methods to detect and attribute text generated by LLMs. This paper investigates {``}Cross-Model Detection,{''} by evaluating whether a classifier trained to distinguish between source LLM-generated and human-written text can also detect text from a target LLM without further training. The study comprehensively explores various LLM sizes and families and assesses the impact of conversational fine-tuning techniques, quantization, and watermarking on classifier generalization. The research also explores Model Attribution, encompassing source model identification, model family, and model size classification, in addition to quantization and watermarking detection. Our results reveal several key findings: a clear inverse relationship between classifier effectiveness and model size, with larger LLMs being more challenging to detect, especially when the classifier is trained on data from smaller models. Training on data from similarly sized LLMs can improve detection performance from larger models but may lead to decreased performance when dealing with smaller models. Additionally, model attribution experiments show promising results in identifying source models and model families, highlighting detectable signatures in LLM-generated text, with particularly remarkable outcomes in watermarking detection, while no detectable signatures of quantization were observed. Overall, our study contributes valuable insights into the interplay of model size, family, and training data in LLM detection and attribution.

From Text to Source: Results in Detecting Large Language Model-Generated Content

lrec-main.665

Poster

2309.13322

https://aclanthology.org/2024.lrec-main.666.bib

https://aclanthology.org/2024.lrec-main.666/

@inproceedings{titung-alm-2024-fuse, title = "{FUSE} - {F}r{U}stration and Surprise Expressions: A Subtle Emotional Multimodal Language Corpus", author = "Titung, Rajesh and Alm, Cecilia Ovesdotter", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.666", pages = "7544--7555", abstract = "This study introduces a novel multimodal corpus for expressive task-based spoken language and dialogue, focused on language use under frustration and surprise, elicited from three tasks motivated by prior research and collected in an IRB-approved experiment. The resource is unique both because these are understudied affect states for emotion modeling in language, and also because it provides both individual and dyadic multimodally grounded language. The study includes a detailed analysis of annotations and performance results for multimodal emotion inference in language use.", }

This study introduces a novel multimodal corpus for expressive task-based spoken language and dialogue, focused on language use under frustration and surprise, elicited from three tasks motivated by prior research and collected in an IRB-approved experiment. The resource is unique both because these are understudied affect states for emotion modeling in language, and also because it provides both individual and dyadic multimodally grounded language. The study includes a detailed analysis of annotations and performance results for multimodal emotion inference in language use.

FUSE - FrUstration and Surprise Expressions: A Subtle Emotional Multimodal Language Corpus

lrec-main.666

Poster

https://aclanthology.org/2024.lrec-main.667.bib

https://aclanthology.org/2024.lrec-main.667/

@inproceedings{yu-etal-2024-fusion, title = "Fusion-in-T5: Unifying Variant Signals for Simple and Effective Document Ranking with Attention Fusion", author = "Yu, Shi and Fan, Chenghao and Xiong, Chenyan and Jin, David and Liu, Zhiyuan and Liu, Zhenghao", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.667", pages = "7556--7561", abstract = "Common document ranking pipelines in search systems are cascade systems that involve multiple ranking layers to integrate different information step-by-step. In this paper, we propose a novel re-ranker Fusion-in-T5 (FiT5), which integrates text matching information, ranking features, and global document information into one single unified model via templated-based input and global attention. Experiments on passage ranking benchmarks MS MARCO and TREC DL show that FiT5, as one single model, significantly improves ranking performance over complex cascade pipelines. Analysis finds that through attention fusion, FiT5 jointly utilizes various forms of ranking information via gradually attending to related documents and ranking features, and improves the detection of subtle nuances. Our code is open-sourced at https://github.com/OpenMatch/FiT5 . Keywords: document ranking, attention, fusion", }

Common document ranking pipelines in search systems are cascade systems that involve multiple ranking layers to integrate different information step-by-step. In this paper, we propose a novel re-ranker Fusion-in-T5 (FiT5), which integrates text matching information, ranking features, and global document information into one single unified model via templated-based input and global attention. Experiments on passage ranking benchmarks MS MARCO and TREC DL show that FiT5, as one single model, significantly improves ranking performance over complex cascade pipelines. Analysis finds that through attention fusion, FiT5 jointly utilizes various forms of ranking information via gradually attending to related documents and ranking features, and improves the detection of subtle nuances. Our code is open-sourced at https://github.com/OpenMatch/FiT5 . Keywords: document ranking, attention, fusion

Fusion-in-T5: Unifying Variant Signals for Simple and Effective Document Ranking with Attention Fusion

lrec-main.667

Poster

https://aclanthology.org/2024.lrec-main.668.bib

https://aclanthology.org/2024.lrec-main.668/

@inproceedings{jon-bojar-2024-gaatme, title = "{GAATME}: A Genetic Algorithm for Adversarial Translation Metrics Evaluation", author = "Jon, Josef and Bojar, Ond{\v{r}}ej", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.668", pages = "7562--7569", abstract = "Building on a recent method for decoding translation candidates from a Machine Translation (MT) model via a genetic algorithm, we modify it to generate adversarial translations to test and challenge MT evaluation metrics. The produced translations score very well in an arbitrary MT evaluation metric selected beforehand, despite containing serious, deliberately introduced errors. The method can be used to create adversarial test sets to analyze the biases and shortcomings of the metrics. We publish various such test sets for the Czech to English language pair, as well as the code to convert any parallel data into a similar adversarial test set.", }

Building on a recent method for decoding translation candidates from a Machine Translation (MT) model via a genetic algorithm, we modify it to generate adversarial translations to test and challenge MT evaluation metrics. The produced translations score very well in an arbitrary MT evaluation metric selected beforehand, despite containing serious, deliberately introduced errors. The method can be used to create adversarial test sets to analyze the biases and shortcomings of the metrics. We publish various such test sets for the Czech to English language pair, as well as the code to convert any parallel data into a similar adversarial test set.

GAATME: A Genetic Algorithm for Adversarial Translation Metrics Evaluation

lrec-main.668

Poster

https://aclanthology.org/2024.lrec-main.669.bib

https://aclanthology.org/2024.lrec-main.669/

@inproceedings{zhou-etal-2024-gcnet, title = "{GCN}et: Global-and-Context Collaborative Learning for Aspect-Based Sentiment Analysis", author = "Zhou, Ting and Shen, Ying and Li, Yinghui", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.669", pages = "7570--7580", abstract = "Aspect-Based Sentiment Analysis (ABSA) aims to determine the sentiment polarities of specified aspect terms in a sentence. Most previous approaches mainly use an attention mechanism or graph neural networks based on dependency trees to explicitly model the connections between aspect terms and opinion words. However, these methods may not effectively address cases where the sentiment of an aspect term is implicitly described, as the corresponding opinion words may not directly appear in the sentence. To alleviate this issue, in this paper, we propose a GCNet that explicitly leverages global semantic information to guide context encoding. Particularly, we design a semantics encoding module that incorporates global semantic features into sequential modeling process to enable the consideration of the overall sentiment tendency of a sentence, while the global semantic features are also refined by adaptively focusing on different parts of the sentence. Moreover, for a comprehensive sentence analysis, we also include a syntactic feature encoding module along with a pre-fusion module to integrate the refined global features with the syntactic representations. Extensive experiments on three public datasets demonstrate that our model outperforms state-of-the-art methods, indicating the robustness and effectiveness of our approach.", }

Aspect-Based Sentiment Analysis (ABSA) aims to determine the sentiment polarities of specified aspect terms in a sentence. Most previous approaches mainly use an attention mechanism or graph neural networks based on dependency trees to explicitly model the connections between aspect terms and opinion words. However, these methods may not effectively address cases where the sentiment of an aspect term is implicitly described, as the corresponding opinion words may not directly appear in the sentence. To alleviate this issue, in this paper, we propose a GCNet that explicitly leverages global semantic information to guide context encoding. Particularly, we design a semantics encoding module that incorporates global semantic features into sequential modeling process to enable the consideration of the overall sentiment tendency of a sentence, while the global semantic features are also refined by adaptively focusing on different parts of the sentence. Moreover, for a comprehensive sentence analysis, we also include a syntactic feature encoding module along with a pre-fusion module to integrate the refined global features with the syntactic representations. Extensive experiments on three public datasets demonstrate that our model outperforms state-of-the-art methods, indicating the robustness and effectiveness of our approach.

GCNet: Global-and-Context Collaborative Learning for Aspect-Based Sentiment Analysis

lrec-main.669

Poster

https://aclanthology.org/2024.lrec-main.670.bib

https://aclanthology.org/2024.lrec-main.670/

@inproceedings{xie-etal-2024-gecsum, title = "{GECS}um: Generative Evaluation-Driven Sequence Level Contrastive Learning for Abstractive Summarization", author = "Xie, Jiawen and Zhang, Shaoting and Zhang, Xiaofan", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.670", pages = "7581--7595", abstract = "While dominant in abstractive summarization, transformer-based language models with the standard maximum likelihood estimation (MLE) training remain challenged by two discrepancies: the misalignment between token-level training and sequence-level evaluation, and the divergence between teacher-forcing training manner and auto-regressive generation behavior. Recent studies have shown that sequence-level contrastive learning, which utilizes the quality differences between multiple summaries as prior information, can effectively mitigate these issues. However, as certain evaluation metrics often determine the contrastive signals in existing methods, this leads to the model performance aligning with the preferences of these metrics being limited by the evaluation capabilities of these metrics. Inspired by prior works that treat the evaluation of generated text as a text generation problem, we propose a generative evaluation-driven contrastive learning framework, which leverages the semantic understanding capabilities of the abstractive model itself to evaluate summary in reference-based settings. In this way, our method establishes a connection between the model{'}s reference-based evaluation and reference-free generation scenarios, allowing them to share the benefits of model capability enhancements. Extensive experiments on four summarization datasets demonstrate that our method outperforms the previous state-of-the-art regarding comprehensive performance. Various empirical analyses further substantiate the effectiveness of our method.", }

While dominant in abstractive summarization, transformer-based language models with the standard maximum likelihood estimation (MLE) training remain challenged by two discrepancies: the misalignment between token-level training and sequence-level evaluation, and the divergence between teacher-forcing training manner and auto-regressive generation behavior. Recent studies have shown that sequence-level contrastive learning, which utilizes the quality differences between multiple summaries as prior information, can effectively mitigate these issues. However, as certain evaluation metrics often determine the contrastive signals in existing methods, this leads to the model performance aligning with the preferences of these metrics being limited by the evaluation capabilities of these metrics. Inspired by prior works that treat the evaluation of generated text as a text generation problem, we propose a generative evaluation-driven contrastive learning framework, which leverages the semantic understanding capabilities of the abstractive model itself to evaluate summary in reference-based settings. In this way, our method establishes a connection between the model{'}s reference-based evaluation and reference-free generation scenarios, allowing them to share the benefits of model capability enhancements. Extensive experiments on four summarization datasets demonstrate that our method outperforms the previous state-of-the-art regarding comprehensive performance. Various empirical analyses further substantiate the effectiveness of our method.

GECSum: Generative Evaluation-Driven Sequence Level Contrastive Learning for Abstractive Summarization

lrec-main.670

Poster

https://aclanthology.org/2024.lrec-main.671.bib

https://aclanthology.org/2024.lrec-main.671/

@inproceedings{fridriksdottir-einarsson-2024-gendered, title = "Gendered Grammar or Ingrained Bias? Exploring Gender Bias in {I}celandic Language Models", author = "Fri{\dh}riksd{\'o}ttir, Steinunn Rut and Einarsson, Hafsteinn", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.671", pages = "7596--7610", abstract = "Large language models, trained on vast datasets, exhibit increased output quality in proportion to the amount of data that is used to train them. This data-driven learning process has brought forth a pressing issue where these models may not only reflect but also amplify gender bias, racism, religious prejudice, and queerphobia present in their training data that may not always be recent. This study explores gender bias in language models trained on Icelandic, focusing on occupation-related terms. Icelandic is a highly grammatically gendered language that favors the masculine when referring to groups of people with indeterminable genders. Our aim is to explore whether language models merely mirror gender distributions within the corresponding professions or if they exhibit biases tied to their grammatical genders. Results indicate a significant overall predisposition towards the masculine but specific occupation terms consistently lean toward a particular gender, indicating complex interplays of societal and linguistic influences.", }

Large language models, trained on vast datasets, exhibit increased output quality in proportion to the amount of data that is used to train them. This data-driven learning process has brought forth a pressing issue where these models may not only reflect but also amplify gender bias, racism, religious prejudice, and queerphobia present in their training data that may not always be recent. This study explores gender bias in language models trained on Icelandic, focusing on occupation-related terms. Icelandic is a highly grammatically gendered language that favors the masculine when referring to groups of people with indeterminable genders. Our aim is to explore whether language models merely mirror gender distributions within the corresponding professions or if they exhibit biases tied to their grammatical genders. Results indicate a significant overall predisposition towards the masculine but specific occupation terms consistently lean toward a particular gender, indicating complex interplays of societal and linguistic influences.

Gendered Grammar or Ingrained Bias? Exploring Gender Bias in Icelandic Language Models

lrec-main.671

Poster

https://aclanthology.org/2024.lrec-main.672.bib

https://aclanthology.org/2024.lrec-main.672/

@inproceedings{singhal-etal-2024-generating, title = "Generating Clarification Questions for Disambiguating Contracts", author = "Singhal, Anmol and Jain, Chirag and Anish, Preethu Rose and Chakraborty, Arkajyoti and Ghaisas, Smita", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.672", pages = "7611--7622", abstract = "Enterprises frequently enter into commercial contracts that can serve as vital sources of project-specific requirements. Contractual clauses are obligatory, and the requirements derived from contracts can detail the downstream implementation activities that non-legal stakeholders, including requirement analysts, engineers, and delivery personnel, need to conduct. However, comprehending contracts is cognitively demanding and error-prone for such stakeholders due to the extensive use of Legalese and the inherent complexity of contract language. Furthermore, contracts often contain ambiguously worded clauses to ensure comprehensive coverage. In contrast, non-legal stakeholders require a detailed and unambiguous comprehension of contractual clauses to craft actionable requirements. In this work, we introduce a novel legal NLP task that involves generating clarification questions for contracts. These questions aim to identify contract ambiguities on a document level, thereby assisting non-legal stakeholders in obtaining the necessary details for eliciting requirements. This task is challenged by three core issues: (1) data availability, (2) the length and unstructured nature of contracts, and (3) the complexity of legal text. To address these issues, we propose ConRAP, a retrieval-augmented prompting framework for generating clarification questions to disambiguate contractual text. Experiments conducted on contracts sourced from the publicly available CUAD dataset show that ConRAP with ChatGPT can detect ambiguities with an F2 score of 0.87. 70{\%} of the generated clarification questions are deemed useful by human evaluators.", }

Enterprises frequently enter into commercial contracts that can serve as vital sources of project-specific requirements. Contractual clauses are obligatory, and the requirements derived from contracts can detail the downstream implementation activities that non-legal stakeholders, including requirement analysts, engineers, and delivery personnel, need to conduct. However, comprehending contracts is cognitively demanding and error-prone for such stakeholders due to the extensive use of Legalese and the inherent complexity of contract language. Furthermore, contracts often contain ambiguously worded clauses to ensure comprehensive coverage. In contrast, non-legal stakeholders require a detailed and unambiguous comprehension of contractual clauses to craft actionable requirements. In this work, we introduce a novel legal NLP task that involves generating clarification questions for contracts. These questions aim to identify contract ambiguities on a document level, thereby assisting non-legal stakeholders in obtaining the necessary details for eliciting requirements. This task is challenged by three core issues: (1) data availability, (2) the length and unstructured nature of contracts, and (3) the complexity of legal text. To address these issues, we propose ConRAP, a retrieval-augmented prompting framework for generating clarification questions to disambiguate contractual text. Experiments conducted on contracts sourced from the publicly available CUAD dataset show that ConRAP with ChatGPT can detect ambiguities with an F2 score of 0.87. 70{\%} of the generated clarification questions are deemed useful by human evaluators.

Generating Clarification Questions for Disambiguating Contracts

lrec-main.672

Poster

2403.08053

https://aclanthology.org/2024.lrec-main.673.bib

https://aclanthology.org/2024.lrec-main.673/

@inproceedings{mitra-etal-2024-generating, title = "Generating Contextual Images for Long-Form Text", author = "Mitra, Avijit and Gupta, Nalin and Naik, Chetan and Sethy, Abhinav and Bice, Kinsey and Raeesy, Zeynab", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.673", pages = "7623--7633", abstract = "We investigate the problem of synthesizing relevant visual imagery from generic long-form text, leveraging Large Language Models (LLMs) and Text-to-Image Models (TIMs). Current Text-to-Image models require short prompts that describe the image content and style explicitly. Unlike image prompts, generation of images from general long-form text requires the image synthesis system to derive the visual content and style elements from the text. In this paper, we study zero-shot prompting and supervised fine-tuning approaches that use LLMs and TIMs jointly for synthesizing images. We present an empirical study on generating images for Wikipedia articles covering a broad spectrum of topic and image styles. We compare these systems using a suite of metrics, including a novel metric specifically designed to evaluate the semantic correctness of generated images. Our study offers a preliminary understanding of existing models{'} strengths and limitation for the task of image generation from long-form text, and sets up an evaluation framework and establishes baselines for future research.", }

We investigate the problem of synthesizing relevant visual imagery from generic long-form text, leveraging Large Language Models (LLMs) and Text-to-Image Models (TIMs). Current Text-to-Image models require short prompts that describe the image content and style explicitly. Unlike image prompts, generation of images from general long-form text requires the image synthesis system to derive the visual content and style elements from the text. In this paper, we study zero-shot prompting and supervised fine-tuning approaches that use LLMs and TIMs jointly for synthesizing images. We present an empirical study on generating images for Wikipedia articles covering a broad spectrum of topic and image styles. We compare these systems using a suite of metrics, including a novel metric specifically designed to evaluate the semantic correctness of generated images. Our study offers a preliminary understanding of existing models{'} strengths and limitation for the task of image generation from long-form text, and sets up an evaluation framework and establishes baselines for future research.

Generating Contextual Images for Long-Form Text

lrec-main.673

Poster

https://aclanthology.org/2024.lrec-main.674.bib

https://aclanthology.org/2024.lrec-main.674/

@inproceedings{li-etal-2024-generating, title = "Generating Hard-Negative Out-of-Scope Data with {C}hat{GPT} for Intent Classification", author = "Li, Zhijian and Larson, Stefan and Leach, Kevin", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.674", pages = "7634--7646", abstract = "Intent classifiers must be able to distinguish when a user{'}s utterance does not belong to any supported intent to avoid producing incorrect and unrelated system responses. Although out-of-scope (OOS) detection for intent classifiers has been studied, previous work has not yet studied changes in classifier performance against hard-negative out-of-scope utterances (i.e., inputs that share common features with in-scope data, but are actually out-of-scope). We present an automated technique to generate hard-negative OOS data using ChatGPT. We use our technique to build five new hard-negative OOS datasets, and evaluate each against three benchmark intent classifiers. We show that classifiers struggle to correctly identify hard-negative OOS utterances more than general OOS utterances. Finally, we show that incorporating hard-negative OOS data for training improves model robustness when detecting hard-negative OOS data and general OOS data. Our technique, datasets, and evaluation address an important void in the field, offering a straightforward and inexpensive way to collect hard-negative OOS data and improve intent classifiers{'} robustness.", }

Intent classifiers must be able to distinguish when a user{'}s utterance does not belong to any supported intent to avoid producing incorrect and unrelated system responses. Although out-of-scope (OOS) detection for intent classifiers has been studied, previous work has not yet studied changes in classifier performance against hard-negative out-of-scope utterances (i.e., inputs that share common features with in-scope data, but are actually out-of-scope). We present an automated technique to generate hard-negative OOS data using ChatGPT. We use our technique to build five new hard-negative OOS datasets, and evaluate each against three benchmark intent classifiers. We show that classifiers struggle to correctly identify hard-negative OOS utterances more than general OOS utterances. Finally, we show that incorporating hard-negative OOS data for training improves model robustness when detecting hard-negative OOS data and general OOS data. Our technique, datasets, and evaluation address an important void in the field, offering a straightforward and inexpensive way to collect hard-negative OOS data and improve intent classifiers{'} robustness.

Generating Hard-Negative Out-of-Scope Data with ChatGPT for Intent Classification

lrec-main.674

Poster

2403.05640

https://aclanthology.org/2024.lrec-main.675.bib

https://aclanthology.org/2024.lrec-main.675/

@inproceedings{sileo-etal-2024-generating, title = "Generating Multiple-choice Questions for Medical Question Answering with Distractors and Cue-masking", author = "Sileo, Damien and Uma, Kanimozhi and Moens, Marie-Francine", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.675", pages = "7647--7653", abstract = "Medical multiple-choice question answering (MCQA) is a challenging evaluation for medical natural language processing and a helpful task in itself. Medical questions may describe patient symptoms and ask for the correct diagnosis, which requires domain knowledge and complex reasoning. Standard language modeling pretraining alone is not sufficient to achieve the best results with BERT-base size (Devlin et al., 2019) encoders. Jin et al. (2020) showed that focusing masked language modeling on disease name prediction when using medical encyclopedic paragraphs as input leads to considerable MCQA accuracy improvement. In this work, we show that (1) fine-tuning on generated MCQA dataset outperforms the masked language modeling based objective and (2) correctly masking the cues to the answers is critical for good performance. We release new pretraining datasets and achieve state-of-the-art results on 4 MCQA datasets, notably +5.7{\%} with base-size model on MedQA-USMLE.", }

Medical multiple-choice question answering (MCQA) is a challenging evaluation for medical natural language processing and a helpful task in itself. Medical questions may describe patient symptoms and ask for the correct diagnosis, which requires domain knowledge and complex reasoning. Standard language modeling pretraining alone is not sufficient to achieve the best results with BERT-base size (Devlin et al., 2019) encoders. Jin et al. (2020) showed that focusing masked language modeling on disease name prediction when using medical encyclopedic paragraphs as input leads to considerable MCQA accuracy improvement. In this work, we show that (1) fine-tuning on generated MCQA dataset outperforms the masked language modeling based objective and (2) correctly masking the cues to the answers is critical for good performance. We release new pretraining datasets and achieve state-of-the-art results on 4 MCQA datasets, notably +5.7{\%} with base-size model on MedQA-USMLE.

Generating Multiple-choice Questions for Medical Question Answering with Distractors and Cue-masking

lrec-main.675

Poster

2303.07069

https://aclanthology.org/2024.lrec-main.676.bib

https://aclanthology.org/2024.lrec-main.676/

@inproceedings{shi-etal-2024-generative, title = "Generative Multimodal Entity Linking", author = "Shi, Senbao and Xu, Zhenran and Hu, Baotian and Zhang, Min", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.676", pages = "7654--7665", abstract = "Multimodal Entity Linking (MEL) is the task of mapping mentions with multimodal contexts to the referent entities from a knowledge base. Existing MEL methods mainly focus on designing complex multimodal interaction mechanisms and require fine-tuning all model parameters, which can be prohibitively costly and difficult to scale in the era of Large Language Models (LLMs). In this work, we propose GEMEL, a Generative Multimodal Entity Linking framework based on LLMs, which directly generates target entity names. We keep the vision and language model frozen and only train a feature mapper to enable cross-modality interactions. To adapt LLMs to the MEL task, we leverage the in-context learning capability of LLMs by retrieving multimodal instances as demonstrations. Extensive experiments show that, with only ∼0.3{\%} of the model parameters fine-tuned, GEMEL achieves state-of-the-art results on two well-established MEL datasets (7.7{\%} accuracy gains on WikiDiverse and 8.8{\%} accuracy gains on WikiMEL). The performance gain stems from mitigating the popularity bias of LLM predictions and disambiguating less common entities effectively. Further analysis verifies the generality and scalability of GEMEL. Our framework is compatible with any off-the-shelf language model, paving the way towards an efficient and general solution for utilizing LLMs in the MEL task. Our code is available at https://github.com/HITsz-TMG/GEMEL.", }

Multimodal Entity Linking (MEL) is the task of mapping mentions with multimodal contexts to the referent entities from a knowledge base. Existing MEL methods mainly focus on designing complex multimodal interaction mechanisms and require fine-tuning all model parameters, which can be prohibitively costly and difficult to scale in the era of Large Language Models (LLMs). In this work, we propose GEMEL, a Generative Multimodal Entity Linking framework based on LLMs, which directly generates target entity names. We keep the vision and language model frozen and only train a feature mapper to enable cross-modality interactions. To adapt LLMs to the MEL task, we leverage the in-context learning capability of LLMs by retrieving multimodal instances as demonstrations. Extensive experiments show that, with only ∼0.3{\%} of the model parameters fine-tuned, GEMEL achieves state-of-the-art results on two well-established MEL datasets (7.7{\%} accuracy gains on WikiDiverse and 8.8{\%} accuracy gains on WikiMEL). The performance gain stems from mitigating the popularity bias of LLM predictions and disambiguating less common entities effectively. Further analysis verifies the generality and scalability of GEMEL. Our framework is compatible with any off-the-shelf language model, paving the way towards an efficient and general solution for utilizing LLMs in the MEL task. Our code is available at https://github.com/HITsz-TMG/GEMEL.

Generative Multimodal Entity Linking

lrec-main.676

Poster

2306.12725

https://aclanthology.org/2024.lrec-main.677.bib

https://aclanthology.org/2024.lrec-main.677/

@inproceedings{schirmer-etal-2024-gentrac, title = "{GENTRAC}: A Tool for Tracing Trauma in Genocide and Mass Atrocity Court Transcripts", author = "Schirmer, Miriam and Brechenmacher, Christian and Jashari, Endrit and Pfeffer, Juergen", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.677", pages = "7666--7671", abstract = "This paper introduces GENTRAC, an open-access web-based tool built to interactively detect and analyze potentially traumatic content in witness statements of genocide and mass atrocity trials. Harnessing recent developments in natural language processing (NLP) to detect trauma, GENTRAC processes and formats court transcripts for NLP analysis through a sophisticated parsing algorithm and detects the likelihood of traumatic content for each speaker segment. The tool visualizes the density of such content throughout a trial day and provides statistics on the overall amount of traumatic content and speaker distribution. Capable of processing transcripts from four prominent international criminal courts, including the International Criminal Court (ICC), GENTRAC{'}s reach is vast, tailored to handle millions of pages of documents from past and future trials. Detecting potentially re-traumatizing examination methods can enhance the development of trauma-informed legal procedures. GENTRAC also serves as a reliable resource for legal, human rights, and other professionals, aiding their comprehension of mass atrocities{'} emotional toll on survivors.", }

This paper introduces GENTRAC, an open-access web-based tool built to interactively detect and analyze potentially traumatic content in witness statements of genocide and mass atrocity trials. Harnessing recent developments in natural language processing (NLP) to detect trauma, GENTRAC processes and formats court transcripts for NLP analysis through a sophisticated parsing algorithm and detects the likelihood of traumatic content for each speaker segment. The tool visualizes the density of such content throughout a trial day and provides statistics on the overall amount of traumatic content and speaker distribution. Capable of processing transcripts from four prominent international criminal courts, including the International Criminal Court (ICC), GENTRAC{'}s reach is vast, tailored to handle millions of pages of documents from past and future trials. Detecting potentially re-traumatizing examination methods can enhance the development of trauma-informed legal procedures. GENTRAC also serves as a reliable resource for legal, human rights, and other professionals, aiding their comprehension of mass atrocities{'} emotional toll on survivors.

GENTRAC: A Tool for Tracing Trauma in Genocide and Mass Atrocity Court Transcripts

lrec-main.677

Poster

https://aclanthology.org/2024.lrec-main.678.bib

https://aclanthology.org/2024.lrec-main.678/

@inproceedings{dunn-edwards-brown-2024-geographically, title = "Geographically-Informed Language Identification", author = "Dunn, Jonathan and Edwards-Brown, Lane", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.678", pages = "7672--7682", abstract = "This paper develops an approach to language identification in which the set of languages considered by the model depends on the geographic origin of the text in question. Given that many digital corpora can be geo-referenced at the country level, this paper formulates 16 region-specific models, each of which contains the languages expected to appear in countries within that region. These regional models also each include 31 widely-spoken international languages in order to ensure coverage of these linguae francae regardless of location. An upstream evaluation using traditional language identification testing data shows an improvement in f-score ranging from 1.7 points (Southeast Asia) to as much as 10.4 points (North Africa). A downstream evaluation on social media data shows that this improved performance has a significant impact on the language labels which are applied to large real-world corpora. The result is a highly-accurate model that covers 916 languages at a sample size of 50 characters, the performance improved by incorporating geographic information into the model.", }

This paper develops an approach to language identification in which the set of languages considered by the model depends on the geographic origin of the text in question. Given that many digital corpora can be geo-referenced at the country level, this paper formulates 16 region-specific models, each of which contains the languages expected to appear in countries within that region. These regional models also each include 31 widely-spoken international languages in order to ensure coverage of these linguae francae regardless of location. An upstream evaluation using traditional language identification testing data shows an improvement in f-score ranging from 1.7 points (Southeast Asia) to as much as 10.4 points (North Africa). A downstream evaluation on social media data shows that this improved performance has a significant impact on the language labels which are applied to large real-world corpora. The result is a highly-accurate model that covers 916 languages at a sample size of 50 characters, the performance improved by incorporating geographic information into the model.

Geographically-Informed Language Identification

lrec-main.678

Poster

2403.09892

https://aclanthology.org/2024.lrec-main.679.bib

https://aclanthology.org/2024.lrec-main.679/

@inproceedings{schutz-etal-2024-gerdisdetect, title = "{G}er{DISDETECT}: A {G}erman Multilabel Dataset for Disinformation Detection", author = {Sch{\"u}tz, Mina and Pisoiu, Daniela and Liakhovets, Daria and Schindler, Alexander and Siegel, Melanie}, editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.679", pages = "7683--7695", abstract = "Disinformation has become increasingly relevant in recent years both as a political issue and as object of research. Datasets for training machine learning models, especially for other languages than English, are sparse and the creation costly. Annotated datasets often have only binary or multiclass labels, which provide little information about the grounds and system of such classifications. We propose a novel textual dataset GerDISDETECT for German disinformation. To provide comprehensive analytical insights, a fine-grained taxonomy guided annotation scheme is required. The goal of this dataset, instead of providing a direct assessment regarding true or false, is to provide wide-ranging semantic descriptors that allow for complex interpretation as well as inferred decision-making regarding information and trustworthiness of potentially critical articles. This allows this dataset to be also used for other tasks. The dataset was collected in the first three months of 2022 and contains 39 multilabel classes with 5 top-level categories for a total of 1,890 articles: General View (3 labels), Offensive Language (11 labels), Reporting Style (15 labels), Writing Style (6 labels), and Extremism (4 labels). As a baseline, we further pre-trained a multilingual XLM-R model on around 200,000 unlabeled news articles and fine-tuned it for each category.", }

Disinformation has become increasingly relevant in recent years both as a political issue and as object of research. Datasets for training machine learning models, especially for other languages than English, are sparse and the creation costly. Annotated datasets often have only binary or multiclass labels, which provide little information about the grounds and system of such classifications. We propose a novel textual dataset GerDISDETECT for German disinformation. To provide comprehensive analytical insights, a fine-grained taxonomy guided annotation scheme is required. The goal of this dataset, instead of providing a direct assessment regarding true or false, is to provide wide-ranging semantic descriptors that allow for complex interpretation as well as inferred decision-making regarding information and trustworthiness of potentially critical articles. This allows this dataset to be also used for other tasks. The dataset was collected in the first three months of 2022 and contains 39 multilabel classes with 5 top-level categories for a total of 1,890 articles: General View (3 labels), Offensive Language (11 labels), Reporting Style (15 labels), Writing Style (6 labels), and Extremism (4 labels). As a baseline, we further pre-trained a multilingual XLM-R model on around 200,000 unlabeled news articles and fine-tuned it for each category.

GerDISDETECT: A German Multilabel Dataset for Disinformation Detection

lrec-main.679

Poster

https://aclanthology.org/2024.lrec-main.680.bib

https://aclanthology.org/2024.lrec-main.680/

@inproceedings{mascarell-etal-2024-german, title = "{G}erman Also Hallucinates! Inconsistency Detection in News Summaries with the Absinth Dataset", author = "Mascarell, Laura and Chalumattu, Ribin and Rios, Annette", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.680", pages = "7696--7706", abstract = "The advent of Large Language Models (LLMs) has led to remarkable progress on a wide range of natural language processing tasks. Despite the advances, these large-sized models still suffer from hallucinating information in their output, which poses a major issue in automatic text summarization, as we must guarantee that the generated summary is consistent with the content of the source document. Previous research addresses the challenging task of detecting hallucinations in the output (i.e. inconsistency detection) in order to evaluate the faithfulness of the generated summaries. However, these works primarily focus on English and recent multilingual approaches lack German data. This work presents Absinth, a manually annotated dataset for hallucination detection in German news summarization and explores the capabilities of novel open-source LLMs on this task in both fine-tuning and in-context learning settings. We open-source and release the Absinth dataset to foster further research on hallucination detection in German.", }

The advent of Large Language Models (LLMs) has led to remarkable progress on a wide range of natural language processing tasks. Despite the advances, these large-sized models still suffer from hallucinating information in their output, which poses a major issue in automatic text summarization, as we must guarantee that the generated summary is consistent with the content of the source document. Previous research addresses the challenging task of detecting hallucinations in the output (i.e. inconsistency detection) in order to evaluate the faithfulness of the generated summaries. However, these works primarily focus on English and recent multilingual approaches lack German data. This work presents Absinth, a manually annotated dataset for hallucination detection in German news summarization and explores the capabilities of novel open-source LLMs on this task in both fine-tuning and in-context learning settings. We open-source and release the Absinth dataset to foster further research on hallucination detection in German.

German Also Hallucinates! Inconsistency Detection in News Summaries with the Absinth Dataset

lrec-main.680

Poster

2403.03750

https://aclanthology.org/2024.lrec-main.681.bib

https://aclanthology.org/2024.lrec-main.681/

@inproceedings{abrami-etal-2024-german, title = "{G}erman Parliamentary Corpus ({G}er{P}ar{C}or) Reloaded", author = {Abrami, Giuseppe and Bagci, Mevl{\"u}t and Mehler, Alexander}, editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.681", pages = "7707--7716", abstract = "In 2022, the largest German-speaking corpus of parliamentary protocols from three different centuries, on a national and federal level from the countries of Germany, Austria, Switzerland and Liechtenstein, was collected and published - GerParCor. Through GerParCor, it became possible to provide for the first time various parliamentary protocols which were not available digitally and, moreover, could not be retrieved and processed in a uniform manner. Furthermore, GerParCor was additionally preprocessed using NLP methods and made available in XMI format. In this paper, GerParCor is significantly updated by including all new parliamentary protocols in the corpus, as well as adding and preprocessing further parliamentary protocols previously not covered, so that a period up to 1797 is now covered. Besides the integration of a new, state-of-the-art and appropriate NLP preprocessing for the handling of large text corpora, this update also provides an overview of the further reuse of GerParCor by presenting various provisioning capabilities such as API{'}s, among others.", }

In 2022, the largest German-speaking corpus of parliamentary protocols from three different centuries, on a national and federal level from the countries of Germany, Austria, Switzerland and Liechtenstein, was collected and published - GerParCor. Through GerParCor, it became possible to provide for the first time various parliamentary protocols which were not available digitally and, moreover, could not be retrieved and processed in a uniform manner. Furthermore, GerParCor was additionally preprocessed using NLP methods and made available in XMI format. In this paper, GerParCor is significantly updated by including all new parliamentary protocols in the corpus, as well as adding and preprocessing further parliamentary protocols previously not covered, so that a period up to 1797 is now covered. Besides the integration of a new, state-of-the-art and appropriate NLP preprocessing for the handling of large text corpora, this update also provides an overview of the further reuse of GerParCor by presenting various provisioning capabilities such as API{'}s, among others.

German Parliamentary Corpus (GerParCor) Reloaded

lrec-main.681

Poster

https://aclanthology.org/2024.lrec-main.682.bib

https://aclanthology.org/2024.lrec-main.682/

@inproceedings{konca-etal-2024-german, title = "{G}erman {SRL}: Corpus Construction and Model Training", author = "Konca, Maxim and Luecking, Andy and Mehler, Alexander", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.682", pages = "7717--7727", abstract = "A useful semantic role-annotated resource for training semantic role models for the German language is missing. We point out some problems of previous resources and provide a new one due to a combined translation and alignment process: The gold standard CoNLL-2012 semantic role annotations are translated into German. Semantic role labels are transferred due to alignment models. The resulting dataset is used to train a German semantic role model. With F1-scores around 0.7, the major roles achieve competitive evaluation scores, but avoid limitations of previous approaches. The described procedure can be applied to other languages as well.", }

A useful semantic role-annotated resource for training semantic role models for the German language is missing. We point out some problems of previous resources and provide a new one due to a combined translation and alignment process: The gold standard CoNLL-2012 semantic role annotations are translated into German. Semantic role labels are transferred due to alignment models. The resulting dataset is used to train a German semantic role model. With F1-scores around 0.7, the major roles achieve competitive evaluation scores, but avoid limitations of previous approaches. The described procedure can be applied to other languages as well.

German SRL: Corpus Construction and Model Training

lrec-main.682

Poster

https://aclanthology.org/2024.lrec-main.683.bib

https://aclanthology.org/2024.lrec-main.683/

@inproceedings{krenn-etal-2024-germs, title = "{GERMS}-{AT}: A Sexism/Misogyny Dataset of Forum Comments from an {A}ustrian Online Newspaper", author = "Krenn, Brigitte and Petrak, Johann and Kubina, Marina and Burger, Christian", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.683", pages = "7728--7739", abstract = "Brigitte Krenn, Johann Petrak, Marina Kubina, Christian Burger This paper presents a sexism/misogyny dataset extracted from comments of a large online forum of an Austrian newspaper. The comments are in Austrian German language, and in some cases interspersed with dialectal or English elements. We describe the data collection, the annotation guidelines and the annotation process resulting in a corpus of approximately 8 000 comments which were annotated with 5 levels of sexism/misogyny, ranging from 0 (not sexist/misogynist) to 4 (highly sexist/misogynist). The professional forum moderators (self-identified females and males) of the online newspaper were involved as experts in the creation of the annotation guidelines and the annotation of the user comments. In addition, we also describe first results of training transformer-based classification models for both binarized and original label classification of the corpus.", }

Brigitte Krenn, Johann Petrak, Marina Kubina, Christian Burger This paper presents a sexism/misogyny dataset extracted from comments of a large online forum of an Austrian newspaper. The comments are in Austrian German language, and in some cases interspersed with dialectal or English elements. We describe the data collection, the annotation guidelines and the annotation process resulting in a corpus of approximately 8 000 comments which were annotated with 5 levels of sexism/misogyny, ranging from 0 (not sexist/misogynist) to 4 (highly sexist/misogynist). The professional forum moderators (self-identified females and males) of the online newspaper were involved as experts in the creation of the annotation guidelines and the annotation of the user comments. In addition, we also describe first results of training transformer-based classification models for both binarized and original label classification of the corpus.

GERMS-AT: A Sexism/Misogyny Dataset of Forum Comments from an Austrian Online Newspaper

lrec-main.683

Poster

https://aclanthology.org/2024.lrec-main.684.bib

https://aclanthology.org/2024.lrec-main.684/

@inproceedings{dick-etal-2024-gil, title = "{GIL}-{GAL}a{D}: Gender Inclusive Language - {G}erman Auto-Assembled Large Database", author = "Dick, Anna-Katharina and Drews, Matthias and Pickard, Valentin and Pierz, Victoria", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.684", pages = "7740--7745", abstract = "As the need for gender-inclusive language has become a highly debated topic over the years, gendered biases in speech are unfortunately often picked up and propagated by modern language models trained on large amounts of text. While remedial efforts are underway, grammatically gendered languages such as German pose some unique challenges in generating gender-inclusive language for corrective model training or fine-tuning. We assembled GIL-GALaD, a corpus of German gender-inclusive language from different sources such as social media, news articles, public speeches and academic publications. Our corpus includes the most common types of modifications of generic masculine forms of nouns and spans 30 years (1993-2023), containing over 800,000 instances of gender-inclusive language. Tools for corpus usage and extension are to be included in the release. During corpus assembly, we were also able to gain some insights into which types of gender-inclusive language were used in practice throughout the years and across different domains.", }

As the need for gender-inclusive language has become a highly debated topic over the years, gendered biases in speech are unfortunately often picked up and propagated by modern language models trained on large amounts of text. While remedial efforts are underway, grammatically gendered languages such as German pose some unique challenges in generating gender-inclusive language for corrective model training or fine-tuning. We assembled GIL-GALaD, a corpus of German gender-inclusive language from different sources such as social media, news articles, public speeches and academic publications. Our corpus includes the most common types of modifications of generic masculine forms of nouns and spans 30 years (1993-2023), containing over 800,000 instances of gender-inclusive language. Tools for corpus usage and extension are to be included in the release. During corpus assembly, we were also able to gain some insights into which types of gender-inclusive language were used in practice throughout the years and across different domains.

GIL-GALaD: Gender Inclusive Language - German Auto-Assembled Large Database

lrec-main.684

Poster

https://aclanthology.org/2024.lrec-main.685.bib

https://aclanthology.org/2024.lrec-main.685/

@inproceedings{tu-etal-2024-glamr, title = "{GLAMR}: Augmenting {AMR} with {GL}-{V}erb{N}et Event Structure", author = "Tu, Jingxuan and Obiso, Timothy and Ye, Bingyang and Rim, Kyeongmin and Xu, Keer and Yue, Liulu and Brown, Susan Windisch and Palmer, Martha and Pustejovsky, James", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.685", pages = "7746--7759", abstract = "This paper introduces GLAMR, an Abstract Meaning Representation (AMR) interpretation of Generative Lexicon (GL) semantic components. It includes a structured subeventual interpretation of linguistic predicates, and encoding of the opposition structure of property changes of event arguments. Both of these features are recently encoded in VerbNet (VN), and form the scaffolding for the semantic form associated with VN frame files. We develop a new syntax, concepts, and roles for subevent structure based on VN for connecting subevents to atomic predicates. Our proposed extension is compatible with current AMR specification. We also present an approach to automatically augment AMR graphs by inserting subevent structure of the predicates and identifying the subevent arguments from the semantic roles. A pilot annotation of GLAMR graphs of 65 documents (486 sentences), based on procedural texts as a source, is presented as a public dataset. The annotation includes subevents, argument property change, and document-level anaphoric links. Finally, we provide baseline models for converting text to GLAMR and vice versa, along with the application of GLAMR for generating enriched paraphrases with details on subevent transformation and arguments that are not present in the surface form of the texts.", }

This paper introduces GLAMR, an Abstract Meaning Representation (AMR) interpretation of Generative Lexicon (GL) semantic components. It includes a structured subeventual interpretation of linguistic predicates, and encoding of the opposition structure of property changes of event arguments. Both of these features are recently encoded in VerbNet (VN), and form the scaffolding for the semantic form associated with VN frame files. We develop a new syntax, concepts, and roles for subevent structure based on VN for connecting subevents to atomic predicates. Our proposed extension is compatible with current AMR specification. We also present an approach to automatically augment AMR graphs by inserting subevent structure of the predicates and identifying the subevent arguments from the semantic roles. A pilot annotation of GLAMR graphs of 65 documents (486 sentences), based on procedural texts as a source, is presented as a public dataset. The annotation includes subevents, argument property change, and document-level anaphoric links. Finally, we provide baseline models for converting text to GLAMR and vice versa, along with the application of GLAMR for generating enriched paraphrases with details on subevent transformation and arguments that are not present in the surface form of the texts.

GLAMR: Augmenting AMR with GL-VerbNet Event Structure

lrec-main.685

Poster

https://aclanthology.org/2024.lrec-main.686.bib

https://aclanthology.org/2024.lrec-main.686/

@inproceedings{zeng-etal-2024-global, title = "Global and Local Hierarchical Prompt Tuning Framework for Multi-level Implicit Discourse Relation Recognition", author = "Zeng, Lei and He, Ruifang and Sun, Haowen and Xu, Jing and Liu, Chang and Wang, Bo", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.686", pages = "7760--7773", abstract = "Multi-level implicit discourse relation recognition (MIDRR) is a challenging task to recognize the hierarchical discourse relations between the arguments with the absence of connectives. Recent methods tend to incorporate the static hierarchical structure containing all senses (defined as global hierarchy) into prompt tuning through a path prompt template or hierarchical label refining. Howerver, hierarchical modeling is independent of the verbalizer, resulting in a failure to effectively utilize the output probability distribution information of verbalizer. Besides, they ignore the utilization of the dynamic hierarchical label sequence for each instance (defined as local hierarchy) in prompt tuning. In this paper, we propose a global and local hierarchical prompt tuning (GLHPT) framework, which utilize prior knowledge of PLMs while better incorporating hierarchical information from two aspects. We leverage bottom-up propagated probability as the global hierarchy to inject it into multi-level verbalizer (MLV). Furthermore, we design a local hierarchy-driven contrastive learning (LHCL) to improve the probability distribution of MLV. Finally, our model achieves competitive results on two benchmacks.", }

Multi-level implicit discourse relation recognition (MIDRR) is a challenging task to recognize the hierarchical discourse relations between the arguments with the absence of connectives. Recent methods tend to incorporate the static hierarchical structure containing all senses (defined as global hierarchy) into prompt tuning through a path prompt template or hierarchical label refining. Howerver, hierarchical modeling is independent of the verbalizer, resulting in a failure to effectively utilize the output probability distribution information of verbalizer. Besides, they ignore the utilization of the dynamic hierarchical label sequence for each instance (defined as local hierarchy) in prompt tuning. In this paper, we propose a global and local hierarchical prompt tuning (GLHPT) framework, which utilize prior knowledge of PLMs while better incorporating hierarchical information from two aspects. We leverage bottom-up propagated probability as the global hierarchy to inject it into multi-level verbalizer (MLV). Furthermore, we design a local hierarchy-driven contrastive learning (LHCL) to improve the probability distribution of MLV. Finally, our model achieves competitive results on two benchmacks.

Global and Local Hierarchical Prompt Tuning Framework for Multi-level Implicit Discourse Relation Recognition

lrec-main.686

Poster

https://aclanthology.org/2024.lrec-main.687.bib

https://aclanthology.org/2024.lrec-main.687/

@inproceedings{kargaran-etal-2024-glotscript, title = "{G}lot{S}cript: A Resource and Tool for Low Resource Writing System Identification", author = {Kargaran, Amir Hossein and Yvon, Fran{\c{c}}ois and Sch{\"u}tze, Hinrich}, editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.687", pages = "7774--7784", abstract = "We present GlotScript, an open resource and tool for low resource writing system identification. GlotScript-R is a resource that provides the attested writing systems for more than 7,000 languages. It is compiled by aggregating information from existing writing system resources. GlotScript-T is a writing system identification tool that covers all 161 Unicode 15.0 scripts. For an input text, it returns its script distribution where scripts are identified by ISO 15924 codes. We also present two use cases for GlotScript. First, we demonstrate that GlotScript can help cleaning multilingual corpora such as mC4 and OSCAR. Second, we analyze the tokenization of a number of language models such as GPT-4 using GlotScript and provide insights on the coverage of low resource scripts and languages by each language model. We hope that GlotScript will become a useful resource for work on low resource languages in the NLP community. GlotScript-R and GlotScript-T are available at https://github.com/cisnlp/GlotScript.", }

We present GlotScript, an open resource and tool for low resource writing system identification. GlotScript-R is a resource that provides the attested writing systems for more than 7,000 languages. It is compiled by aggregating information from existing writing system resources. GlotScript-T is a writing system identification tool that covers all 161 Unicode 15.0 scripts. For an input text, it returns its script distribution where scripts are identified by ISO 15924 codes. We also present two use cases for GlotScript. First, we demonstrate that GlotScript can help cleaning multilingual corpora such as mC4 and OSCAR. Second, we analyze the tokenization of a number of language models such as GPT-4 using GlotScript and provide insights on the coverage of low resource scripts and languages by each language model. We hope that GlotScript will become a useful resource for work on low resource languages in the NLP community. GlotScript-R and GlotScript-T are available at https://github.com/cisnlp/GlotScript.

GlotScript: A Resource and Tool for Low Resource Writing System Identification

lrec-main.687

Poster

2309.13320

https://huggingface.co/papers/2309.13320

https://aclanthology.org/2024.lrec-main.688.bib

https://aclanthology.org/2024.lrec-main.688/

@inproceedings{lopez-cortez-etal-2024-gmeg, title = "{GMEG}-{EXP}: A Dataset of Human- and {LLM}-Generated Explanations of Grammatical and Fluency Edits", author = "L{\'o}pez Cortez, S. Magal{\'\i} and Norris, Mark Josef and Duman, Steve", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.688", pages = "7785--7800", abstract = "Recent work has explored the ability of large language models (LLMs) to generate explanations of existing labeled data. In this work, we investigate the ability of LLMs to explain revisions in sentences. We introduce a new dataset demonstrating a novel task, which we call explaining text revisions. We collected human- and LLM-generated explanations of grammatical and fluency edits and defined criteria for the human evaluation of the explanations along three dimensions: Coverage, Informativeness, and Correctness. The results of a side-by-side evaluation show an Overall preference for human explanations, but there are many instances in which annotators show no preference. Annotators prefer human-generated explanations for Informativeness and Correctness, but they show no preference for Coverage. We also examined the extent to which the number of revisions in a sentence influences annotators{'} Overall preference for the explanations. We found that the preference for human explanations increases as the number of revisions in the sentence increases. Additionally, we show that the Overall preference for human explanations depends on the type of error being explained. We discuss explanation styles based on a qualitative analysis of 300 explanations. We release our dataset and annotation guidelines to encourage future research.", }

Recent work has explored the ability of large language models (LLMs) to generate explanations of existing labeled data. In this work, we investigate the ability of LLMs to explain revisions in sentences. We introduce a new dataset demonstrating a novel task, which we call explaining text revisions. We collected human- and LLM-generated explanations of grammatical and fluency edits and defined criteria for the human evaluation of the explanations along three dimensions: Coverage, Informativeness, and Correctness. The results of a side-by-side evaluation show an Overall preference for human explanations, but there are many instances in which annotators show no preference. Annotators prefer human-generated explanations for Informativeness and Correctness, but they show no preference for Coverage. We also examined the extent to which the number of revisions in a sentence influences annotators{'} Overall preference for the explanations. We found that the preference for human explanations increases as the number of revisions in the sentence increases. Additionally, we show that the Overall preference for human explanations depends on the type of error being explained. We discuss explanation styles based on a qualitative analysis of 300 explanations. We release our dataset and annotation guidelines to encourage future research.

GMEG-EXP: A Dataset of Human- and LLM-Generated Explanations of Grammatical and Fluency Edits

lrec-main.688

Poster

https://aclanthology.org/2024.lrec-main.689.bib

https://aclanthology.org/2024.lrec-main.689/

@inproceedings{yarlott-etal-2024-golem, title = "{GOLEM}: {GO}ld Standard for Learning and Evaluation of Motifs", author = "Yarlott, W. Victor and Acharya, Anurag and Castro Estrada, Diego and Gomez, Diana and Finlayson, Mark", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.689", pages = "7801--7813", abstract = "Motifs are distinctive, recurring, widely used idiom-like words or phrases, often originating from folklore, whose meaning are anchored in a narrative. Motifs have significance as communicative devices because they concisely imply a constellation of culturally relevant information. Their broad usage suggests their cognitive importance as touchstones of cultural knowledge. We present GOLEM, the first dataset annotated for motific information. The dataset comprises 7,955 English articles (2,039,424 words). The corpus identifies 26,078 motif candidates across 34 motif types from three cultural or national groups: Jewish, Irish, and Puerto Rican. Each motif candidate is labeled with the type of usage (Motific, Referential, Eponymic, or Unrelated), resulting in 1,723 actual motific instances. Annotation was performed by individuals identifying as members of each group and achieved a Fleiss{'} kappa of {\textgreater}0.55. We demonstrate that classification of candidate type is a challenging task for LLMs using a few-shot approach; recent models such as T5, FLAN-T5, GPT-2, and Llama 2 (7B) achieved a performance of 41{\%} accuracy at best. These data will support development of new models and approaches for detecting (and reasoning about) motific information in text. We release the corpus, the annotation guide, and the code to support other researchers building on this work.", }

Motifs are distinctive, recurring, widely used idiom-like words or phrases, often originating from folklore, whose meaning are anchored in a narrative. Motifs have significance as communicative devices because they concisely imply a constellation of culturally relevant information. Their broad usage suggests their cognitive importance as touchstones of cultural knowledge. We present GOLEM, the first dataset annotated for motific information. The dataset comprises 7,955 English articles (2,039,424 words). The corpus identifies 26,078 motif candidates across 34 motif types from three cultural or national groups: Jewish, Irish, and Puerto Rican. Each motif candidate is labeled with the type of usage (Motific, Referential, Eponymic, or Unrelated), resulting in 1,723 actual motific instances. Annotation was performed by individuals identifying as members of each group and achieved a Fleiss{'} kappa of {\textgreater}0.55. We demonstrate that classification of candidate type is a challenging task for LLMs using a few-shot approach; recent models such as T5, FLAN-T5, GPT-2, and Llama 2 (7B) achieved a performance of 41{\%} accuracy at best. These data will support development of new models and approaches for detecting (and reasoning about) motific information in text. We release the corpus, the annotation guide, and the code to support other researchers building on this work.

GOLEM: GOld Standard for Learning and Evaluation of Motifs

lrec-main.689

Poster

https://aclanthology.org/2024.lrec-main.690.bib

https://aclanthology.org/2024.lrec-main.690/

@inproceedings{debess-etal-2024-good, title = "Good or Bad News? Exploring {GPT}-4 for Sentiment Analysis for {F}aroese on a Public News Corpora", author = "Debess, Iben Nyholm and Simonsen, Annika and Einarsson, Hafsteinn", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.690", pages = "7814--7824", abstract = "Sentiment analysis in low-resource languages presents unique challenges that Large Language Models may help address. This study explores the efficacy of GPT-4 for sentiment analysis on Faroese news texts, an uncharted task for this language. On the basis of guidelines presented, the sentiment analysis was performed with a multi-class approach at the sentence and document level with 225 sentences analysed in 170 articles. When comparing GPT-4 to human annotators, we observe that GPT-4 performs remarkably well. We explored two prompt configurations and observed a benefit from having clear instructions for the sentiment analysis task, but no benefit from translating the articles to English before the sentiment analysis task. Our results indicate that GPT-4 can be considered as a valuable tool for generating Faroese test data. Furthermore, our investigation reveals the intricacy of news sentiment. This motivates a more nuanced approach going forward, and we suggest a multi-label approach for future research in this domain. We further explored the efficacy of GPT-4 in topic classification on news texts and observed more negative sentiments expressed in international than national news. Overall, this work demonstrates GPT-4{'}s proficiency on a novel task and its utility for augmenting resources in low-data languages.", }

Sentiment analysis in low-resource languages presents unique challenges that Large Language Models may help address. This study explores the efficacy of GPT-4 for sentiment analysis on Faroese news texts, an uncharted task for this language. On the basis of guidelines presented, the sentiment analysis was performed with a multi-class approach at the sentence and document level with 225 sentences analysed in 170 articles. When comparing GPT-4 to human annotators, we observe that GPT-4 performs remarkably well. We explored two prompt configurations and observed a benefit from having clear instructions for the sentiment analysis task, but no benefit from translating the articles to English before the sentiment analysis task. Our results indicate that GPT-4 can be considered as a valuable tool for generating Faroese test data. Furthermore, our investigation reveals the intricacy of news sentiment. This motivates a more nuanced approach going forward, and we suggest a multi-label approach for future research in this domain. We further explored the efficacy of GPT-4 in topic classification on news texts and observed more negative sentiments expressed in international than national news. Overall, this work demonstrates GPT-4{'}s proficiency on a novel task and its utility for augmenting resources in low-data languages.

Good or Bad News? Exploring GPT-4 for Sentiment Analysis for Faroese on a Public News Corpora

lrec-main.690

Poster

https://aclanthology.org/2024.lrec-main.691.bib

https://aclanthology.org/2024.lrec-main.691/

@inproceedings{verdonik-etal-2024-gos, title = "Gos 2: A New Reference Corpus of Spoken {S}lovenian", author = "Verdonik, Darinka and Dobrovoljc, Kaja and Erjavec, Toma{\v{z}} and Ljube{\v{s}}i{\'c}, Nikola", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.691", pages = "7825--7830", abstract = "This paper introduces a new version of the Gos reference corpus of spoken Slovenian, which was recently extended to more than double the original size (300 hours, 2.4 million words) by adding speech recordings and transcriptions from two related initiatives, the Gos VideoLectures corpus of public academic speech, and the Artur speech recognition database. We describe this process by first presenting the criteria guiding the balanced selection of the newly added data and the challenges encountered when merging language resources with divergent designs, followed by the presentation of other major enhancements of the new Gos corpus, such as improvements in lemmatization and morphosyntactic annotation, word-level speech alignment, a new XML schema and the development of a specialized online concordancer.", }

This paper introduces a new version of the Gos reference corpus of spoken Slovenian, which was recently extended to more than double the original size (300 hours, 2.4 million words) by adding speech recordings and transcriptions from two related initiatives, the Gos VideoLectures corpus of public academic speech, and the Artur speech recognition database. We describe this process by first presenting the criteria guiding the balanced selection of the newly added data and the challenges encountered when merging language resources with divergent designs, followed by the presentation of other major enhancements of the new Gos corpus, such as improvements in lemmatization and morphosyntactic annotation, word-level speech alignment, a new XML schema and the development of a specialized online concordancer.

Gos 2: A New Reference Corpus of Spoken Slovenian

lrec-main.691

Poster

https://aclanthology.org/2024.lrec-main.692.bib

https://aclanthology.org/2024.lrec-main.692/

@inproceedings{katinskaia-yangarber-2024-gpt, title = "{GPT}-3.5 for Grammatical Error Correction", author = "Katinskaia, Anisia and Yangarber, Roman", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.692", pages = "7831--7843", abstract = "This paper investigates the application of GPT-3.5 for Grammatical Error Correction (GEC) in multiple languages in several settings: zero-shot GEC, fine-tuning for GEC, and using GPT-3.5 to re-rank correction hypotheses generated by other GEC models. In the zero-shot setting, we conduct automatic evaluations of the corrections proposed by GPT-3.5 using several methods: estimating grammaticality with language models (LMs), the Scribendy test, and comparing the semantic embeddings of sentences. GPT-3.5 has a known tendency to over-correct erroneous sentences and propose alternative corrections. For several languages, such as Czech, German, Russian, Spanish, and Ukrainian, GPT-3.5 substantially alters the source sentences, including their semantics, which presents significant challenges for evaluation with reference-based metrics. For English, GPT-3.5 demonstrates high recall, generates fluent corrections, and generally preserves sentence semantics. However, human evaluation for both English and Russian reveals that, despite its strong error-detection capabilities, GPT-3.5 struggles with several error types, including punctuation mistakes, tense errors, syntactic dependencies between words, and lexical compatibility at the sentence level.", }

This paper investigates the application of GPT-3.5 for Grammatical Error Correction (GEC) in multiple languages in several settings: zero-shot GEC, fine-tuning for GEC, and using GPT-3.5 to re-rank correction hypotheses generated by other GEC models. In the zero-shot setting, we conduct automatic evaluations of the corrections proposed by GPT-3.5 using several methods: estimating grammaticality with language models (LMs), the Scribendy test, and comparing the semantic embeddings of sentences. GPT-3.5 has a known tendency to over-correct erroneous sentences and propose alternative corrections. For several languages, such as Czech, German, Russian, Spanish, and Ukrainian, GPT-3.5 substantially alters the source sentences, including their semantics, which presents significant challenges for evaluation with reference-based metrics. For English, GPT-3.5 demonstrates high recall, generates fluent corrections, and generally preserves sentence semantics. However, human evaluation for both English and Russian reveals that, despite its strong error-detection capabilities, GPT-3.5 struggles with several error types, including punctuation mistakes, tense errors, syntactic dependencies between words, and lexical compatibility at the sentence level.

GPT-3.5 for Grammatical Error Correction

lrec-main.692

Poster

2405.08469

https://aclanthology.org/2024.lrec-main.693.bib

https://aclanthology.org/2024.lrec-main.693/

@inproceedings{mao-etal-2024-gpteval, title = "{GPTE}val: A Survey on Assessments of {C}hat{GPT} and {GPT}-4", author = "Mao, Rui and Chen, Guanyi and Zhang, Xulang and Guerin, Frank and Cambria, Erik", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.693", pages = "7844--7866", abstract = "The emergence of ChatGPT has generated much speculation in the press about its potential to disrupt social and economic systems. Its astonishing language ability has aroused strong curiosity among scholars about its performance in different domains. There have been many studies evaluating the ability of ChatGPT and GPT-4 in different tasks and disciplines. However, a comprehensive review summarizing the collective assessment findings is lacking. The objective of this survey is to thoroughly analyze prior assessments of ChatGPT and GPT-4, focusing on its language and reasoning abilities, scientific knowledge, and ethical considerations. Furthermore, an examination of the existing evaluation methods is conducted, offering several recommendations for future research.", }

The emergence of ChatGPT has generated much speculation in the press about its potential to disrupt social and economic systems. Its astonishing language ability has aroused strong curiosity among scholars about its performance in different domains. There have been many studies evaluating the ability of ChatGPT and GPT-4 in different tasks and disciplines. However, a comprehensive review summarizing the collective assessment findings is lacking. The objective of this survey is to thoroughly analyze prior assessments of ChatGPT and GPT-4, focusing on its language and reasoning abilities, scientific knowledge, and ethical considerations. Furthermore, an examination of the existing evaluation methods is conducted, offering several recommendations for future research.

GPTEval: A Survey on Assessments of ChatGPT and GPT-4

lrec-main.693

Poster

2308.12488

https://aclanthology.org/2024.lrec-main.694.bib

https://aclanthology.org/2024.lrec-main.694/

@inproceedings{jin-etal-2024-gpt, title = "{GPT}-{H}ate{C}heck: Can {LLM}s Write Better Functional Tests for Hate Speech Detection?", author = "Jin, Yiping and Wanner, Leo and Shvets, Alexander", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.694", pages = "7867--7885", abstract = "Online hate detection suffers from biases incurred in data sampling, annotation, and model pre-training. Therefore, measuring the averaged performance over all examples in held-out test data is inadequate. Instead, we must identify specific model weaknesses and be informed when it is more likely to fail. A recent proposal in this direction is HateCheck, a suite for testing fine-grained model functionalities on synthesized data generated using templates of the kind {``}You are just a [slur] to me.{''} However, despite enabling more detailed diagnostic insights, the HateCheck test cases are often generic and have simplistic sentence structures that do not match the real-world data. To address this limitation, we propose GPT-HateCheck, a framework to generate more diverse and realistic functional tests from scratch by instructing large language models (LLMs). We employ an additional natural language inference (NLI) model to verify the generations. Crowd-sourced annotation demonstrates that the generated test cases are of high quality. Using the new functional tests, we can uncover model weaknesses that would be overlooked using the original HateCheck dataset.", }

Online hate detection suffers from biases incurred in data sampling, annotation, and model pre-training. Therefore, measuring the averaged performance over all examples in held-out test data is inadequate. Instead, we must identify specific model weaknesses and be informed when it is more likely to fail. A recent proposal in this direction is HateCheck, a suite for testing fine-grained model functionalities on synthesized data generated using templates of the kind {``}You are just a [slur] to me.{''} However, despite enabling more detailed diagnostic insights, the HateCheck test cases are often generic and have simplistic sentence structures that do not match the real-world data. To address this limitation, we propose GPT-HateCheck, a framework to generate more diverse and realistic functional tests from scratch by instructing large language models (LLMs). We employ an additional natural language inference (NLI) model to verify the generations. Crowd-sourced annotation demonstrates that the generated test cases are of high quality. Using the new functional tests, we can uncover model weaknesses that would be overlooked using the original HateCheck dataset.

GPT-HateCheck: Can LLMs Write Better Functional Tests for Hate Speech Detection?

lrec-main.694

Poster

2402.15238

https://aclanthology.org/2024.lrec-main.695.bib

https://aclanthology.org/2024.lrec-main.695/

@inproceedings{ekgren-etal-2024-gpt, title = "{GPT}-{SW}3: An Autoregressive Language Model for the {S}candinavian Languages", author = {Ekgren, Ariel and Cuba Gyllensten, Amaru and Stollenwerk, Felix and {\"O}hman, Joey and Isbister, Tim and Gogoulou, Evangelia and Carlsson, Fredrik and Casademont, Judit and Sahlgren, Magnus}, editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.695", pages = "7886--7900", abstract = "This paper details the process of developing the first native large generative language model for the North Germanic languages, GPT-SW3. We cover all parts of the development process, from data collection and processing, training configuration and instruction finetuning, to evaluation, applications, and considerations for release strategies. We discuss pros and cons of developing large language models for smaller languages and in relatively peripheral regions of the globe, and we hope that this paper can serve as a guide and reference for other researchers that undertake the development of large generative models for smaller languages.", }

This paper details the process of developing the first native large generative language model for the North Germanic languages, GPT-SW3. We cover all parts of the development process, from data collection and processing, training configuration and instruction finetuning, to evaluation, applications, and considerations for release strategies. We discuss pros and cons of developing large language models for smaller languages and in relatively peripheral regions of the globe, and we hope that this paper can serve as a guide and reference for other researchers that undertake the development of large generative models for smaller languages.

GPT-SW3: An Autoregressive Language Model for the Scandinavian Languages

lrec-main.695

Poster

https://aclanthology.org/2024.lrec-main.696.bib

https://aclanthology.org/2024.lrec-main.696/

@inproceedings{huo-etal-2024-gradient, title = "Gradient Consistency-based Parameter Allocation for Multilingual Neural Machine Translation", author = "Huo, Wenshuai and Feng, Xiaocheng and Huang, Yichong and Fu, Chengpeng and Wang, Hui and Qin, Bing", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.696", pages = "7901--7912", abstract = "Multilingual neural machine translation handles the translation of multiple languages with one unified model. However, this joint-training paradigm incurs the notorious issue of parameter interference, where the model compromises with the language diversity to find a common solution. Recent research has explored avoiding this problem by selecting certain parameters for each language direction from the original model to form language-specific sub-networks. However, determining how many parameters to choose and which parameters to select is still a serious challenge. In this work, we propose an approach called CaPA (Consistency-based Parameter Allocation), which dynamically allocates parameters of appropriate scale to each language direction based on the consistency between the gradient of the individual language and the average gradient. Specifically, CaPA allocates more parameters to languages with higher gradient consistency as these languages tend to have a more positive impact on other languages. Furthermore, considering the varying levels of interference across different parts of the model, we propose an adaptive parameter allocation based on module-level gradient consistency. Experimental results show the correlation between gradient consistency and parameter interference, as well as the effectiveness of our proposed method.", }

Multilingual neural machine translation handles the translation of multiple languages with one unified model. However, this joint-training paradigm incurs the notorious issue of parameter interference, where the model compromises with the language diversity to find a common solution. Recent research has explored avoiding this problem by selecting certain parameters for each language direction from the original model to form language-specific sub-networks. However, determining how many parameters to choose and which parameters to select is still a serious challenge. In this work, we propose an approach called CaPA (Consistency-based Parameter Allocation), which dynamically allocates parameters of appropriate scale to each language direction based on the consistency between the gradient of the individual language and the average gradient. Specifically, CaPA allocates more parameters to languages with higher gradient consistency as these languages tend to have a more positive impact on other languages. Furthermore, considering the varying levels of interference across different parts of the model, we propose an adaptive parameter allocation based on module-level gradient consistency. Experimental results show the correlation between gradient consistency and parameter interference, as well as the effectiveness of our proposed method.

Gradient Consistency-based Parameter Allocation for Multilingual Neural Machine Translation

lrec-main.696

Poster

https://aclanthology.org/2024.lrec-main.697.bib

https://aclanthology.org/2024.lrec-main.697/

@inproceedings{littell-etal-2024-gramble, title = "Gramble: A Tabular Programming Language for Collaborative Linguistic Modeling", author = "Littell, Patrick and Stewart, Darlene and Davis, Fineen and Pine, Aidan and Kuhn, Roland", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.697", pages = "7913--7925", abstract = "We introduce Gramble, a domain-specific programming language for linguistic parsing and generation, in the tradition of XFST, TWOLC, and Kleene. Gramble features an intuitive tabular syntax and supports live group programming, allowing community experts to participate more directly in system development without having to be programmers themselves. A cross-platform interpreter is available for Windows, MacOS, and UNIX, supports collaborative programming on the web via Google Sheets, and is released open-source under the MIT license.", }

We introduce Gramble, a domain-specific programming language for linguistic parsing and generation, in the tradition of XFST, TWOLC, and Kleene. Gramble features an intuitive tabular syntax and supports live group programming, allowing community experts to participate more directly in system development without having to be programmers themselves. A cross-platform interpreter is available for Windows, MacOS, and UNIX, supports collaborative programming on the web via Google Sheets, and is released open-source under the MIT license.

Gramble: A Tabular Programming Language for Collaborative Linguistic Modeling

lrec-main.697

Poster

https://aclanthology.org/2024.lrec-main.698.bib

https://aclanthology.org/2024.lrec-main.698/

@inproceedings{chan-etal-2024-grammatical, title = "Grammatical Error Correction for Code-Switched Sentences by Learners of {E}nglish", author = "Chan, Kelvin Wey Han and Bryant, Christopher and Nguyen, Li and Caines, Andrew and Yuan, Zheng", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.698", pages = "7926--7938", abstract = "Code-switching (CSW) is a common phenomenon among multilingual speakers where multiple languages are used in a single discourse or utterance. Mixed language utterances may still contain grammatical errors however, yet most existing Grammar Error Correction (GEC) systems have been trained on monolingual data and not developed with CSW in mind. In this work, we conduct the first exploration into the use of GEC systems on CSW text. Through this exploration, we propose a novel method of generating synthetic CSW GEC datasets by translating different spans of text within existing GEC corpora. We then investigate different methods of selecting these spans based on CSW ratio, switch-point factor and linguistic constraints, and identify how they affect the performance of GEC systems on CSW text. Our best model achieves an average increase of 1.57 F0.5 across 3 CSW test sets (English-Chinese, English-Korean and English-Japanese) without affecting the model{'}s performance on a monolingual dataset. We furthermore discovered that models trained on one CSW language generalise relatively well to other typologically similar CSW languages.", }

Code-switching (CSW) is a common phenomenon among multilingual speakers where multiple languages are used in a single discourse or utterance. Mixed language utterances may still contain grammatical errors however, yet most existing Grammar Error Correction (GEC) systems have been trained on monolingual data and not developed with CSW in mind. In this work, we conduct the first exploration into the use of GEC systems on CSW text. Through this exploration, we propose a novel method of generating synthetic CSW GEC datasets by translating different spans of text within existing GEC corpora. We then investigate different methods of selecting these spans based on CSW ratio, switch-point factor and linguistic constraints, and identify how they affect the performance of GEC systems on CSW text. Our best model achieves an average increase of 1.57 F0.5 across 3 CSW test sets (English-Chinese, English-Korean and English-Japanese) without affecting the model{'}s performance on a monolingual dataset. We furthermore discovered that models trained on one CSW language generalise relatively well to other typologically similar CSW languages.

Grammatical Error Correction for Code-Switched Sentences by Learners of English

lrec-main.698

Poster

2404.12489

https://aclanthology.org/2024.lrec-main.699.bib

https://aclanthology.org/2024.lrec-main.699/

@inproceedings{aksu-chen-2024-granular, title = "Granular Change Accuracy: A More Accurate Performance Metric for Dialogue State Tracking", author = "Aksu, Taha and Chen, Nancy", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.699", pages = "7939--7948", abstract = "Current metrics for evaluating Dialogue State Tracking (DST) systems exhibit three primary limitations. They: i) erroneously presume a uniform distribution of slots throughout the dialog, ii) neglect to assign partial scores for individual turns, iii) frequently overestimate or underestimate performance by repeatedly counting the models{'} successful or failed predictions. To address these shortcomings, we introduce a novel metric: Granular Change Accuracy (GCA). GCA focuses on evaluating the predicted changes in dialogue state over the entire dialogue history. Benchmarking reveals that GCA effectively reduces biases arising from distribution uniformity and the positioning of errors across turns, resulting in a more precise evaluation. Notably, we find that these biases are particularly pronounced when evaluating few-shot or zero-shot trained models, becoming even more evident as the model{'}s error rate increases. Hence, GCA offers significant promise, particularly for assessing models trained with limited resources. Our GCA implementation is a useful addition to the pool of DST metrics.", }

Current metrics for evaluating Dialogue State Tracking (DST) systems exhibit three primary limitations. They: i) erroneously presume a uniform distribution of slots throughout the dialog, ii) neglect to assign partial scores for individual turns, iii) frequently overestimate or underestimate performance by repeatedly counting the models{'} successful or failed predictions. To address these shortcomings, we introduce a novel metric: Granular Change Accuracy (GCA). GCA focuses on evaluating the predicted changes in dialogue state over the entire dialogue history. Benchmarking reveals that GCA effectively reduces biases arising from distribution uniformity and the positioning of errors across turns, resulting in a more precise evaluation. Notably, we find that these biases are particularly pronounced when evaluating few-shot or zero-shot trained models, becoming even more evident as the model{'}s error rate increases. Hence, GCA offers significant promise, particularly for assessing models trained with limited resources. Our GCA implementation is a useful addition to the pool of DST metrics.

Granular Change Accuracy: A More Accurate Performance Metric for Dialogue State Tracking

lrec-main.699

Poster

2403.11123

https://aclanthology.org/2024.lrec-main.700.bib

https://aclanthology.org/2024.lrec-main.700/

@inproceedings{evdaimon-etal-2024-greekbart, title = "{G}reek{BART}: The First Pretrained {G}reek Sequence-to-Sequence Model", author = "Evdaimon, Iakovos and Abdine, Hadi and Xypolopoulos, Christos and Outsios, Stamatis and Vazirgiannis, Michalis and Stamou, Giorgos", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.700", pages = "7949--7962", abstract = "The era of transfer learning has revolutionized the fields of Computer Vision and Natural Language Processing, bringing powerful pretrained models with exceptional performance across a variety of tasks. Specifically, Natural Language Processing tasks have been dominated by transformer-based language models. In Natural Language Inference and Natural Language Generation tasks, the BERT model and its variants, as well as the GPT model and its successors, demonstrated exemplary performance. However, the majority of these models are pretrained and assessed primarily for the English language or on a multilingual corpus. In this paper, we introduce GreekBART, the first Seq2Seq model based on BART-base architecture and pretrained on a large-scale Greek corpus. We evaluate and compare GreekBART against BART-random, Greek-BERT, and XLM-R on a variety of discriminative tasks. In addition, we examine its performance on two NLG tasks from GreekSUM, a newly introduced summarization dataset for the Greek language. The model, the code, and the new summarization dataset will be publicly available.", }

The era of transfer learning has revolutionized the fields of Computer Vision and Natural Language Processing, bringing powerful pretrained models with exceptional performance across a variety of tasks. Specifically, Natural Language Processing tasks have been dominated by transformer-based language models. In Natural Language Inference and Natural Language Generation tasks, the BERT model and its variants, as well as the GPT model and its successors, demonstrated exemplary performance. However, the majority of these models are pretrained and assessed primarily for the English language or on a multilingual corpus. In this paper, we introduce GreekBART, the first Seq2Seq model based on BART-base architecture and pretrained on a large-scale Greek corpus. We evaluate and compare GreekBART against BART-random, Greek-BERT, and XLM-R on a variety of discriminative tasks. In addition, we examine its performance on two NLG tasks from GreekSUM, a newly introduced summarization dataset for the Greek language. The model, the code, and the new summarization dataset will be publicly available.

GreekBART: The First Pretrained Greek Sequence-to-Sequence Model

lrec-main.700

Poster

2304.00869

https://huggingface.co/papers/2304.00869