Datasets:

NAACL

/

NAACL2024-papers

like 0

Follow

North American Chapter of the Association for Computational Linguistics 1

https://aclanthology.org/2024.bea-1.15.bib

https://aclanthology.org/2024.bea-1.15/

@inproceedings{koutcheme-etal-2024-using, title = "Using Program Repair as a Proxy for Language Models{'} Feedback Ability in Programming Education", author = "Koutcheme, Charles and Dainese, Nicola and Hellas, Arto", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.15", pages = "165--181", abstract = "One of the key challenges in programming education is being able to provide high-quality feedback to learners. Such feedback often includes explanations of the issues in students{'} programs coupled with suggestions on how to fix these issues. Large language models (LLMs) have recently emerged as valuable tools that can help in this effort. In this article, we explore the relationship between the program repair ability of LLMs and their proficiency in providing natural language explanations of coding mistakes. We outline a benchmarking study that evaluates leading LLMs (including open-source ones) on program repair and explanation tasks. Our experiments study the capabilities of LLMs both on a course level and on a programming concept level, allowing us to assess whether the programming concepts practised in exercises with faulty student programs relate to the performance of the models. Our results highlight that LLMs proficient in repairing student programs tend to provide more complete and accurate natural language explanations of code issues. Overall, these results enhance our understanding of the role and capabilities of LLMs in programming education. Using program repair as a proxy for explanation evaluation opens the door for cost-effective assessment methods.", }

One of the key challenges in programming education is being able to provide high-quality feedback to learners. Such feedback often includes explanations of the issues in students{'} programs coupled with suggestions on how to fix these issues. Large language models (LLMs) have recently emerged as valuable tools that can help in this effort. In this article, we explore the relationship between the program repair ability of LLMs and their proficiency in providing natural language explanations of coding mistakes. We outline a benchmarking study that evaluates leading LLMs (including open-source ones) on program repair and explanation tasks. Our experiments study the capabilities of LLMs both on a course level and on a programming concept level, allowing us to assess whether the programming concepts practised in exercises with faulty student programs relate to the performance of the models. Our results highlight that LLMs proficient in repairing student programs tend to provide more complete and accurate natural language explanations of code issues. Overall, these results enhance our understanding of the role and capabilities of LLMs in programming education. Using program repair as a proxy for explanation evaluation opens the door for cost-effective assessment methods.

Using Program Repair as a Proxy for Language Models' Feedback Ability in Programming Education

bea-1.15

Poster

https://aclanthology.org/2024.bea-1.16.bib

https://aclanthology.org/2024.bea-1.16/

@inproceedings{ilagan-etal-2024-automated, title = "Automated Evaluation of Teacher Encouragement of Student-to-Student Interactions in a Simulated Classroom Discussion", author = "Ilagan, Michael and Beigman Klebanov, Beata and Mikeska, Jamie", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.16", pages = "182--198", abstract = "Leading students to engage in argumentation-focused discussions is a challenge for elementary school teachers, as doing so requires facilitating group discussions with student-to-student interaction. The Mystery Powder (MP) Task was designed to be used in online simulated classrooms to develop teachers{'} skill in facilitating small group science discussions. In order to provide timely and scaleable feedback to teachers facilitating a discussion in the simulated classroom, we employ a hybrid modeling approach that successfully combines fine-tuned large language models with features capturing important elements of the discourse dynamic to evaluate MP discussion transcripts. To our knowledge, this is the first application of a hybrid model to automate evaluation of teacher discourse.", }

Leading students to engage in argumentation-focused discussions is a challenge for elementary school teachers, as doing so requires facilitating group discussions with student-to-student interaction. The Mystery Powder (MP) Task was designed to be used in online simulated classrooms to develop teachers{'} skill in facilitating small group science discussions. In order to provide timely and scaleable feedback to teachers facilitating a discussion in the simulated classroom, we employ a hybrid modeling approach that successfully combines fine-tuned large language models with features capturing important elements of the discourse dynamic to evaluate MP discussion transcripts. To our knowledge, this is the first application of a hybrid model to automate evaluation of teacher discourse.

Automated Evaluation of Teacher Encouragement of Student-to-Student Interactions in a Simulated Classroom Discussion

bea-1.16

Poster

https://aclanthology.org/2024.bea-1.17.bib

https://aclanthology.org/2024.bea-1.17/

@inproceedings{ribeiro-flucht-etal-2024-explainable, title = "Explainable {AI} in Language Learning: Linking Empirical Evidence and Theoretical Concepts in Proficiency and Readability Modeling of {P}ortuguese", author = "Ribeiro-Flucht, Luisa and Chen, Xiaobin and Meurers, Detmar", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.17", pages = "199--209", abstract = "While machine learning methods have supported significantly improved results in education research, a common deficiency lies in the explainability of the result. Explainable AI (XAI) aims to fill that gap by providing transparent, conceptually understandable explanations for the classification decisions, enhancing human comprehension and trust in the outcomes. This paper explores an XAI approach to proficiency and readability assessment employing a comprehensive set of 465 linguistic complexity measures. We identify theoretical descriptions associating such measures with varying levels of proficiency and readability and validate them using cross-corpus experiments employing supervised machine learning and Shapley Additive Explanations. The results not only highlight the utility of a diverse set of complexity measures in effectively modeling proficiency and readability in Portuguese, achieving a state-of-the-art accuracy of 0.70 in the proficiency classification task and of 0.84 in the readability classification task, but they largely corroborate the theoretical research assumptions, especially in the lexical domain.", }

While machine learning methods have supported significantly improved results in education research, a common deficiency lies in the explainability of the result. Explainable AI (XAI) aims to fill that gap by providing transparent, conceptually understandable explanations for the classification decisions, enhancing human comprehension and trust in the outcomes. This paper explores an XAI approach to proficiency and readability assessment employing a comprehensive set of 465 linguistic complexity measures. We identify theoretical descriptions associating such measures with varying levels of proficiency and readability and validate them using cross-corpus experiments employing supervised machine learning and Shapley Additive Explanations. The results not only highlight the utility of a diverse set of complexity measures in effectively modeling proficiency and readability in Portuguese, achieving a state-of-the-art accuracy of 0.70 in the proficiency classification task and of 0.84 in the readability classification task, but they largely corroborate the theoretical research assumptions, especially in the lexical domain.

Explainable AI in Language Learning: Linking Empirical Evidence and Theoretical Concepts in Proficiency and Readability Modeling of Portuguese

bea-1.17

Poster

https://aclanthology.org/2024.bea-1.18.bib

https://aclanthology.org/2024.bea-1.18/

@inproceedings{schaller-etal-2024-fairness, title = "Fairness in Automated Essay Scoring: A Comparative Analysis of Algorithms on {G}erman Learner Essays from Secondary Education", author = "Schaller, Nils-Jonathan and Ding, Yuning and Horbach, Andrea and Meyer, Jennifer and Jansen, Thorben", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.18", pages = "210--221", abstract = "Pursuing educational equity, particularly in writing instruction, requires that all students receive fair (i.e., accurate and unbiased) assessment and feedback on their texts. Automated Essay Scoring (AES) algorithms have so far focused on optimizing the mean accuracy of their scores and paid less attention to fair scores for all subgroups, although research shows that students receive unfair scores on their essays in relation to demographic variables, which in turn are related to their writing competence. We add to the literature arguing that AES should also optimize for fairness by presenting insights on the fairness of scoring algorithms on a corpus of learner texts in the German language and introduce the novelty of examining fairness on psychological and demographic differences in addition to demographic differences. We compare shallow learning, deep learning, and large language models with full and skewed subsets of training data to investigate what is needed for fair scoring. The results show that training on a skewed subset of higher and lower cognitive ability students shows no bias but very low accuracy for students outside the training set. Our results highlight the need for specific training data on all relevant user groups, not only for demographic background variables but also for cognitive abilities as psychological student characteristics.", }

Pursuing educational equity, particularly in writing instruction, requires that all students receive fair (i.e., accurate and unbiased) assessment and feedback on their texts. Automated Essay Scoring (AES) algorithms have so far focused on optimizing the mean accuracy of their scores and paid less attention to fair scores for all subgroups, although research shows that students receive unfair scores on their essays in relation to demographic variables, which in turn are related to their writing competence. We add to the literature arguing that AES should also optimize for fairness by presenting insights on the fairness of scoring algorithms on a corpus of learner texts in the German language and introduce the novelty of examining fairness on psychological and demographic differences in addition to demographic differences. We compare shallow learning, deep learning, and large language models with full and skewed subsets of training data to investigate what is needed for fair scoring. The results show that training on a skewed subset of higher and lower cognitive ability students shows no bias but very low accuracy for students outside the training set. Our results highlight the need for specific training data on all relevant user groups, not only for demographic background variables but also for cognitive abilities as psychological student characteristics.

Fairness in Automated Essay Scoring: A Comparative Analysis of Algorithms on German Learner Essays from Secondary Education

bea-1.18

Poster

https://aclanthology.org/2024.bea-1.19.bib

https://aclanthology.org/2024.bea-1.19/

@inproceedings{scarlatos-etal-2024-improving, title = "Improving Automated Distractor Generation for Math Multiple-choice Questions with Overgenerate-and-rank", author = "Scarlatos, Alexander and Feng, Wanyong and Lan, Andrew and Woodhead, Simon and Smith, Digory", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.19", pages = "222--231", abstract = "Multiple-choice questions (MCQs) are commonly used across all levels of math education since they can be deployed and graded at a large scale. A critical component of MCQs is the distractors, i.e., incorrect answers crafted to reflect student errors or misconceptions. Automatically generating them in math MCQs, e.g., with large language models, has been challenging. In this work, we propose a novel method to enhance the quality of generated distractors through overgenerate-and-rank, training a ranking model to predict how likely distractors are to be selected by real students. Experimental results on a real-world dataset and human evaluation with math teachers show that our ranking model increases alignment with human-authored distractors, although human-authored ones are still preferred over generated ones.", }

Multiple-choice questions (MCQs) are commonly used across all levels of math education since they can be deployed and graded at a large scale. A critical component of MCQs is the distractors, i.e., incorrect answers crafted to reflect student errors or misconceptions. Automatically generating them in math MCQs, e.g., with large language models, has been challenging. In this work, we propose a novel method to enhance the quality of generated distractors through overgenerate-and-rank, training a ranking model to predict how likely distractors are to be selected by real students. Experimental results on a real-world dataset and human evaluation with math teachers show that our ranking model increases alignment with human-authored distractors, although human-authored ones are still preferred over generated ones.

Improving Automated Distractor Generation for Math Multiple-choice Questions with Overgenerate-and-rank

bea-1.19

Poster

2405.05144

https://aclanthology.org/2024.bea-1.20.bib

https://aclanthology.org/2024.bea-1.20/

@inproceedings{stowe-2024-identifying, title = "Identifying Fairness Issues in Automatically Generated Testing Content", author = "Stowe, Kevin and Longwill, Benny and Francis, Alyssa and Aoyama, Tatsuya and Ghosh, Debanjan and Somasundaran, Swapna", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.20", pages = "232--250", abstract = "Natural language generation tools are powerful and effective for generating content. However, language models are known to display bias and fairness issues, making them impractical to deploy for many use cases. We here focus on how fairness issues impact automatically generated test content, which can have stringent requirements to ensure the test measures only what it was intended to measure. Specifically, we review test content generated for a large-scale standardized English proficiency test with the goal of identifying content that only pertains to a certain subset of the test population as well as content that has the potential to be upsetting or distracting to some test takers. Issues like these could inadvertently impact a test taker{'}s score and thus should be avoided. This kind of content does not reflect the more commonly-acknowledged biases, making it challenging even for modern models that contain safeguards. We build a dataset of 601 generated texts annotated for fairness and explore a variety of methods for classification: fine-tuning, topic-based classification, and prompting, including few-shot and self-correcting prompts. We find that combining prompt self-correction and few-shot learning performs best, yielding an F1 score of 0.79 on our held-out test set, while much smaller BERT- and topic-based models have competitive performance on out-of-domain data.", }

Natural language generation tools are powerful and effective for generating content. However, language models are known to display bias and fairness issues, making them impractical to deploy for many use cases. We here focus on how fairness issues impact automatically generated test content, which can have stringent requirements to ensure the test measures only what it was intended to measure. Specifically, we review test content generated for a large-scale standardized English proficiency test with the goal of identifying content that only pertains to a certain subset of the test population as well as content that has the potential to be upsetting or distracting to some test takers. Issues like these could inadvertently impact a test taker{'}s score and thus should be avoided. This kind of content does not reflect the more commonly-acknowledged biases, making it challenging even for modern models that contain safeguards. We build a dataset of 601 generated texts annotated for fairness and explore a variety of methods for classification: fine-tuning, topic-based classification, and prompting, including few-shot and self-correcting prompts. We find that combining prompt self-correction and few-shot learning performs best, yielding an F1 score of 0.79 on our held-out test set, while much smaller BERT- and topic-based models have competitive performance on out-of-domain data.

Identifying Fairness Issues in Automatically Generated Testing Content

bea-1.20

Poster

2404.15104

https://aclanthology.org/2024.bea-1.21.bib

https://aclanthology.org/2024.bea-1.21/

@inproceedings{mita-etal-2024-towards, title = "Towards Automated Document Revision: Grammatical Error Correction, Fluency Edits, and Beyond", author = "Mita, Masato and Sakaguchi, Keisuke and Hagiwara, Masato and Mizumoto, Tomoya and Suzuki, Jun and Inui, Kentaro", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.21", pages = "251--265", abstract = "Natural language processing (NLP) technology has rapidly improved automated grammatical error correction (GEC) tasks, and the GEC community has begun to explore document-level revision. However, there are two major obstacles to going beyond automated \textit{sentence-level} GEC to NLP-based document-level revision support: (1) there are few public corpora with document-level revisions annotated by professional editors, and (2) it is infeasible to obtain all possible references and evaluate revision quality using such references because there are infinite revision possibilities. To address these challenges, this paper proposes a new document revision corpus, Text Revision of ACL papers (TETRA), in which multiple professional editors have revised academic papers sampled from the ACL anthology. This corpus enables us to focus on document-level and paragraph-level edits, such as edits related to coherence and consistency. Additionally, as a case study using the TETRA corpus, we investigate reference-less and interpretable methods for meta-evaluation to detect quality improvements according to document revisions. We show the uniqueness of TETRA compared with existing document revision corpora and demonstrate that a fine-tuned pre-trained language model can discriminate the quality of documents after revision even when the difference is subtle.", }

Natural language processing (NLP) technology has rapidly improved automated grammatical error correction (GEC) tasks, and the GEC community has begun to explore document-level revision. However, there are two major obstacles to going beyond automated \textit{sentence-level} GEC to NLP-based document-level revision support: (1) there are few public corpora with document-level revisions annotated by professional editors, and (2) it is infeasible to obtain all possible references and evaluate revision quality using such references because there are infinite revision possibilities. To address these challenges, this paper proposes a new document revision corpus, Text Revision of ACL papers (TETRA), in which multiple professional editors have revised academic papers sampled from the ACL anthology. This corpus enables us to focus on document-level and paragraph-level edits, such as edits related to coherence and consistency. Additionally, as a case study using the TETRA corpus, we investigate reference-less and interpretable methods for meta-evaluation to detect quality improvements according to document revisions. We show the uniqueness of TETRA compared with existing document revision corpora and demonstrate that a fine-tuned pre-trained language model can discriminate the quality of documents after revision even when the difference is subtle.

Towards Automated Document Revision: Grammatical Error Correction, Fluency Edits, and Beyond

bea-1.21

Poster

2205.11484

https://aclanthology.org/2024.bea-1.22.bib

https://aclanthology.org/2024.bea-1.22/

@inproceedings{durward-thomson-2024-evaluating, title = "Evaluating Vocabulary Usage in {LLM}s", author = "Durward, Matthew and Thomson, Christopher", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.22", pages = "266--282", abstract = "The paper focuses on investigating vocabulary usage for AI and human-generated text. We define vocabulary usage in two ways: structural differences and keyword differences. Structural differences are evaluated by converting text into Vocabulary-Managment Profiles, initially used for discourse analysis. Through VMPs, we can treat the text data as a time series, allowing an evaluation by implementing Dynamic time-warping distance measures and subsequently deriving similarity scores to provide an indication of whether the structural dynamics in AI texts resemble human texts. To analyze keywords, we use a measure that emphasizes frequency and dispersion to source {`}key{'} keywords. A qualitative approach is then applied, noting thematic differences between human and AI writing.", }

The paper focuses on investigating vocabulary usage for AI and human-generated text. We define vocabulary usage in two ways: structural differences and keyword differences. Structural differences are evaluated by converting text into Vocabulary-Managment Profiles, initially used for discourse analysis. Through VMPs, we can treat the text data as a time series, allowing an evaluation by implementing Dynamic time-warping distance measures and subsequently deriving similarity scores to provide an indication of whether the structural dynamics in AI texts resemble human texts. To analyze keywords, we use a measure that emphasizes frequency and dispersion to source {`}key{'} keywords. A qualitative approach is then applied, noting thematic differences between human and AI writing.

Evaluating Vocabulary Usage in LLMs

bea-1.22

Poster

https://aclanthology.org/2024.bea-1.23.bib

https://aclanthology.org/2024.bea-1.23/

@inproceedings{stahl-etal-2024-exploring, title = "Exploring {LLM} Prompting Strategies for Joint Essay Scoring and Feedback Generation", author = "Stahl, Maja and Biermann, Leon and Nehring, Andreas and Wachsmuth, Henning", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.23", pages = "283--298", abstract = "Individual feedback can help students improve their essay writing skills. However, the manual effort required to provide such feedback limits individualization in practice. Automatically-generated essay feedback may serve as an alternative to guide students at their own pace, convenience, and desired frequency. Large language models (LLMs) have demonstrated strong performance in generating coherent and contextually relevant text. Yet, their ability to provide helpful essay feedback is unclear. This work explores several prompting strategies for LLM-based zero-shot and few-shot generation of essay feedback. Inspired by Chain-of-Thought prompting, we study how and to what extent automated essay scoring (AES) can benefit the quality of generated feedback. We evaluate both the AES performance that LLMs can achieve with prompting only and the helpfulness of the generated essay feedback. Our results suggest that tackling AES and feedback generation jointly improves AES performance. However, while our manual evaluation emphasizes the quality of the generated essay feedback, the impact of essay scoring on the generated feedback remains low ultimately.", }

Individual feedback can help students improve their essay writing skills. However, the manual effort required to provide such feedback limits individualization in practice. Automatically-generated essay feedback may serve as an alternative to guide students at their own pace, convenience, and desired frequency. Large language models (LLMs) have demonstrated strong performance in generating coherent and contextually relevant text. Yet, their ability to provide helpful essay feedback is unclear. This work explores several prompting strategies for LLM-based zero-shot and few-shot generation of essay feedback. Inspired by Chain-of-Thought prompting, we study how and to what extent automated essay scoring (AES) can benefit the quality of generated feedback. We evaluate both the AES performance that LLMs can achieve with prompting only and the helpfulness of the generated essay feedback. Our results suggest that tackling AES and feedback generation jointly improves AES performance. However, while our manual evaluation emphasizes the quality of the generated essay feedback, the impact of essay scoring on the generated feedback remains low ultimately.

Exploring LLM Prompting Strategies for Joint Essay Scoring and Feedback Generation

bea-1.23

Poster

2404.15845

https://aclanthology.org/2024.bea-1.24.bib

https://aclanthology.org/2024.bea-1.24/

@inproceedings{glandorf-meurers-2024-towards, title = "Towards Fine-Grained Pedagogical Control over {E}nglish Grammar Complexity in Educational Text Generation", author = "Glandorf, Dominik and Meurers, Detmar", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.24", pages = "299--308", abstract = "Teaching foreign languages and fostering language awareness in subject matter teaching requires a profound knowledge of grammar structures. Yet, while Large Language Models can act as tutors, it is unclear how effectively they can control grammar in generated text and adapt to learner needs. In this study, we investigate the ability of these models to exemplify pedagogically relevant grammar patterns, detect instances of grammar in a given text, and constrain text generation to grammar characteristic of a proficiency level. Concretely, we (1) evaluate the ability of GPT3.5 and GPT4 to generate example sentences for the standard English Grammar Profile CEFR taxonomy using few-shot in-context learning, (2) train BERT-based detectors with these generated examples of grammatical patterns, and (3) control the grammatical complexity of text generated by the open Mistral model by ranking sentence candidates with these detectors. We show that the grammar pattern instantiation quality is accurate but too homogeneous, and our classifiers successfully detect these patterns. A GPT-generated dataset of almost 1 million positive and negative examples for the English Grammar Profile is released with this work. With our method, Mistral{'}s output significantly increases the number of characteristic grammar constructions on the desired level, outperforming GPT4. This showcases how language domain knowledge can enhance Large Language Models for specific education needs, facilitating their effective use for intelligent tutor development and AI-generated materials. Code, models, and data are available at https://github.com/dominikglandorf/LLM-grammar.", }

Teaching foreign languages and fostering language awareness in subject matter teaching requires a profound knowledge of grammar structures. Yet, while Large Language Models can act as tutors, it is unclear how effectively they can control grammar in generated text and adapt to learner needs. In this study, we investigate the ability of these models to exemplify pedagogically relevant grammar patterns, detect instances of grammar in a given text, and constrain text generation to grammar characteristic of a proficiency level. Concretely, we (1) evaluate the ability of GPT3.5 and GPT4 to generate example sentences for the standard English Grammar Profile CEFR taxonomy using few-shot in-context learning, (2) train BERT-based detectors with these generated examples of grammatical patterns, and (3) control the grammatical complexity of text generated by the open Mistral model by ranking sentence candidates with these detectors. We show that the grammar pattern instantiation quality is accurate but too homogeneous, and our classifiers successfully detect these patterns. A GPT-generated dataset of almost 1 million positive and negative examples for the English Grammar Profile is released with this work. With our method, Mistral{'}s output significantly increases the number of characteristic grammar constructions on the desired level, outperforming GPT4. This showcases how language domain knowledge can enhance Large Language Models for specific education needs, facilitating their effective use for intelligent tutor development and AI-generated materials. Code, models, and data are available at https://github.com/dominikglandorf/LLM-grammar.

Towards Fine-Grained Pedagogical Control over English Grammar Complexity in Educational Text Generation

bea-1.24

Poster

https://aclanthology.org/2024.bea-1.25.bib

https://aclanthology.org/2024.bea-1.25/

@inproceedings{chamieh-etal-2024-llms, title = "{LLM}s in Short Answer Scoring: Limitations and Promise of Zero-Shot and Few-Shot Approaches", author = "Chamieh, Imran and Zesch, Torsten and Giebermann, Klaus", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.25", pages = "309--315", abstract = "In this work, we investigate the potential of Large Language Models (LLMs) for automated short answer scoring. We test zero-shot and few-shot settings, and compare with fine-tuned models and a supervised upper-bound, across three diverse datasets. Our results, in zero-shot and few-shot settings, show that LLMs perform poorly in these settings: LLMs have difficulty with tasks that require complex reasoning or domain-specific knowledge. While the models show promise on general knowledge tasks. The fine-tuned model come close to the supervised results but are still not feasible for application, highlighting potential overfitting issues. Overall, our study highlights the challenges and limitations of LLMs in short answer scoring and indicates that there currently seems to be no basis for applying LLMs for short answer scoring.", }

In this work, we investigate the potential of Large Language Models (LLMs) for automated short answer scoring. We test zero-shot and few-shot settings, and compare with fine-tuned models and a supervised upper-bound, across three diverse datasets. Our results, in zero-shot and few-shot settings, show that LLMs perform poorly in these settings: LLMs have difficulty with tasks that require complex reasoning or domain-specific knowledge. While the models show promise on general knowledge tasks. The fine-tuned model come close to the supervised results but are still not feasible for application, highlighting potential overfitting issues. Overall, our study highlights the challenges and limitations of LLMs in short answer scoring and indicates that there currently seems to be no basis for applying LLMs for short answer scoring.

LLMs in Short Answer Scoring: Limitations and Promise of Zero-Shot and Few-Shot Approaches

bea-1.25

Poster

https://aclanthology.org/2024.bea-1.26.bib

https://aclanthology.org/2024.bea-1.26/

@inproceedings{doi-etal-2024-automated, title = "Automated Essay Scoring Using Grammatical Variety and Errors with Multi-Task Learning and Item Response Theory", author = "Doi, Kosuke and Sudoh, Katsuhito and Nakamura, Satoshi", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.26", pages = "316--329", abstract = "This study examines the effect of grammatical features in automatic essay scoring (AES). We use two kinds of grammatical features as input to an AES model: (1) grammatical items that writers used correctly in essays, and (2) the number of grammatical errors. Experimental results show that grammatical features improve the performance of AES models that predict the holistic scores of essays. Multi-task learning with the holistic and grammar scores, alongside using grammatical features, resulted in a larger improvement in model performance. We also show that a model using grammar abilities estimated using Item Response Theory (IRT) as the labels for the auxiliary task achieved comparable performance to when we used grammar scores assigned by human raters. In addition, we weight the grammatical features using IRT to consider the difficulty of grammatical items and writers{'} grammar abilities. We found that weighting grammatical features with the difficulty led to further improvement in performance.", }

This study examines the effect of grammatical features in automatic essay scoring (AES). We use two kinds of grammatical features as input to an AES model: (1) grammatical items that writers used correctly in essays, and (2) the number of grammatical errors. Experimental results show that grammatical features improve the performance of AES models that predict the holistic scores of essays. Multi-task learning with the holistic and grammar scores, alongside using grammatical features, resulted in a larger improvement in model performance. We also show that a model using grammar abilities estimated using Item Response Theory (IRT) as the labels for the auxiliary task achieved comparable performance to when we used grammar scores assigned by human raters. In addition, we weight the grammatical features using IRT to consider the difficulty of grammatical items and writers{'} grammar abilities. We found that weighting grammatical features with the difficulty led to further improvement in performance.

Automated Essay Scoring Using Grammatical Variety and Errors with Multi-Task Learning and Item Response Theory

bea-1.26

Poster

2406.08817

https://aclanthology.org/2024.bea-1.27.bib

https://aclanthology.org/2024.bea-1.27/

@inproceedings{shaka-etal-2024-error, title = "Error Tracing in Programming: A Path to Personalised Feedback", author = "Shaka, Martha and Carraro, Diego and Brown, Kenneth", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.27", pages = "330--342", abstract = "Knowledge tracing, the process of estimating students{'} mastery over concepts from their past performance and predicting future outcomes, often relies on binary pass/fail predictions. This hinders the provision of specific feedback by failing to diagnose precise errors. We present an error-tracing model for learning programming that advances traditional knowledge tracing by employing multi-label classification to forecast exact errors students may generate. Through experiments on a real student dataset, we validate our approach and compare it to two baseline knowledge-tracing methods. We demonstrate an improved ability to predict specific errors, for first attempts and for subsequent attempts at individual problems.", }

Knowledge tracing, the process of estimating students{'} mastery over concepts from their past performance and predicting future outcomes, often relies on binary pass/fail predictions. This hinders the provision of specific feedback by failing to diagnose precise errors. We present an error-tracing model for learning programming that advances traditional knowledge tracing by employing multi-label classification to forecast exact errors students may generate. Through experiments on a real student dataset, we validate our approach and compare it to two baseline knowledge-tracing methods. We demonstrate an improved ability to predict specific errors, for first attempts and for subsequent attempts at individual problems.

Error Tracing in Programming: A Path to Personalised Feedback

bea-1.27

Poster

https://aclanthology.org/2024.bea-1.28.bib

https://aclanthology.org/2024.bea-1.28/

@inproceedings{lim-lee-2024-improving, title = "Improving Readability Assessment with Ordinal Log-Loss", author = "Lim, Ho Hung and Lee, John", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.28", pages = "343--350", abstract = "Automatic Readability Assessment (ARA) predicts the level of difficulty of a text, e.g. at Grade 1 to Grade 12. ARA is an ordinal classification task since the predicted levels follow an underlying order, from easy to difficult. However, most neural ARA models ignore the distance between the gold level and predicted level, treating all levels as independent labels. This paper investigates whether distance-sensitive loss functions can improve ARA performance. We evaluate a variety of loss functions on neural ARA models, and show that ordinal log-loss can produce statistically significant improvement over the standard cross-entropy loss in terms of adjacent accuracy in a majority of our datasets.", }

Automatic Readability Assessment (ARA) predicts the level of difficulty of a text, e.g. at Grade 1 to Grade 12. ARA is an ordinal classification task since the predicted levels follow an underlying order, from easy to difficult. However, most neural ARA models ignore the distance between the gold level and predicted level, treating all levels as independent labels. This paper investigates whether distance-sensitive loss functions can improve ARA performance. We evaluate a variety of loss functions on neural ARA models, and show that ordinal log-loss can produce statistically significant improvement over the standard cross-entropy loss in terms of adjacent accuracy in a majority of our datasets.

Improving Readability Assessment with Ordinal Log-Loss

bea-1.28

Poster

https://aclanthology.org/2024.bea-1.29.bib

https://aclanthology.org/2024.bea-1.29/

@inproceedings{paddags-etal-2024-automated, title = "Automated Sentence Generation for a Spaced Repetition Software", author = "Paddags, Benjamin and Hershcovich, Daniel and Savage, Valkyrie", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.29", pages = "351--364", abstract = "This paper presents and tests AllAI, an app that utilizes state-of-the-art NLP technology to assist second language acquisition through a novel method of sentence-based spaced repetition. Diverging from current single word or fixed sentence repetition, AllAI dynamically combines words due for repetition into sentences, enabling learning words in context while scheduling them independently. This research explores various suitable NLP paradigms and finds a few-shot prompting approach and retrieval of existing sentences from a corpus to yield the best correctness and scheduling accuracy. Subsequently, it evaluates these methods on 26 learners of Danish, finding a four-fold increase in the speed at which new words are learned, compared to conventional spaced repetition. Users of the retrieval method also reported significantly higher enjoyment, hinting at a higher user engagement.", }

This paper presents and tests AllAI, an app that utilizes state-of-the-art NLP technology to assist second language acquisition through a novel method of sentence-based spaced repetition. Diverging from current single word or fixed sentence repetition, AllAI dynamically combines words due for repetition into sentences, enabling learning words in context while scheduling them independently. This research explores various suitable NLP paradigms and finds a few-shot prompting approach and retrieval of existing sentences from a corpus to yield the best correctness and scheduling accuracy. Subsequently, it evaluates these methods on 26 learners of Danish, finding a four-fold increase in the speed at which new words are learned, compared to conventional spaced repetition. Users of the retrieval method also reported significantly higher enjoyment, hinting at a higher user engagement.

Automated Sentence Generation for a Spaced Repetition Software

bea-1.29

Poster

https://aclanthology.org/2024.bea-1.30.bib

https://aclanthology.org/2024.bea-1.30/

@inproceedings{li-etal-2024-using, title = "Using Large Language Models to Assess Young Students{'} Writing Revisions", author = "Li, Tianwen and Liu, Zhexiong and Matsumura, Lindsay and Wang, Elaine and Litman, Diane and Correnti, Richard", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.30", pages = "365--380", abstract = "Although effective revision is the crucial component of writing instruction, few automated writing evaluation (AWE) systems specifically focus on the quality of the revisions students undertake. In this study we investigate the use of a large language model (GPT-4) with Chain-of-Thought (CoT) prompting for assessing the quality of young students{'} essay revisions aligned with the automated feedback messages they received. Results indicate that GPT-4 has significant potential for evaluating revision quality, particularly when detailed rubrics are included that describe common revision patterns shown by young writers. However, the addition of CoT prompting did not significantly improve performance. Further examination of GPT-4{'}s scoring performance across various levels of student writing proficiency revealed variable agreement with human ratings. The implications for improving AWE systems focusing on young students are discussed.", }

Although effective revision is the crucial component of writing instruction, few automated writing evaluation (AWE) systems specifically focus on the quality of the revisions students undertake. In this study we investigate the use of a large language model (GPT-4) with Chain-of-Thought (CoT) prompting for assessing the quality of young students{'} essay revisions aligned with the automated feedback messages they received. Results indicate that GPT-4 has significant potential for evaluating revision quality, particularly when detailed rubrics are included that describe common revision patterns shown by young writers. However, the addition of CoT prompting did not significantly improve performance. Further examination of GPT-4{'}s scoring performance across various levels of student writing proficiency revealed variable agreement with human ratings. The implications for improving AWE systems focusing on young students are discussed.

Using Large Language Models to Assess Young Students' Writing Revisions

bea-1.30

Poster

https://aclanthology.org/2024.bea-1.31.bib

https://aclanthology.org/2024.bea-1.31/

@inproceedings{berruti-etal-2024-automatic, title = "Automatic Crossword Clues Extraction for Language Learning", author = "Berruti, Santiago and Collazo, Arturo and Sellanes, Diego and Ros{\'a}, Aiala and Chiruzzo, Luis", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.31", pages = "381--390", abstract = "Crosswords are a powerful tool that could be used in educational contexts, but they are not that easy to build. In this work, we present experiments on automatically extracting clues from simple texts that could be used to create crosswords, with the aim of using them in the context of teaching English at the beginner level. We present a series of heuristic patterns based on NLP tools for extracting clues, and use them to create a set of 2209 clues from a collection of 400 simple texts. Human annotators labeled the clues, and this dataset is used to evaluate the performance of our heuristics, and also to create a classifier that predicts if an extracted clue is correct. Our best classifier achieves an accuracy of 84{\%}.", }

Crosswords are a powerful tool that could be used in educational contexts, but they are not that easy to build. In this work, we present experiments on automatically extracting clues from simple texts that could be used to create crosswords, with the aim of using them in the context of teaching English at the beginner level. We present a series of heuristic patterns based on NLP tools for extracting clues, and use them to create a set of 2209 clues from a collection of 400 simple texts. Human annotators labeled the clues, and this dataset is used to evaluate the performance of our heuristics, and also to create a classifier that predicts if an extracted clue is correct. Our best classifier achieves an accuracy of 84{\%}.

Automatic Crossword Clues Extraction for Language Learning

bea-1.31

Poster

https://aclanthology.org/2024.bea-1.32.bib

https://aclanthology.org/2024.bea-1.32/

@inproceedings{gurin-schleifer-etal-2024-anna, title = "Anna Karenina Strikes Again: Pre-Trained {LLM} Embeddings May Favor High-Performing Learners", author = "Gurin Schleifer, Abigail and Beigman Klebanov, Beata and Ariely, Moriah and Alexandron, Giora", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.32", pages = "391--402", abstract = "Unsupervised clustering of student responses to open-ended questions into behavioral and cognitive profiles using pre-trained LLM embeddings is an emerging technique, but little is known about how well this captures pedagogically meaningful information. We investigate this in the context of student responses to open-ended questions in biology, which were previously analyzed and clustered by experts into theory-driven Knowledge Profiles (KPs).Comparing these KPs to ones discovered by purely data-driven clustering techniques, we report poor discoverability of most KPs, except for the ones including the correct answers. We trace this {`}discoverability bias{'} to the representations of KPs in the pre-trained LLM embeddings space.", }

Unsupervised clustering of student responses to open-ended questions into behavioral and cognitive profiles using pre-trained LLM embeddings is an emerging technique, but little is known about how well this captures pedagogically meaningful information. We investigate this in the context of student responses to open-ended questions in biology, which were previously analyzed and clustered by experts into theory-driven Knowledge Profiles (KPs).Comparing these KPs to ones discovered by purely data-driven clustering techniques, we report poor discoverability of most KPs, except for the ones including the correct answers. We trace this {`}discoverability bias{'} to the representations of KPs in the pre-trained LLM embeddings space.

Anna Karenina Strikes Again: Pre-Trained LLM Embeddings May Favor High-Performing Learners

bea-1.32

Poster

2406.06599

https://aclanthology.org/2024.bea-1.33.bib

https://aclanthology.org/2024.bea-1.33/

@inproceedings{carpenter-etal-2024-assessing, title = "Assessing Student Explanations with Large Language Models Using Fine-Tuning and Few-Shot Learning", author = "Carpenter, Dan and Min, Wookhee and Lee, Seung and Ozogul, Gamze and Zheng, Xiaoying and Lester, James", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.33", pages = "403--413", abstract = "The practice of soliciting self-explanations from students is widely recognized for its pedagogical benefits. However, the labor-intensive effort required to manually assess students{'} explanations makes it impractical for classroom settings. As a result, many current solutions to gauge students{'} understanding during class are often limited to multiple choice or fill-in-the-blank questions, which are less effective at exposing misconceptions or helping students to understand and integrate new concepts. Recent advances in large language models (LLMs) present an opportunity to assess student explanations in real-time, making explanation-based classroom response systems feasible for implementation. In this work, we investigate LLM-based approaches for assessing the correctness of students{'} explanations in response to undergraduate computer science questions. We investigate alternative prompting approaches for multiple LLMs (i.e., Llama 2, GPT-3.5, and GPT-4) and compare their performance to FLAN-T5 models trained in a fine-tuning manner. The results suggest that the highest accuracy and weighted F1 score were achieved by fine-tuning FLAN-T5, while an in-context learning approach with GPT-4 attains the highest macro F1 score.", }

The practice of soliciting self-explanations from students is widely recognized for its pedagogical benefits. However, the labor-intensive effort required to manually assess students{'} explanations makes it impractical for classroom settings. As a result, many current solutions to gauge students{'} understanding during class are often limited to multiple choice or fill-in-the-blank questions, which are less effective at exposing misconceptions or helping students to understand and integrate new concepts. Recent advances in large language models (LLMs) present an opportunity to assess student explanations in real-time, making explanation-based classroom response systems feasible for implementation. In this work, we investigate LLM-based approaches for assessing the correctness of students{'} explanations in response to undergraduate computer science questions. We investigate alternative prompting approaches for multiple LLMs (i.e., Llama 2, GPT-3.5, and GPT-4) and compare their performance to FLAN-T5 models trained in a fine-tuning manner. The results suggest that the highest accuracy and weighted F1 score were achieved by fine-tuning FLAN-T5, while an in-context learning approach with GPT-4 attains the highest macro F1 score.

Assessing Student Explanations with Large Language Models Using Fine-Tuning and Few-Shot Learning

bea-1.33

Poster

https://aclanthology.org/2024.bea-1.34.bib

https://aclanthology.org/2024.bea-1.34/

@inproceedings{munoz-sanchez-etal-2024-harnessing, title = "Harnessing {GPT} to Study Second Language Learner Essays: Can We Use Perplexity to Determine Linguistic Competence?", author = "Mu{\~n}oz S{\'a}nchez, Ricardo and Dobnik, Simon and Volodina, Elena", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.34", pages = "414--427", abstract = "Generative language models have been used to study a wide variety of phenomena in NLP. This allows us to better understand the linguistic capabilities of those models and to better analyse the texts that we are working with. However, these studies have mainly focused on text generated by L1 speakers of English. In this paper we study whether linguistic competence of L2 learners of Swedish (through their performance on essay tasks) correlates with the perplexity of a decoder-only model (GPT-SW3). We run two sets of experiments, doing both quantitative and qualitative analyses for each of them. In the first one, we analyse the perplexities of the essays and compare them with the CEFR level of the essays, both from an essay-wide level and from a token level. In our second experiment, we compare the perplexity of an L2 learner essay with a normalised version of it. We find that the perplexity of essays tends to be lower for higher CEFR levels and that normalised essays have a lower perplexity than the original versions. Moreover, we find that different factors can lead to spikes in perplexity, not all of them being related to L2 learner language.", }

Generative language models have been used to study a wide variety of phenomena in NLP. This allows us to better understand the linguistic capabilities of those models and to better analyse the texts that we are working with. However, these studies have mainly focused on text generated by L1 speakers of English. In this paper we study whether linguistic competence of L2 learners of Swedish (through their performance on essay tasks) correlates with the perplexity of a decoder-only model (GPT-SW3). We run two sets of experiments, doing both quantitative and qualitative analyses for each of them. In the first one, we analyse the perplexities of the essays and compare them with the CEFR level of the essays, both from an essay-wide level and from a token level. In our second experiment, we compare the perplexity of an L2 learner essay with a normalised version of it. We find that the perplexity of essays tends to be lower for higher CEFR levels and that normalised essays have a lower perplexity than the original versions. Moreover, we find that different factors can lead to spikes in perplexity, not all of them being related to L2 learner language.

Harnessing GPT to Study Second Language Learner Essays: Can We Use Perplexity to Determine Linguistic Competence?

bea-1.34

Poster

https://aclanthology.org/2024.bea-1.35.bib

https://aclanthology.org/2024.bea-1.35/

@inproceedings{yancey-etal-2024-bert, title = "{BERT}-{IRT}: Accelerating Item Piloting with {BERT} Embeddings and Explainable {IRT} Models", author = "Yancey, Kevin P. and Runge, Andrew and LaFlair, Geoffrey and Mulcaire, Phoebe", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.35", pages = "428--438", abstract = "Estimating item parameters (e.g., the difficulty of a question) is an important part of modern high-stakes tests. Conventional methods require lengthy pilots to collect response data from a representative population of test-takers. The need for these pilots limit item bank size and how often those item banks can be refreshed, impacting test security, while increasing costs needed to support the test and taking up the test-taker{'}s valuable time. Our paper presents a novel explanatory item response theory (IRT) model, BERT-IRT, that has been used on the Duolingo English Test (DET), a high-stakes test of English, to reduce the length of pilots by a factor of 10. Our evaluation shows how the model uses BERT embeddings and engineered NLP features to accelerate item piloting without sacrificing criterion validity or reliability.", }

Estimating item parameters (e.g., the difficulty of a question) is an important part of modern high-stakes tests. Conventional methods require lengthy pilots to collect response data from a representative population of test-takers. The need for these pilots limit item bank size and how often those item banks can be refreshed, impacting test security, while increasing costs needed to support the test and taking up the test-taker{'}s valuable time. Our paper presents a novel explanatory item response theory (IRT) model, BERT-IRT, that has been used on the Duolingo English Test (DET), a high-stakes test of English, to reduce the length of pilots by a factor of 10. Our evaluation shows how the model uses BERT embeddings and engineered NLP features to accelerate item piloting without sacrificing criterion validity or reliability.

BERT-IRT: Accelerating Item Piloting with BERT Embeddings and Explainable IRT Models

bea-1.35

Poster

https://aclanthology.org/2024.bea-1.36.bib

https://aclanthology.org/2024.bea-1.36/

@inproceedings{ding-etal-2024-transfer, title = "Transfer Learning of Argument Mining in Student Essays", author = "Ding, Yuning and Lohmann, Julian and Schaller, Nils-Jonathan and Jansen, Thorben and Horbach, Andrea", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.36", pages = "439--449", abstract = "This paper explores the transferability of a cross-prompt argument mining model trained on argumentative essays authored by native English-speaking learners (EN-L1) across educational contexts and languages. Specifically, the adaptability of a multilingual transformer model is assessed through its application to comparable argumentative essays authored by English-as-a-foreign-language learners (EN-L2) for context transfer, and a dataset composed of essays written by native German learners (DE) for both language and task transfer. To separate language effects from educational context effects, we also perform experiments on a machine-translated version of the German dataset (DE-MT). Our findings demonstrate that, even under zero-shot conditions, a model trained on native English speakers exhibits satisfactory performance on the EN-L2/DE datasets. Machine translation does not substantially enhance this performance, suggesting that distinct writing styles across educational contexts impact performance more than language differences.", }

This paper explores the transferability of a cross-prompt argument mining model trained on argumentative essays authored by native English-speaking learners (EN-L1) across educational contexts and languages. Specifically, the adaptability of a multilingual transformer model is assessed through its application to comparable argumentative essays authored by English-as-a-foreign-language learners (EN-L2) for context transfer, and a dataset composed of essays written by native German learners (DE) for both language and task transfer. To separate language effects from educational context effects, we also perform experiments on a machine-translated version of the German dataset (DE-MT). Our findings demonstrate that, even under zero-shot conditions, a model trained on native English speakers exhibits satisfactory performance on the EN-L2/DE datasets. Machine translation does not substantially enhance this performance, suggesting that distinct writing styles across educational contexts impact performance more than language differences.

Transfer Learning of Argument Mining in Student Essays

bea-1.36

Poster

https://aclanthology.org/2024.bea-1.37.bib

https://aclanthology.org/2024.bea-1.37/

@inproceedings{bradford-etal-2024-building, title = "Building Robust Content Scoring Models for Student Explanations of Social Justice Science Issues", author = "Bradford, Allison and Steimel, Kenneth and Riordan, Brian and Linn, Marcia", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.37", pages = "450--458", abstract = "With increased attention to connecting science topics to real-world contexts, like issues of social justice, teachers need support to assess student progress in explaining such issues. In this work, we explore the robustness of NLP-based automatic content scoring models that provide insight into student ability to integrate their science and social justice ideas in two different environmental science contexts. We leverage encoder-only transformer models to capture the degree to which students explain a science phenomenon, understand the intersecting justice issues, and integrate their understanding of science and social justice. We developed models training on data from each of the contexts as well as from a combined dataset. We found that the models developed in one context generate educationally useful scores in the other context. The model trained on the combined dataset performed as well as or better than the models trained on separate datasets in most cases. Quadratic weighted kappas demonstrate that these models are above threshold for use in classrooms.", }

With increased attention to connecting science topics to real-world contexts, like issues of social justice, teachers need support to assess student progress in explaining such issues. In this work, we explore the robustness of NLP-based automatic content scoring models that provide insight into student ability to integrate their science and social justice ideas in two different environmental science contexts. We leverage encoder-only transformer models to capture the degree to which students explain a science phenomenon, understand the intersecting justice issues, and integrate their understanding of science and social justice. We developed models training on data from each of the contexts as well as from a combined dataset. We found that the models developed in one context generate educationally useful scores in the other context. The model trained on the combined dataset performed as well as or better than the models trained on separate datasets in most cases. Quadratic weighted kappas demonstrate that these models are above threshold for use in classrooms.

Building Robust Content Scoring Models for Student Explanations of Social Justice Science Issues

bea-1.37

Poster

https://aclanthology.org/2024.bea-1.38.bib

https://aclanthology.org/2024.bea-1.38/

@inproceedings{beigman-klebanov-etal-2024-miscue, title = "From Miscue to Evidence of Difficulty: Analysis of Automatically Detected Miscues in Oral Reading for Feedback Potential", author = "Beigman Klebanov, Beata and Suhan, Michael and O{'}Reilly, Tenaha and Wang, Zuowei", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.38", pages = "459--469", abstract = "This research is situated in the space between an existing NLP capability and its use(s) in an educational context. We analyze oral reading data collected with a deployed automated speech analysis software and consider how the results of automated speech analysis can be interpreted and used to inform the ideation and design of a new feature {--} feedback to learners and teachers. Our analysis shows how the details of the system{'}s performance and the details of the context of use both significantly impact the ideation process.", }

This research is situated in the space between an existing NLP capability and its use(s) in an educational context. We analyze oral reading data collected with a deployed automated speech analysis software and consider how the results of automated speech analysis can be interpreted and used to inform the ideation and design of a new feature {--} feedback to learners and teachers. Our analysis shows how the details of the system{'}s performance and the details of the context of use both significantly impact the ideation process.

From Miscue to Evidence of Difficulty: Analysis of Automatically Detected Miscues in Oral Reading for Feedback Potential

bea-1.38

Poster

https://aclanthology.org/2024.bea-1.39.bib

https://aclanthology.org/2024.bea-1.39/

@inproceedings{yaneva-etal-2024-findings, title = "Findings from the First Shared Task on Automated Prediction of Difficulty and Response Time for Multiple-Choice Questions", author = "Yaneva, Victoria and North, Kai and Baldwin, Peter and Ha, Le An and Rezayi, Saed and Zhou, Yiyun and Ray Choudhury, Sagnik and Harik, Polina and Clauser, Brian", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.39", pages = "470--482", abstract = "This paper reports findings from the First Shared Task on Automated Prediction of Difficulty and Response Time for Multiple-Choice Questions. The task was organized as part of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA{'}24), held in conjunction with NAACL 2024, and called upon the research community to contribute solutions to the problem of modeling difficulty and response time for clinical multiple-choice questions (MCQs). A set of 667 previously used and now retired MCQs from the United States Medical Licensing Examination (USMLE®) and their corresponding difficulties and mean response times were made available for experimentation. A total of 17 teams submitted solutions and 12 teams submitted system report papers describing their approaches. This paper summarizes the findings from the shared task and analyzes the main approaches proposed by the participants.", }

This paper reports findings from the First Shared Task on Automated Prediction of Difficulty and Response Time for Multiple-Choice Questions. The task was organized as part of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA{'}24), held in conjunction with NAACL 2024, and called upon the research community to contribute solutions to the problem of modeling difficulty and response time for clinical multiple-choice questions (MCQs). A set of 667 previously used and now retired MCQs from the United States Medical Licensing Examination (USMLE®) and their corresponding difficulties and mean response times were made available for experimentation. A total of 17 teams submitted solutions and 12 teams submitted system report papers describing their approaches. This paper summarizes the findings from the shared task and analyzes the main approaches proposed by the participants.

Findings from the First Shared Task on Automated Prediction of Difficulty and Response Time for Multiple-Choice Questions

bea-1.39

Poster

https://aclanthology.org/2024.bea-1.40.bib

https://aclanthology.org/2024.bea-1.40/

@inproceedings{gombert-etal-2024-predicting, title = "Predicting Item Difficulty and Item Response Time with Scalar-mixed Transformer Encoder Models and Rational Network Regression Heads", author = "Gombert, Sebastian and Menzel, Lukas and Di Mitri, Daniele and Drachsler, Hendrik", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.40", pages = "483--492", abstract = "This paper describes a contribution to the BEA 2024 Shared Task on Automated Prediction of Item Difficulty and Response Time. The participants in this shared task are to develop models for predicting the difficulty and response time of multiple-choice items in the medical field. These items were taken from the United States Medical Licensing Examination® (USMLE®), a high-stakes medical exam. For this purpose, we evaluated multiple BERT-like pre-trained transformer encoder models, which we combined with Scalar Mixing and two custom 2-layer classification heads using learnable Rational Activations as an activation function, each for predicting one of the two variables of interest in a multi-task setup. Our best models placed first out of 43 for predicting item difficulty and fifth out of 34 for predicting Item Response Time.", }

This paper describes a contribution to the BEA 2024 Shared Task on Automated Prediction of Item Difficulty and Response Time. The participants in this shared task are to develop models for predicting the difficulty and response time of multiple-choice items in the medical field. These items were taken from the United States Medical Licensing Examination® (USMLE®), a high-stakes medical exam. For this purpose, we evaluated multiple BERT-like pre-trained transformer encoder models, which we combined with Scalar Mixing and two custom 2-layer classification heads using learnable Rational Activations as an activation function, each for predicting one of the two variables of interest in a multi-task setup. Our best models placed first out of 43 for predicting item difficulty and fifth out of 34 for predicting Item Response Time.

Predicting Item Difficulty and Item Response Time with Scalar-mixed Transformer Encoder Models and Rational Network Regression Heads

bea-1.40

Poster

https://aclanthology.org/2024.bea-1.41.bib

https://aclanthology.org/2024.bea-1.41/

@inproceedings{rogoz-ionescu-2024-unibucllm, title = "{U}nibuc{LLM}: Harnessing {LLM}s for Automated Prediction of Item Difficulty and Response Time for Multiple-Choice Questions", author = "Rogoz, Ana-Cristina and Ionescu, Radu Tudor", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.41", pages = "493--502", abstract = "This work explores a novel data augmentation method based on Large Language Models (LLMs) for predicting item difficulty and response time of retired USMLE Multiple-Choice Questions (MCQs) in the BEA 2024 Shared Task. Our approach is based on augmenting the dataset with answers from zero-shot LLMs (Falcon, Meditron, Mistral) and employing transformer-based models based on six alternative feature combinations. The results suggest that predicting the difficulty of questions is more challenging. Notably, our top performing methods consistently include the question text, and benefit from the variability of LLM answers, highlighting the potential of LLMs for improving automated assessment in medical licensing exams. We make our code available at: https://github.com/ana-rogoz/BEA-2024.", }

This work explores a novel data augmentation method based on Large Language Models (LLMs) for predicting item difficulty and response time of retired USMLE Multiple-Choice Questions (MCQs) in the BEA 2024 Shared Task. Our approach is based on augmenting the dataset with answers from zero-shot LLMs (Falcon, Meditron, Mistral) and employing transformer-based models based on six alternative feature combinations. The results suggest that predicting the difficulty of questions is more challenging. Notably, our top performing methods consistently include the question text, and benefit from the variability of LLM answers, highlighting the potential of LLMs for improving automated assessment in medical licensing exams. We make our code available at: https://github.com/ana-rogoz/BEA-2024.

UnibucLLM: Harnessing LLMs for Automated Prediction of Item Difficulty and Response Time for Multiple-Choice Questions

bea-1.41

Poster

2404.13343

https://aclanthology.org/2024.bea-1.42.bib

https://aclanthology.org/2024.bea-1.42/

@inproceedings{felice-duran-karaoz-2024-british, title = "The {B}ritish Council submission to the {BEA} 2024 shared task", author = "Felice, Mariano and Duran Karaoz, Zeynep", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.42", pages = "503--511", abstract = "This paper describes our submission to the item difficulty prediction track of the BEA 2024 shared task. Our submission included the output of three systems: 1) a feature-based linear regression model, 2) a RoBERTa-based model and 3) a linear regression ensemble built on the predictions of the two previous models. Our systems ranked 7th, 8th and 5th respectively, demonstrating that simple models can achieve optimal results. A closer look at the results shows that predictions are more accurate for items in the middle of the difficulty range, with no other obvious relationships between difficulty and the accuracy of predictions.", }

This paper describes our submission to the item difficulty prediction track of the BEA 2024 shared task. Our submission included the output of three systems: 1) a feature-based linear regression model, 2) a RoBERTa-based model and 3) a linear regression ensemble built on the predictions of the two previous models. Our systems ranked 7th, 8th and 5th respectively, demonstrating that simple models can achieve optimal results. A closer look at the results shows that predictions are more accurate for items in the middle of the difficulty range, with no other obvious relationships between difficulty and the accuracy of predictions.

The British Council submission to the BEA 2024 shared task

bea-1.42

Poster

https://aclanthology.org/2024.bea-1.43.bib

https://aclanthology.org/2024.bea-1.43/

@inproceedings{tack-etal-2024-itec, title = "{ITEC} at {BEA} 2024 Shared Task: Predicting Difficulty and Response Time of Medical Exam Questions with Statistical, Machine Learning, and Language Models", author = {Tack, Ana{\"\i}s and Buseyne, Siem and Chen, Changsheng and D{'}hondt, Robbe and De Vrindt, Michiel and Gharahighehi, Alireza and Metwaly, Sameh and Nakano, Felipe Kenji and Noreillie, Ann-Sophie}, editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.43", pages = "512--521", abstract = "This paper presents the results of our participation in the BEA 2024 shared task on the automated prediction of item difficulty and item response time (APIDIRT), hosted by the NBME (National Board of Medical Examiners). During this task, practice multiple-choice questions from the United States Medical Licensing Examination® (USMLE®) were shared, and research teams were tasked with devising systems capable of predicting the difficulty and average response time for new exam questions.Our team, part of the interdisciplinary itec research group, participated in the task. We extracted linguistic features and clinical embeddings from question items and tested various modeling techniques, including statistical regression, machine learning, language models, and ensemble methods. Surprisingly, simplermodels such as Lasso and random forest regression, utilizing principal component features from linguistic and clinical embeddings, outperformed more complex models. In the competition, our random forest model ranked 4th out of 43 submissions for difficulty prediction, while the Lasso model secured the 2nd position out of 34 submissions for response time prediction. Further analysis suggests that had we submitted the Lasso model for difficulty prediction, we would have achieved an even higher ranking. We also observed that predicting response time is easier than predicting difficulty, with features such as item length, type, exam step, and analytical thinking influencing response time prediction more significantly.", }

This paper presents the results of our participation in the BEA 2024 shared task on the automated prediction of item difficulty and item response time (APIDIRT), hosted by the NBME (National Board of Medical Examiners). During this task, practice multiple-choice questions from the United States Medical Licensing Examination® (USMLE®) were shared, and research teams were tasked with devising systems capable of predicting the difficulty and average response time for new exam questions.Our team, part of the interdisciplinary itec research group, participated in the task. We extracted linguistic features and clinical embeddings from question items and tested various modeling techniques, including statistical regression, machine learning, language models, and ensemble methods. Surprisingly, simplermodels such as Lasso and random forest regression, utilizing principal component features from linguistic and clinical embeddings, outperformed more complex models. In the competition, our random forest model ranked 4th out of 43 submissions for difficulty prediction, while the Lasso model secured the 2nd position out of 34 submissions for response time prediction. Further analysis suggests that had we submitted the Lasso model for difficulty prediction, we would have achieved an even higher ranking. We also observed that predicting response time is easier than predicting difficulty, with features such as item length, type, exam step, and analytical thinking influencing response time prediction more significantly.

ITEC at BEA 2024 Shared Task: Predicting Difficulty and Response Time of Medical Exam Questions with Statistical, Machine Learning, and Language Models

bea-1.43

Poster

https://aclanthology.org/2024.bea-1.44.bib

https://aclanthology.org/2024.bea-1.44/

@inproceedings{bulut-etal-2024-item, title = "Item Difficulty and Response Time Prediction with Large Language Models: An Empirical Analysis of {USMLE} Items", author = "Bulut, Okan and Gorgun, Guher and Tan, Bin", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.44", pages = "522--527", abstract = "This paper summarizes our methodology and results for the BEA 2024 Shared Task. This competition focused on predicting item difficulty and response time for retired multiple-choice items from the United States Medical Licensing Examination® (USMLE®). We extracted linguistic features from the item stem and response options using multiple methods, including the BiomedBERT model, FastText embeddings, and Coh-Metrix. The extracted features were combined with additional features available in item metadata (e.g., item type) to predict item difficulty and average response time. The results showed that the BiomedBERT model was the most effective in predicting item difficulty, while the fine-tuned model based on FastText word embeddings was the best model for predicting response time.", }

This paper summarizes our methodology and results for the BEA 2024 Shared Task. This competition focused on predicting item difficulty and response time for retired multiple-choice items from the United States Medical Licensing Examination® (USMLE®). We extracted linguistic features from the item stem and response options using multiple methods, including the BiomedBERT model, FastText embeddings, and Coh-Metrix. The extracted features were combined with additional features available in item metadata (e.g., item type) to predict item difficulty and average response time. The results showed that the BiomedBERT model was the most effective in predicting item difficulty, while the fine-tuned model based on FastText word embeddings was the best model for predicting response time.

Item Difficulty and Response Time Prediction with Large Language Models: An Empirical Analysis of USMLE Items

bea-1.44

Poster

https://aclanthology.org/2024.bea-1.45.bib

https://aclanthology.org/2024.bea-1.45/

@inproceedings{fulari-rusert-2024-utilizing, title = "Utilizing Machine Learning to Predict Question Difficulty and Response Time for Enhanced Test Construction", author = "Fulari, Rishikesh and Rusert, Jonathan", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.45", pages = "528--533", abstract = "In this paper, we present the details of ourcontribution to the BEA Shared Task on Automated Prediction of Item Difficulty and Response Time. Participants in this collaborativeeffort are tasked with developing models to predict the difficulty and response time of multiplechoice items within the medical domain. Theseitems are sourced from the United States Medical Licensing Examination® (USMLE®), asignificant medical assessment. In order toachieve this, we experimented with two featurization techniques, one using lingusitic features and the other using embeddings generated by BERT fine-tuned over MS-MARCOdataset. Further, we tried several different machine learning models such as Linear Regression, Decision Trees, KNN and Boosting models such as XGBoost and GBDT. We found thatout of all the models we experimented withRandom Forest Regressor trained on Linguisticfeatures gave the least root mean squared error.", }

In this paper, we present the details of ourcontribution to the BEA Shared Task on Automated Prediction of Item Difficulty and Response Time. Participants in this collaborativeeffort are tasked with developing models to predict the difficulty and response time of multiplechoice items within the medical domain. Theseitems are sourced from the United States Medical Licensing Examination® (USMLE®), asignificant medical assessment. In order toachieve this, we experimented with two featurization techniques, one using lingusitic features and the other using embeddings generated by BERT fine-tuned over MS-MARCOdataset. Further, we tried several different machine learning models such as Linear Regression, Decision Trees, KNN and Boosting models such as XGBoost and GBDT. We found thatout of all the models we experimented withRandom Forest Regressor trained on Linguisticfeatures gave the least root mean squared error.

Utilizing Machine Learning to Predict Question Difficulty and Response Time for Enhanced Test Construction

bea-1.45

Poster

https://aclanthology.org/2024.bea-1.46.bib

https://aclanthology.org/2024.bea-1.46/

@inproceedings{venkata-ravi-ram-etal-2024-leveraging, title = "Leveraging Physical and Semantic Features of text item for Difficulty and Response Time Prediction of {USMLE} Questions", author = "Venkata Ravi Ram, Gummuluri and Kesanam, Ashinee and M, Anand Kumar", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.46", pages = "534--541", abstract = "This paper presents our system developed for the Shared Task on Automated Prediction of Item Difficulty and Item Response Time for USMLE questions, organized by the Association for Computational Linguistics (ACL) Special Interest Group for building Educational Applications (BEA SIGEDU). The Shared Task, held as a workshop at the North American Chapter of the Association for Computational Linguistics (NAACL) 2024 conference, aimed to advance the state-of-the-art in predicting item characteristics directly from item text, with implications for the fairness and validity of standardized exams. We compared various methods ranging from BERT for regression to Random forest, Gradient Boosting(GB), Linear Regression, Support Vector Regressor (SVR), k-nearest neighbours (KNN) Regressor, MultiLayer Perceptron(MLP) to custom-ANN using BioBERT and Word2Vec embeddings and provided inferences on which performed better. This paper also explains the importance of data augmentation to balance the data in order to get better results. We also proposed five hypotheses regarding factors impacting difficulty and response time for a question and also verified it thereby helping researchers to derive meaningful numerical attributes for accurate prediction. We achieved a RSME score of 0.315 for Difficulty prediction and 26.945 for Response Time.", }

This paper presents our system developed for the Shared Task on Automated Prediction of Item Difficulty and Item Response Time for USMLE questions, organized by the Association for Computational Linguistics (ACL) Special Interest Group for building Educational Applications (BEA SIGEDU). The Shared Task, held as a workshop at the North American Chapter of the Association for Computational Linguistics (NAACL) 2024 conference, aimed to advance the state-of-the-art in predicting item characteristics directly from item text, with implications for the fairness and validity of standardized exams. We compared various methods ranging from BERT for regression to Random forest, Gradient Boosting(GB), Linear Regression, Support Vector Regressor (SVR), k-nearest neighbours (KNN) Regressor, MultiLayer Perceptron(MLP) to custom-ANN using BioBERT and Word2Vec embeddings and provided inferences on which performed better. This paper also explains the importance of data augmentation to balance the data in order to get better results. We also proposed five hypotheses regarding factors impacting difficulty and response time for a question and also verified it thereby helping researchers to derive meaningful numerical attributes for accurate prediction. We achieved a RSME score of 0.315 for Difficulty prediction and 26.945 for Response Time.

Leveraging Physical and Semantic Features of text item for Difficulty and Response Time Prediction of USMLE Questions

bea-1.46

Poster

https://aclanthology.org/2024.bea-1.47.bib

https://aclanthology.org/2024.bea-1.47/

@inproceedings{duenas-etal-2024-upn, title = "{UPN}-{ICC} at {BEA} 2024 Shared Task: Leveraging {LLM}s for Multiple-Choice Questions Difficulty Prediction", author = "Duenas, George and Jimenez, Sergio and Mateus Ferro, Geral", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.47", pages = "542--550", abstract = "We describe the second-best run for the shared task on predicting the difficulty of Multi-Choice Questions (MCQs) in the medical domain. Our approach leverages prompting Large Language Models (LLMs). Rather than straightforwardly querying difficulty, we simulate medical candidate{'}s responses to questions across various scenarios. For this, more than 10,000 prompts were required for the 467 training questions and the 200 test questions. From the answers to these prompts, we extracted a set of features which we combined with a Ridge Regression to which we only adjusted the regularization parameter using the training set. Our motivation stems from the belief that MCQ difficulty is influenced more by the respondent population than by item-specific content features. We conclude that the approach is promising and has the potential to improve other item-based systems on this task, which turned out to be extremely challenging and has ample room for future improvement.", }

We describe the second-best run for the shared task on predicting the difficulty of Multi-Choice Questions (MCQs) in the medical domain. Our approach leverages prompting Large Language Models (LLMs). Rather than straightforwardly querying difficulty, we simulate medical candidate{'}s responses to questions across various scenarios. For this, more than 10,000 prompts were required for the 467 training questions and the 200 test questions. From the answers to these prompts, we extracted a set of features which we combined with a Ridge Regression to which we only adjusted the regularization parameter using the training set. Our motivation stems from the belief that MCQ difficulty is influenced more by the respondent population than by item-specific content features. We conclude that the approach is promising and has the potential to improve other item-based systems on this task, which turned out to be extremely challenging and has ample room for future improvement.

UPN-ICC at BEA 2024 Shared Task: Leveraging LLMs for Multiple-Choice Questions Difficulty Prediction

bea-1.47

Poster

https://aclanthology.org/2024.bea-1.48.bib

https://aclanthology.org/2024.bea-1.48/

@inproceedings{yousefpoori-naeim-etal-2024-using, title = "Using Machine Learning to Predict Item Difficulty and Response Time in Medical Tests", author = "Yousefpoori-Naeim, Mehrdad and Zargari, Shayan and Hatami, Zahra", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.48", pages = "551--560", abstract = "Prior knowledge of item characteristics, such as difficulty and response time, without pretesting items can substantially save time and cost in high-standard test development. Using a variety of machine learning (ML) algorithms, the present study explored several (non-)linguistic features (such as Coh-Metrix indices) along with MPNet word embeddings to predict the difficulty and response time of a sample of medical test items. In both prediction tasks, the contribution of embeddings to models already containing other features was found to be extremely limited. Moreover, a comparison of feature importance scores across the two prediction tasks revealed that cohesion-based features were the strongest predictors of difficulty, while the prediction of response time was primarily dependent on length-related features.", }

Prior knowledge of item characteristics, such as difficulty and response time, without pretesting items can substantially save time and cost in high-standard test development. Using a variety of machine learning (ML) algorithms, the present study explored several (non-)linguistic features (such as Coh-Metrix indices) along with MPNet word embeddings to predict the difficulty and response time of a sample of medical test items. In both prediction tasks, the contribution of embeddings to models already containing other features was found to be extremely limited. Moreover, a comparison of feature importance scores across the two prediction tasks revealed that cohesion-based features were the strongest predictors of difficulty, while the prediction of response time was primarily dependent on length-related features.

Using Machine Learning to Predict Item Difficulty and Response Time in Medical Tests

bea-1.48

Poster

https://aclanthology.org/2024.bea-1.49.bib

https://aclanthology.org/2024.bea-1.49/

@inproceedings{veeramani-etal-2024-large, title = "Large Language Model-based Pipeline for Item Difficulty and Response Time Estimation for Educational Assessments", author = "Veeramani, Hariram and Thapa, Surendrabikram and Shankar, Natarajan Balaji and Alwan, Abeer", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.49", pages = "561--566", abstract = "This work presents a novel framework for the automated prediction of item difficulty and response time within educational assessments. Utilizing data from the BEA 2024 Shared Task, we integrate Named Entity Recognition, Semantic Role Labeling, and linguistic features to prompt a Large Language Model (LLM). Our best approach achieves an RMSE of 0.308 for item difficulty and 27.474 for response time prediction, improving on the provided baseline. The framework{'}s adaptability is demonstrated on audio recordings of 3rd-8th graders from the Atlanta, Georgia area responding to the Test of Narrative Language. These results highlight the framework{'}s potential to enhance test development efficiency.", }

This work presents a novel framework for the automated prediction of item difficulty and response time within educational assessments. Utilizing data from the BEA 2024 Shared Task, we integrate Named Entity Recognition, Semantic Role Labeling, and linguistic features to prompt a Large Language Model (LLM). Our best approach achieves an RMSE of 0.308 for item difficulty and 27.474 for response time prediction, improving on the provided baseline. The framework{'}s adaptability is demonstrated on audio recordings of 3rd-8th graders from the Atlanta, Georgia area responding to the Test of Narrative Language. These results highlight the framework{'}s potential to enhance test development efficiency.

Large Language Model-based Pipeline for Item Difficulty and Response Time Estimation for Educational Assessments

bea-1.49

Poster

https://aclanthology.org/2024.bea-1.50.bib

https://aclanthology.org/2024.bea-1.50/

@inproceedings{rodrigo-etal-2024-uned, title = "{UNED} team at {BEA} 2024 Shared Task: Testing different Input Formats for predicting Item Difficulty and Response Time in Medical Exams", author = "Rodrigo, Alvaro and Moreno-{\'A}lvarez, Sergio and Pe{\~n}as, Anselmo", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.50", pages = "567--570", abstract = "This paper presents the description and primary outcomes of our team{'}s participation in the BEA 2024 shared task. Our primary exploration involved employing transformer-based systems, particularly BERT models, due to their suitability for Natural Language Processing tasks and efficiency with computational resources. We experimented with various input formats, including concatenating all text elements and incorporating only the clinical case. Surprisingly, our results revealed different impacts on predicting difficulty versus response time, with the former favoring clinical text only and the latter benefiting from including the correct answer. Despite moderate performance in difficulty prediction, our models excelled in response time prediction, ranking highest among all participants. This study lays the groundwork for future investigations into more complex approaches and configurations, aiming to advance the automatic prediction of exam difficulty and response time.", }

This paper presents the description and primary outcomes of our team{'}s participation in the BEA 2024 shared task. Our primary exploration involved employing transformer-based systems, particularly BERT models, due to their suitability for Natural Language Processing tasks and efficiency with computational resources. We experimented with various input formats, including concatenating all text elements and incorporating only the clinical case. Surprisingly, our results revealed different impacts on predicting difficulty versus response time, with the former favoring clinical text only and the latter benefiting from including the correct answer. Despite moderate performance in difficulty prediction, our models excelled in response time prediction, ranking highest among all participants. This study lays the groundwork for future investigations into more complex approaches and configurations, aiming to advance the automatic prediction of exam difficulty and response time.

UNED team at BEA 2024 Shared Task: Testing different Input Formats for predicting Item Difficulty and Response Time in Medical Exams

bea-1.50

Poster

https://aclanthology.org/2024.bea-1.51.bib

https://aclanthology.org/2024.bea-1.51/

@inproceedings{shardlow-etal-2024-bea, title = "The {BEA} 2024 Shared Task on the Multilingual Lexical Simplification Pipeline", author = {Shardlow, Matthew and Alva-Manchego, Fernando and Batista-Navarro, Riza and Bott, Stefan and Calderon Ramirez, Saul and Cardon, R{\'e}mi and Fran{\c{c}}ois, Thomas and Hayakawa, Akio and Horbach, Andrea and H{\"u}lsing, Anna and Ide, Yusuke and Imperial, Joseph Marvin and Nohejl, Adam and North, Kai and Occhipinti, Laura and Rojas, Nelson Per{\'e}z and Raihan, Nishat and Ranasinghe, Tharindu and Salazar, Martin Solis and {\v{S}}tajner, Sanja and Zampieri, Marcos and Saggion, Horacio}, editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.51", pages = "571--589", abstract = "We report the findings of the 2024 Multilingual Lexical Simplification Pipeline shared task. We released a new dataset comprising 5,927 instances of lexical complexity prediction and lexical simplification on common contexts across 10 languages, split into trial (300) and test (5,627). 10 teams participated across 2 tracks and 10 languages with 233 runs evaluated across all systems. Five teams participated in all languages for the lexical complexity prediction task and 4 teams participated in all languages for the lexical simplification task. Teams employed a range of strategies, making use of open and closed source large language models for lexical simplification, as well as feature-based approaches for lexical complexity prediction. The highest scoring team on the combined multilingual data was able to obtain a Pearson{'}s correlation of 0.6241 and an ACC@1@Top1 of 0.3772, both demonstrating that there is still room for improvement on two difficult sub-tasks of the lexical simplification pipeline.", }

We report the findings of the 2024 Multilingual Lexical Simplification Pipeline shared task. We released a new dataset comprising 5,927 instances of lexical complexity prediction and lexical simplification on common contexts across 10 languages, split into trial (300) and test (5,627). 10 teams participated across 2 tracks and 10 languages with 233 runs evaluated across all systems. Five teams participated in all languages for the lexical complexity prediction task and 4 teams participated in all languages for the lexical simplification task. Teams employed a range of strategies, making use of open and closed source large language models for lexical simplification, as well as feature-based approaches for lexical complexity prediction. The highest scoring team on the combined multilingual data was able to obtain a Pearson{'}s correlation of 0.6241 and an ACC@1@Top1 of 0.3772, both demonstrating that there is still room for improvement on two difficult sub-tasks of the lexical simplification pipeline.

The BEA 2024 Shared Task on the Multilingual Lexical Simplification Pipeline

bea-1.51

Poster

https://aclanthology.org/2024.bea-1.52.bib

https://aclanthology.org/2024.bea-1.52/

@inproceedings{enomoto-etal-2024-tmu, title = "{TMU}-{HIT} at {MLSP} 2024: How Well Can {GPT}-4 Tackle Multilingual Lexical Simplification?", author = "Enomoto, Taisei and Kim, Hwichan and Hirasawa, Tosho and Nagai, Yoshinari and Sato, Ayako and Nakajima, Kyotaro and Komachi, Mamoru", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.52", pages = "590--598", abstract = "Lexical simplification (LS) is a process of replacing complex words with simpler alternatives to help readers understand sentences seamlessly. This process is divided into two primary subtasks: assessing word complexities and replacing high-complexity words with simpler alternatives. Employing task-specific supervised data to train models is a prevalent strategy for addressing these subtasks. However, such approach cannot be employed for low-resource languages. Therefore, this paper introduces a multilingual LS pipeline system that does not rely on supervised data. Specifically, we have developed systems based on GPT-4 for each subtask. Our systems demonstrated top-class performance on both tasks in many languages. The results indicate that GPT-4 can effectively assess lexical complexity and simplify complex words in a multilingual context with high quality.", }

Lexical simplification (LS) is a process of replacing complex words with simpler alternatives to help readers understand sentences seamlessly. This process is divided into two primary subtasks: assessing word complexities and replacing high-complexity words with simpler alternatives. Employing task-specific supervised data to train models is a prevalent strategy for addressing these subtasks. However, such approach cannot be employed for low-resource languages. Therefore, this paper introduces a multilingual LS pipeline system that does not rely on supervised data. Specifically, we have developed systems based on GPT-4 for each subtask. Our systems demonstrated top-class performance on both tasks in many languages. The results indicate that GPT-4 can effectively assess lexical complexity and simplify complex words in a multilingual context with high quality.

TMU-HIT at MLSP 2024: How Well Can GPT-4 Tackle Multilingual Lexical Simplification?

bea-1.52

Poster

https://aclanthology.org/2024.bea-1.53.bib

https://aclanthology.org/2024.bea-1.53/

@inproceedings{seneviratne-suominen-2024-anu, title = "{ANU} at {MLSP}-2024: Prompt-based Lexical Simplification for {E}nglish and {S}inhala", author = "Seneviratne, Sandaru and Suominen, Hanna", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.53", pages = "599--604", abstract = "Lexical simplification, the process of simplifying complex content in text without any modifications to the syntactical structure of text, plays a crucial role in enhancing comprehension and accessibility. This paper presents an approach to lexical simplification that relies on the capabilities of generative Artificial Intelligence (AI) models to predict the complexity of words and substitute complex words with simpler alternatives. Early lexical simplification methods predominantly relied on rule-based approaches, transitioning gradually to machine learning and deep learning techniques, leveraging contextual embeddings from large language models. However, the the emergence of generative AI models revolutionized the landscape of natural language processing, including lexical simplification. In this study, we proposed a straightforward yet effective method that employs generative AI models for both predicting lexical complexity and generating appropriate substitutions. To predict lexical complexity, we adopted three distinct types of prompt templates, while for lexical substitution, we employed three prompt templates alongside an ensemble approach. Extending our experimentation to include both English and Sinhala data, our approach demonstrated comparable performance across both languages, with particular strengths in lexical substitution.", }

Lexical simplification, the process of simplifying complex content in text without any modifications to the syntactical structure of text, plays a crucial role in enhancing comprehension and accessibility. This paper presents an approach to lexical simplification that relies on the capabilities of generative Artificial Intelligence (AI) models to predict the complexity of words and substitute complex words with simpler alternatives. Early lexical simplification methods predominantly relied on rule-based approaches, transitioning gradually to machine learning and deep learning techniques, leveraging contextual embeddings from large language models. However, the the emergence of generative AI models revolutionized the landscape of natural language processing, including lexical simplification. In this study, we proposed a straightforward yet effective method that employs generative AI models for both predicting lexical complexity and generating appropriate substitutions. To predict lexical complexity, we adopted three distinct types of prompt templates, while for lexical substitution, we employed three prompt templates alongside an ensemble approach. Extending our experimentation to include both English and Sinhala data, our approach demonstrated comparable performance across both languages, with particular strengths in lexical substitution.

ANU at MLSP-2024: Prompt-based Lexical Simplification for English and Sinhala

bea-1.53

Poster

https://aclanthology.org/2024.bea-1.54.bib

https://aclanthology.org/2024.bea-1.54/

@inproceedings{dutilleul-etal-2024-isep, title = "{ISEP}{\_}{P}residency{\_}{U}niversity at {MLSP} 2024 Shared Task: Using {GPT}-3.5 to Generate Substitutes for Lexical Simplification", author = "Dutilleul, Benjamin and Debaillon, Mathis and Mathias, Sandeep", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.54", pages = "605--609", abstract = "Lexical substitute generation is a task where we generate substitutes for a given word to fit in the required context. It is one of the main steps for automatic lexical simplifcation. In this paper, we introduce an automatic lexical simplification system using the GPT-3 large language model. The system generates simplified candidate substitutions for complex words to aid readability and comprehension for the reader. The paper describes the system that we submitted for the Multilingual Lexical Simplification Pipeline Shared Task at the 2024 BEA Workshop. During the shared task, we experimented with Catalan, English, French, Italian, Portuguese, and German for the Lexical Simplification Shared Task. We achieved the best results in Catalan and Portuguese, and were runners-up in English, French and Italian. To further research in this domain, we also release our code upon acceptance of the paper.", }

Lexical substitute generation is a task where we generate substitutes for a given word to fit in the required context. It is one of the main steps for automatic lexical simplifcation. In this paper, we introduce an automatic lexical simplification system using the GPT-3 large language model. The system generates simplified candidate substitutions for complex words to aid readability and comprehension for the reader. The paper describes the system that we submitted for the Multilingual Lexical Simplification Pipeline Shared Task at the 2024 BEA Workshop. During the shared task, we experimented with Catalan, English, French, Italian, Portuguese, and German for the Lexical Simplification Shared Task. We achieved the best results in Catalan and Portuguese, and were runners-up in English, French and Italian. To further research in this domain, we also release our code upon acceptance of the paper.

ISEP_Presidency_University at MLSP 2024 Shared Task: Using GPT-3.5 to Generate Substitutes for Lexical Simplification

bea-1.54

Poster

https://aclanthology.org/2024.bea-1.55.bib

https://aclanthology.org/2024.bea-1.55/

@inproceedings{cristea-nisioi-2024-machine, title = "Archaeology at MLSP 2024: Machine Translation for Lexical Complexity Prediction and Lexical Simplification", author = "Cristea, Petru and Nisioi, Sergiu", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.55", pages = "610--617", abstract = "We present the submissions of team Archaeology for the Lexical Simplification and Lexical Complexity Prediction Shared Tasks at BEA2024. Our approach for this shared task consists in creating two pipelines for generating lexical substitutions and estimating the complexity: one using machine translation texts into English and one using the original language.For the LCP subtask, our xgb regressor is trained with engineered features (based primarily on English language resources) and shallow word structure features. For the LS subtask we use a locally-executed quantized LLM to generate candidates and sort them by complexity score computed using the pipeline designed for LCP.These pipelines provide distinct perspectives on the lexical simplification process, offering insights into the efficacy and limitations of employing Machine Translation versus direct processing on the original language data.", }

We present the submissions of team Archaeology for the Lexical Simplification and Lexical Complexity Prediction Shared Tasks at BEA2024. Our approach for this shared task consists in creating two pipelines for generating lexical substitutions and estimating the complexity: one using machine translation texts into English and one using the original language.For the LCP subtask, our xgb regressor is trained with engineered features (based primarily on English language resources) and shallow word structure features. For the LS subtask we use a locally-executed quantized LLM to generate candidates and sort them by complexity score computed using the pipeline designed for LCP.These pipelines provide distinct perspectives on the lexical simplification process, offering insights into the efficacy and limitations of employing Machine Translation versus direct processing on the original language data.

Archaeology at MLSP 2024: Machine Translation for Lexical Complexity Prediction and Lexical Simplification

bea-1.55

Poster

https://aclanthology.org/2024.bea-1.56.bib

https://aclanthology.org/2024.bea-1.56/

@inproceedings{sastre-etal-2024-retuyt, title = "{RETUYT}-{INCO} at {MLSP} 2024: Experiments on Language Simplification using Embeddings, Classifiers and Large Language Models", author = "Sastre, Ignacio and Alfonso, Leandro and Fleitas, Facundo and Gil, Federico and Lucas, Andr{\'e}s and Spoturno, Tom{\'a}s and G{\'o}ngora, Santiago and Ros{\'a}, Aiala and Chiruzzo, Luis", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.56", pages = "618--626", abstract = "In this paper we present the participation of the RETUYT-INCO team at the BEA-MLSP 2024 shared task. We followed different approaches, from Multilayer Perceptron models with word embeddings to Large Language Models fine-tuned on different datasets: already existing, crowd-annotated, and synthetic.Our best models are based on fine-tuning Mistral-7B, either with a manually annotated dataset or with synthetic data.", }

In this paper we present the participation of the RETUYT-INCO team at the BEA-MLSP 2024 shared task. We followed different approaches, from Multilayer Perceptron models with word embeddings to Large Language Models fine-tuned on different datasets: already existing, crowd-annotated, and synthetic.Our best models are based on fine-tuning Mistral-7B, either with a manually annotated dataset or with synthetic data.

RETUYT-INCO at MLSP 2024: Experiments on Language Simplification using Embeddings, Classifiers and Large Language Models

bea-1.56

Poster

https://aclanthology.org/2024.bea-1.57.bib

https://aclanthology.org/2024.bea-1.57/

@inproceedings{goswami-etal-2024-gmu, title = "{GMU} at {MLSP} 2024: Multilingual Lexical Simplification with Transformer Models", author = "Goswami, Dhiman and North, Kai and Zampieri, Marcos", editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.57", pages = "627--634", abstract = "This paper presents GMU{'}s submission to the Multilingual Lexical Simplification Pipeline (MLSP) shared task at the BEA workshop 2024. The task includes Lexical Complexity Prediction (LCP) and Lexical Simplification (LS) sub-tasks across 10 languages. Our submissions achieved rankings ranging from 1st to 5th in LCP and from 1st to 3rd in LS. Our best performing approach for LCP is a weighted ensemble based on Pearson correlation of language specific transformer models trained on all languages combined. For LS, GPT4-turbo zero-shot prompting achieved the best performance.", }

This paper presents GMU{'}s submission to the Multilingual Lexical Simplification Pipeline (MLSP) shared task at the BEA workshop 2024. The task includes Lexical Complexity Prediction (LCP) and Lexical Simplification (LS) sub-tasks across 10 languages. Our submissions achieved rankings ranging from 1st to 5th in LCP and from 1st to 3rd in LS. Our best performing approach for LCP is a weighted ensemble based on Pearson correlation of language specific transformer models trained on all languages combined. For LS, GPT4-turbo zero-shot prompting achieved the best performance.

GMU at MLSP 2024: Multilingual Lexical Simplification with Transformer Models

bea-1.57

Poster

https://aclanthology.org/2024.bea-1.58.bib

https://aclanthology.org/2024.bea-1.58/

@inproceedings{tack-2024-itec, title = "{ITEC} at {MLSP} 2024: Transferring Predictions of Lexical Difficulty from Non-Native Readers", author = {Tack, Ana{\"\i}s}, editor = {Kochmar, Ekaterina and Bexte, Marie and Burstein, Jill and Horbach, Andrea and Laarmann-Quante, Ronja and Tack, Ana{\"\i}s and Yaneva, Victoria and Yuan, Zheng}, booktitle = "Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.bea-1.58", pages = "635--639", abstract = "This paper presents the results of our team{'}s participation in the BEA 2024 shared task on the multilingual lexical simplification pipeline (MLSP; Shardlow et al., 2024). During the task, organizers supplied data that combined two components of the simplification pipeline: lexical complexity prediction and lexical substitution. This dataset encompassed ten languages, including French. Given the absence of dedicated training data, teams were challenged with employing systems trained on pre-existing resources and evaluating their performance on unexplored test data.Our team contributed to the task using previously developed models for predicting lexical difficulty in French (Tack, 2021). These models were built on deep learning architectures, adding to our participation in the CWI 2018 shared task (De Hertog and Tack, 2018). The training dataset comprised 262,054 binary decision annotations, capturing perceived lexical difficulty, collected from a sample of 56 non-native French readers. Two pre-trained neural logistic models were used: (1) a model for predicting difficulty for words within their sentence context, and (2) a model for predicting difficulty for isolated words.The findings revealed that despite being trained for a distinct prediction task (as indicated by a negative R2 fit), transferring the logistic predictions of lexical difficulty to continuous scores of lexical complexity exhibited a positive correlation. Specifically, the results indicated that isolated predictions exhibited a higher correlation (r = .36) compared to contextualized predictions (r = .33). Moreover, isolated predictions demonstrated a remarkably higher Spearman rank correlation (ρ = .50) than contextualized predictions (ρ = .35). These results align with earlier observations by Tack (2021), suggesting that the ground truth primarily captures more lexical access difficulties than word-to-context integration problems.", }

This paper presents the results of our team{'}s participation in the BEA 2024 shared task on the multilingual lexical simplification pipeline (MLSP; Shardlow et al., 2024). During the task, organizers supplied data that combined two components of the simplification pipeline: lexical complexity prediction and lexical substitution. This dataset encompassed ten languages, including French. Given the absence of dedicated training data, teams were challenged with employing systems trained on pre-existing resources and evaluating their performance on unexplored test data.Our team contributed to the task using previously developed models for predicting lexical difficulty in French (Tack, 2021). These models were built on deep learning architectures, adding to our participation in the CWI 2018 shared task (De Hertog and Tack, 2018). The training dataset comprised 262,054 binary decision annotations, capturing perceived lexical difficulty, collected from a sample of 56 non-native French readers. Two pre-trained neural logistic models were used: (1) a model for predicting difficulty for words within their sentence context, and (2) a model for predicting difficulty for isolated words.The findings revealed that despite being trained for a distinct prediction task (as indicated by a negative R2 fit), transferring the logistic predictions of lexical difficulty to continuous scores of lexical complexity exhibited a positive correlation. Specifically, the results indicated that isolated predictions exhibited a higher correlation (r = .36) compared to contextualized predictions (r = .33). Moreover, isolated predictions demonstrated a remarkably higher Spearman rank correlation (ρ = .50) than contextualized predictions (ρ = .35). These results align with earlier observations by Tack (2021), suggesting that the ground truth primarily captures more lexical access difficulties than word-to-context integration problems.

ITEC at MLSP 2024: Transferring Predictions of Lexical Difficulty from Non-Native Readers

bea-1.58

Poster

https://aclanthology.org/2024.clinicalnlp-1.1.bib

https://aclanthology.org/2024.clinicalnlp-1.1/

@inproceedings{chen-hirschberg-2024-exploring, title = "Exploring Robustness in Doctor-Patient Conversation Summarization: An Analysis of Out-of-Domain {SOAP} Notes", author = "Chen, Yu-Wen and Hirschberg, Julia", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.1", doi = "10.18653/v1/2024.clinicalnlp-1.1", pages = "1--9", abstract = "Summarizing medical conversations poses unique challenges due to the specialized domain and the difficulty of collecting in-domain training data. In this study, we investigate the performance of state-of-the-art doctor-patient conversation generative summarization models on the out-of-domain data. We divide the summarization model of doctor-patient conversation into two configurations: (1) a general model, without specifying subjective (S), objective (O), and assessment (A) and plan (P) notes; (2) a SOAP-oriented model that generates a summary with SOAP sections. We analyzed the limitations and strengths of the fine-tuning language model-based methods and GPTs on both configurations. We also conducted a Linguistic Inquiry and Word Count analysis to compare the SOAP notes from different datasets. The results exhibit a strong correlation for reference notes across different datasets, indicating that format mismatch (i.e., discrepancies in word distribution) is not the main cause of performance decline on out-of-domain data. Lastly, a detailed analysis of SOAP notes is included to provide insights into missing information and hallucinations introduced by the models.", }

Summarizing medical conversations poses unique challenges due to the specialized domain and the difficulty of collecting in-domain training data. In this study, we investigate the performance of state-of-the-art doctor-patient conversation generative summarization models on the out-of-domain data. We divide the summarization model of doctor-patient conversation into two configurations: (1) a general model, without specifying subjective (S), objective (O), and assessment (A) and plan (P) notes; (2) a SOAP-oriented model that generates a summary with SOAP sections. We analyzed the limitations and strengths of the fine-tuning language model-based methods and GPTs on both configurations. We also conducted a Linguistic Inquiry and Word Count analysis to compare the SOAP notes from different datasets. The results exhibit a strong correlation for reference notes across different datasets, indicating that format mismatch (i.e., discrepancies in word distribution) is not the main cause of performance decline on out-of-domain data. Lastly, a detailed analysis of SOAP notes is included to provide insights into missing information and hallucinations introduced by the models.

Exploring Robustness in Doctor-Patient Conversation Summarization: An Analysis of Out-of-Domain SOAP Notes

clinicalnlp-1.1

Poster

2406.02826

https://aclanthology.org/2024.clinicalnlp-1.2.bib

https://aclanthology.org/2024.clinicalnlp-1.2/

@inproceedings{khlaut-etal-2024-efficient, title = "Efficient Medical Question Answering with Knowledge-Augmented Question Generation", author = "Khlaut, Julien and Dancette, Corentin and Ferreres, Elodie and Alaedine, Benani and Herent, Herent and Manceron, Pierre", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.2", doi = "10.18653/v1/2024.clinicalnlp-1.2", pages = "10--20", abstract = "In the expanding field of language model applications, medical knowledge representation remains a significant challenge due to the specialized nature of the domain. Large language models, such as GPT-4, obtain reasonable scores on medical question-answering tasks, but smaller models are far behind.In this work, we introduce a method to improve the proficiency of a small language model in the medical domain by employing a two-fold approach. We first fine-tune the model on a corpus of medical textbooks. Then, we use GPT-4 to generate questions similar to the downstream task, prompted with textbook knowledge, and use them to fine-tune the model. Additionally, we introduce ECN-QA, a novel Medical QA dataset containing {``}progressive questions{''} composed of related sequential questions. We show the benefits of our training strategy on this dataset.The study{'}s findings highlight the potential of small language models in the medical domain when appropriately fine-tuned.", }

In the expanding field of language model applications, medical knowledge representation remains a significant challenge due to the specialized nature of the domain. Large language models, such as GPT-4, obtain reasonable scores on medical question-answering tasks, but smaller models are far behind.In this work, we introduce a method to improve the proficiency of a small language model in the medical domain by employing a two-fold approach. We first fine-tune the model on a corpus of medical textbooks. Then, we use GPT-4 to generate questions similar to the downstream task, prompted with textbook knowledge, and use them to fine-tune the model. Additionally, we introduce ECN-QA, a novel Medical QA dataset containing {``}progressive questions{''} composed of related sequential questions. We show the benefits of our training strategy on this dataset.The study{'}s findings highlight the potential of small language models in the medical domain when appropriately fine-tuned.

Efficient Medical Question Answering with Knowledge-Augmented Question Generation

clinicalnlp-1.2

Poster

2405.14654

https://aclanthology.org/2024.clinicalnlp-1.3.bib

https://aclanthology.org/2024.clinicalnlp-1.3/

@inproceedings{pal-sankarasubbu-2024-gemini, title = "Gemini Goes to {M}ed School: Exploring the Capabilities of Multimodal Large Language Models on Medical Challenge Problems {\&} Hallucinations", author = "Pal, Ankit and Sankarasubbu, Malaikannan", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.3", doi = "10.18653/v1/2024.clinicalnlp-1.3", pages = "21--46", abstract = "Large language models have the potential to be valuable in the healthcare industry, but it{'}s crucial to verify their safety and effectiveness through rigorous evaluation. In our study, we evaluated LLMs, including Google{'}s Gemini, across various medical tasks. Despite Gemini{'}s capabilities, it underperformed compared to leading models like MedPaLM 2 and GPT-4, particularly in medical visual question answering (VQA), with a notable accuracy gap (Gemini at 61.45{\%} vs. GPT-4V at 88{\%}). Our analysis revealed that Gemini is highly susceptible to hallucinations, overconfidence, and knowledge gaps, which indicate risks if deployed uncritically. We also performed a detailed analysis by medical subject and test type, providing actionable feedback for developers and clinicians. To mitigate risks, we implemented effective prompting strategies, improving performance, and contributed to the field by releasing a Python module for medical LLM evaluation and establishing a leaderboard on Hugging Face for ongoing research and development. Python module can be found at https://github.com/promptslab/RosettaEval", }

Large language models have the potential to be valuable in the healthcare industry, but it{'}s crucial to verify their safety and effectiveness through rigorous evaluation. In our study, we evaluated LLMs, including Google{'}s Gemini, across various medical tasks. Despite Gemini{'}s capabilities, it underperformed compared to leading models like MedPaLM 2 and GPT-4, particularly in medical visual question answering (VQA), with a notable accuracy gap (Gemini at 61.45{\%} vs. GPT-4V at 88{\%}). Our analysis revealed that Gemini is highly susceptible to hallucinations, overconfidence, and knowledge gaps, which indicate risks if deployed uncritically. We also performed a detailed analysis by medical subject and test type, providing actionable feedback for developers and clinicians. To mitigate risks, we implemented effective prompting strategies, improving performance, and contributed to the field by releasing a Python module for medical LLM evaluation and establishing a leaderboard on Hugging Face for ongoing research and development. Python module can be found at https://github.com/promptslab/RosettaEval

Gemini Goes to Med School: Exploring the Capabilities of Multimodal Large Language Models on Medical Challenge Problems & Hallucinations

clinicalnlp-1.3

Poster

2402.07023

https://huggingface.co/papers/2402.07023

https://aclanthology.org/2024.clinicalnlp-1.4.bib

https://aclanthology.org/2024.clinicalnlp-1.4/

@inproceedings{ziletti-dambrosi-2024-retrieval, title = "Retrieval augmented text-to-{SQL} generation for epidemiological question answering using electronic health records", author = "Ziletti, Angelo and DAmbrosi, Leonardo", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.4", doi = "10.18653/v1/2024.clinicalnlp-1.4", pages = "47--53", abstract = "Electronic health records (EHR) and claims data are rich sources of real-world data that reflect patient health status and healthcare utilization. Querying these databases to answer epidemiological questions is challenging due to the intricacy of medical terminology and the need for complex SQL queries. Here, we introduce an end-to-end methodology that combines text-to-SQL generation with retrieval augmented generation (RAG) to answer epidemiological questions using EHR and claims data. We show that our approach, which integrates a medical coding step into the text-to-SQL process, significantly improves the performance over simple prompting. Our findings indicate that although current language models are not yet sufficiently accurate for unsupervised use, RAG offers a promising direction for improving their capabilities, as shown in a realistic industry setting.", }

Electronic health records (EHR) and claims data are rich sources of real-world data that reflect patient health status and healthcare utilization. Querying these databases to answer epidemiological questions is challenging due to the intricacy of medical terminology and the need for complex SQL queries. Here, we introduce an end-to-end methodology that combines text-to-SQL generation with retrieval augmented generation (RAG) to answer epidemiological questions using EHR and claims data. We show that our approach, which integrates a medical coding step into the text-to-SQL process, significantly improves the performance over simple prompting. Our findings indicate that although current language models are not yet sufficiently accurate for unsupervised use, RAG offers a promising direction for improving their capabilities, as shown in a realistic industry setting.

Retrieval augmented text-to-SQL generation for epidemiological question answering using electronic health records

clinicalnlp-1.4

Poster

2403.09226

https://aclanthology.org/2024.clinicalnlp-1.5.bib

https://aclanthology.org/2024.clinicalnlp-1.5/

@inproceedings{yang-etal-2024-clinicalmamba, title = "{C}linical{M}amba: A Generative Clinical Language Model on Longitudinal Clinical Notes", author = "Yang, Zhichao and Mitra, Avijit and Kwon, Sunjae and Yu, Hong", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.5", doi = "10.18653/v1/2024.clinicalnlp-1.5", pages = "54--63", abstract = "The advancement of natural language processing (NLP) systems in healthcare hinges on language models{'} ability to interpret the intricate information contained within clinical notes. This process often requires integrating information from various time points in a patient{'}s medical history. However, most earlier clinical language models were pretrained with a context length limited to roughly one clinical document. In this study, We introduce ClinicalMamba, a specialized version of the Mamba language model, pretrained on a vast corpus of longitudinal clinical notes to address the unique linguistic characteristics and information processing needs of the medical domain. ClinicalMamba models, with 130 million and 2.8 billion parameters, demonstrate superior performance in modeling clinical language across extended text lengths compared to Mamba and other clinical models based on longformer and Llama. With few-shot learning, ClinicalMamba achieves notable benchmarks in speed and performance, outperforming existing clinical language models and large language models like GPT-4 in longitudinal clinical tasks.", }

The advancement of natural language processing (NLP) systems in healthcare hinges on language models{'} ability to interpret the intricate information contained within clinical notes. This process often requires integrating information from various time points in a patient{'}s medical history. However, most earlier clinical language models were pretrained with a context length limited to roughly one clinical document. In this study, We introduce ClinicalMamba, a specialized version of the Mamba language model, pretrained on a vast corpus of longitudinal clinical notes to address the unique linguistic characteristics and information processing needs of the medical domain. ClinicalMamba models, with 130 million and 2.8 billion parameters, demonstrate superior performance in modeling clinical language across extended text lengths compared to Mamba and other clinical models based on longformer and Llama. With few-shot learning, ClinicalMamba achieves notable benchmarks in speed and performance, outperforming existing clinical language models and large language models like GPT-4 in longitudinal clinical tasks.

ClinicalMamba: A Generative Clinical Language Model on Longitudinal Clinical Notes

clinicalnlp-1.5

Poster

2403.05795

https://huggingface.co/papers/2403.05795

https://aclanthology.org/2024.clinicalnlp-1.6.bib

https://aclanthology.org/2024.clinicalnlp-1.6/

@inproceedings{lin-etal-2024-working, title = "Working Alliance Transformer for Psychotherapy Dialogue Classification", author = "Lin, Baihan and Cecchi, Guillermo and Bouneffouf, Djallel", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.6", doi = "10.18653/v1/2024.clinicalnlp-1.6", pages = "64--69", abstract = "As a predictive measure of the treatment outcome in psychotherapy, the working alliance measures the agreement of the patient and the therapist in terms of their bond, task and goal. Long been a clinical quantity estimated by the patients{'} and therapists{'} self-evaluative reports, we believe that the working alliance can be better characterized using natural language processing technique directly in the dialogue transcribed in each therapy session. In this work, we propose the Working Alliance Transformer (WAT), a Transformer-based classification model that has a psychological state encoder which infers the working alliance scores by projecting the embedding of the dialogues turns onto the embedding space of the clinical inventory for working alliance. We evaluate our method in a real-world dataset with over 950 therapy sessions with anxiety, depression, schizophrenia and suicidal patients and demonstrate an empirical advantage of using information about therapeutic states in the sequence classification task of psychotherapy dialogues.", }

As a predictive measure of the treatment outcome in psychotherapy, the working alliance measures the agreement of the patient and the therapist in terms of their bond, task and goal. Long been a clinical quantity estimated by the patients{'} and therapists{'} self-evaluative reports, we believe that the working alliance can be better characterized using natural language processing technique directly in the dialogue transcribed in each therapy session. In this work, we propose the Working Alliance Transformer (WAT), a Transformer-based classification model that has a psychological state encoder which infers the working alliance scores by projecting the embedding of the dialogues turns onto the embedding space of the clinical inventory for working alliance. We evaluate our method in a real-world dataset with over 950 therapy sessions with anxiety, depression, schizophrenia and suicidal patients and demonstrate an empirical advantage of using information about therapeutic states in the sequence classification task of psychotherapy dialogues.

Working Alliance Transformer for Psychotherapy Dialogue Classification

clinicalnlp-1.6

Poster

2210.15603

https://aclanthology.org/2024.clinicalnlp-1.7.bib

https://aclanthology.org/2024.clinicalnlp-1.7/

@inproceedings{liang-sonntag-2024-building, title = "Building A {G}erman Clinical Named Entity Recognition System without In-domain Training Data", author = "Liang, Siting and Sonntag, Daniel", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.7", doi = "10.18653/v1/2024.clinicalnlp-1.7", pages = "70--81", abstract = "Clinical Named Entity Recognition (NER) is essential for extracting important medical insights from clinical narratives. Given the challenges in obtaining expert training datasets for real-world clinical applications related to data protection regulations and the lack of standardised entity types, this work represents a collaborative initiative aimed at building a German clinical NER system with a focus on addressing these obstacles effectively. In response to the challenge of training data scarcity, we propose a Conditional Relevance Learning (CRL) approach in low-resource transfer learning scenarios. CRL effectively leverages a pre-trained language model and domain-specific open resources, enabling the acquisition of a robust base model tailored for clinical NER tasks, particularly in the face of changing label sets. This flexibility empowers the implementation of a Multilayered Semantic Annotation (MSA) schema in our NER system, capable of organizing a diverse array of entity types, thus significantly boosting the NER system{'}s adaptability and utility across various clinical domains. In the case study, we demonstrate how our NER system can be applied to overcome resource constraints and comply with data privacy regulations. Lacking prior training on in-domain data, feedback from expert users in respective domains is essential in identifying areas for system refinement. Future work will focus on the integration of expert feedback to improve system performance in specific clinical contexts.", }

Clinical Named Entity Recognition (NER) is essential for extracting important medical insights from clinical narratives. Given the challenges in obtaining expert training datasets for real-world clinical applications related to data protection regulations and the lack of standardised entity types, this work represents a collaborative initiative aimed at building a German clinical NER system with a focus on addressing these obstacles effectively. In response to the challenge of training data scarcity, we propose a Conditional Relevance Learning (CRL) approach in low-resource transfer learning scenarios. CRL effectively leverages a pre-trained language model and domain-specific open resources, enabling the acquisition of a robust base model tailored for clinical NER tasks, particularly in the face of changing label sets. This flexibility empowers the implementation of a Multilayered Semantic Annotation (MSA) schema in our NER system, capable of organizing a diverse array of entity types, thus significantly boosting the NER system{'}s adaptability and utility across various clinical domains. In the case study, we demonstrate how our NER system can be applied to overcome resource constraints and comply with data privacy regulations. Lacking prior training on in-domain data, feedback from expert users in respective domains is essential in identifying areas for system refinement. Future work will focus on the integration of expert feedback to improve system performance in specific clinical contexts.

Building A German Clinical Named Entity Recognition System without In-domain Training Data

clinicalnlp-1.7

Poster

https://aclanthology.org/2024.clinicalnlp-1.8.bib

https://aclanthology.org/2024.clinicalnlp-1.8/

@inproceedings{burdisso-etal-2024-daic, title = "{DAIC}-{WOZ}: On the Validity of Using the Therapist{'}s prompts in Automatic Depression Detection from Clinical Interviews", author = "Burdisso, Sergio and Reyes-Ram{\'\i}rez, Ernesto and Villatoro-tello, Esa{\'u} and S{\'a}nchez-Vega, Fernando and Lopez Monroy, Adrian and Motlicek, Petr", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.8", doi = "10.18653/v1/2024.clinicalnlp-1.8", pages = "82--90", abstract = "Automatic depression detection from conversational data has gained significant interest in recent years.The DAIC-WOZ dataset, interviews conducted by a human-controlled virtual agent, has been widely used for this task.Recent studies have reported enhanced performance when incorporating interviewer{'}s prompts into the model.In this work, we hypothesize that this improvement might be mainly due to a bias present in these prompts, rather than the proposed architectures and methods.Through ablation experiments and qualitative analysis, we discover that models using interviewer{'}s prompts learn to focus on a specific region of the interviews, where questions about past experiences with mental health issues are asked, and use them as discriminative shortcuts to detect depressed participants. In contrast, models using participant responses gather evidence from across the entire interview.Finally, to highlight the magnitude of this bias, we achieve a 0.90 F1 score by intentionally exploiting it, the highest result reported to date on this dataset using only textual information.Our findings underline the need for caution when incorporating interviewers{'} prompts into models, as they may inadvertently learn to exploit targeted prompts, rather than learning to characterize the language and behavior that are genuinely indicative of the patient{'}s mental health condition.", }

Automatic depression detection from conversational data has gained significant interest in recent years.The DAIC-WOZ dataset, interviews conducted by a human-controlled virtual agent, has been widely used for this task.Recent studies have reported enhanced performance when incorporating interviewer{'}s prompts into the model.In this work, we hypothesize that this improvement might be mainly due to a bias present in these prompts, rather than the proposed architectures and methods.Through ablation experiments and qualitative analysis, we discover that models using interviewer{'}s prompts learn to focus on a specific region of the interviews, where questions about past experiences with mental health issues are asked, and use them as discriminative shortcuts to detect depressed participants. In contrast, models using participant responses gather evidence from across the entire interview.Finally, to highlight the magnitude of this bias, we achieve a 0.90 F1 score by intentionally exploiting it, the highest result reported to date on this dataset using only textual information.Our findings underline the need for caution when incorporating interviewers{'} prompts into models, as they may inadvertently learn to exploit targeted prompts, rather than learning to characterize the language and behavior that are genuinely indicative of the patient{'}s mental health condition.

DAIC-WOZ: On the Validity of Using the Therapist's prompts in Automatic Depression Detection from Clinical Interviews

clinicalnlp-1.8

Poster

2404.14463

https://aclanthology.org/2024.clinicalnlp-1.9.bib

https://aclanthology.org/2024.clinicalnlp-1.9/

@inproceedings{gema-etal-2024-parameter, title = "Parameter-Efficient Fine-Tuning of {LL}a{MA} for the Clinical Domain", author = "Gema, Aryo and Minervini, Pasquale and Daines, Luke and Hope, Tom and Alex, Beatrice", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.9", doi = "10.18653/v1/2024.clinicalnlp-1.9", pages = "91--104", abstract = "Adapting pretrained language models to novel domains, such as clinical applications, traditionally involves retraining their entire set of parameters. Parameter-Efficient Fine-Tuning (PEFT) techniques for fine-tuning language models significantly reduce computational requirements by selectively fine-tuning small subsets of parameters. In this study, we propose a two-step PEFT framework and evaluate it in the clinical domain. Our approach combines a specialised PEFT adapter layer designed for clinical domain adaptation with another adapter specialised for downstream tasks. We evaluate the framework on multiple clinical outcome prediction datasets, comparing it to clinically trained language models. Our framework achieves a better AUROC score averaged across all clinical downstream tasks compared to clinical language models. In particular, we observe large improvements of 4-5{\%} AUROC in large-scale multilabel classification tasks, such as diagnoses and procedures classification. To our knowledge, this study is the first to provide an extensive empirical analysis of the interplay between PEFT techniques and domain adaptation in an important real-world domain of clinical applications.", }

Adapting pretrained language models to novel domains, such as clinical applications, traditionally involves retraining their entire set of parameters. Parameter-Efficient Fine-Tuning (PEFT) techniques for fine-tuning language models significantly reduce computational requirements by selectively fine-tuning small subsets of parameters. In this study, we propose a two-step PEFT framework and evaluate it in the clinical domain. Our approach combines a specialised PEFT adapter layer designed for clinical domain adaptation with another adapter specialised for downstream tasks. We evaluate the framework on multiple clinical outcome prediction datasets, comparing it to clinically trained language models. Our framework achieves a better AUROC score averaged across all clinical downstream tasks compared to clinical language models. In particular, we observe large improvements of 4-5{\%} AUROC in large-scale multilabel classification tasks, such as diagnoses and procedures classification. To our knowledge, this study is the first to provide an extensive empirical analysis of the interplay between PEFT techniques and domain adaptation in an important real-world domain of clinical applications.

Parameter-Efficient Fine-Tuning of LLaMA for the Clinical Domain

clinicalnlp-1.9

Poster

2307.03042

https://aclanthology.org/2024.clinicalnlp-1.10.bib

https://aclanthology.org/2024.clinicalnlp-1.10/

@inproceedings{sanchez-carmona-etal-2024-multilevel, title = "A Multilevel Analysis of {P}ub{M}ed-only {BERT}-based Biomedical Models", author = "Sanchez Carmona, Vicente and Jiang, Shanshan and Dong, Bin", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.10", doi = "10.18653/v1/2024.clinicalnlp-1.10", pages = "105--110", abstract = "Biomedical NLP models play a big role in the automatic extraction of information from biomedical documents, such as COVID research papers. Three landmark models have led the way in this area: BioBERT, MSR BiomedBERT, and BioLinkBERT. However, their shallow evaluation {--}a single mean score{--} forbid us to better understand how the contributions proposed in each model advance the Biomedical NLP field. We show through a Multilevel Analysis how we can assess these contributions. Our analyses across 5000 fine-tuned models show that, actually, BiomedBERT{'}s true effect is bigger than BioLinkBERT{'}s effect, and the success of BioLinkBERT does not seem to be due to its contribution {--}the Link function{--} but due to an unknown factor.", }

Biomedical NLP models play a big role in the automatic extraction of information from biomedical documents, such as COVID research papers. Three landmark models have led the way in this area: BioBERT, MSR BiomedBERT, and BioLinkBERT. However, their shallow evaluation {--}a single mean score{--} forbid us to better understand how the contributions proposed in each model advance the Biomedical NLP field. We show through a Multilevel Analysis how we can assess these contributions. Our analyses across 5000 fine-tuned models show that, actually, BiomedBERT{'}s true effect is bigger than BioLinkBERT{'}s effect, and the success of BioLinkBERT does not seem to be due to its contribution {--}the Link function{--} but due to an unknown factor.

A Multilevel Analysis of PubMed-only BERT-based Biomedical Models

clinicalnlp-1.10

Poster

https://aclanthology.org/2024.clinicalnlp-1.11.bib

https://aclanthology.org/2024.clinicalnlp-1.11/

@inproceedings{aracena-etal-2024-privacy, title = "A Privacy-Preserving Corpus for Occupational Health in {S}panish: Evaluation for {NER} and Classification Tasks", author = "Aracena, Claudio and Miranda, Luis and Vakili, Thomas and Villena, Fabi{\'a}n and Quiroga, Tamara and N{\'u}{\~n}ez-Torres, Fredy and Rocco, Victor and Dunstan, Jocelyn", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.11", doi = "10.18653/v1/2024.clinicalnlp-1.11", pages = "111--121", abstract = "Annotated corpora are essential to reliable natural language processing. While they are expensive to create, they are essential for building and evaluating systems. This study introduces a new corpus of 2,869 medical and admission reports collected by an occupational insurance and health provider. The corpus has been carefully annotated for personally identifiable information (PII) and is shared, masking this information. Two annotators adhered to annotation guidelines during the annotation process, and a referee later resolved annotation conflicts in a consolidation process to build a gold standard subcorpus. The inter-annotator agreement values, measured in F1, range between 0.86 and 0.93 depending on the selected subcorpus. The value of the corpus is demonstrated by evaluating its use for NER of PII and a classification task. The evaluations find that fine-tuned models and GPT-3.5 reach F1 of 0.911 and 0.720 in NER of PII, respectively. In the case of the insurance coverage classification task, using the original or de-identified corpus results in similar performance. The annotated data are released in de-identified form.", }

Annotated corpora are essential to reliable natural language processing. While they are expensive to create, they are essential for building and evaluating systems. This study introduces a new corpus of 2,869 medical and admission reports collected by an occupational insurance and health provider. The corpus has been carefully annotated for personally identifiable information (PII) and is shared, masking this information. Two annotators adhered to annotation guidelines during the annotation process, and a referee later resolved annotation conflicts in a consolidation process to build a gold standard subcorpus. The inter-annotator agreement values, measured in F1, range between 0.86 and 0.93 depending on the selected subcorpus. The value of the corpus is demonstrated by evaluating its use for NER of PII and a classification task. The evaluations find that fine-tuned models and GPT-3.5 reach F1 of 0.911 and 0.720 in NER of PII, respectively. In the case of the insurance coverage classification task, using the original or de-identified corpus results in similar performance. The annotated data are released in de-identified form.

A Privacy-Preserving Corpus for Occupational Health in Spanish: Evaluation for NER and Classification Tasks

clinicalnlp-1.11

Poster

https://aclanthology.org/2024.clinicalnlp-1.12.bib

https://aclanthology.org/2024.clinicalnlp-1.12/

@inproceedings{nair-etal-2024-dera, title = "{DERA}: Enhancing Large Language Model Completions with Dialog-Enabled Resolving Agents", author = "Nair, Varun and Schumacher, Elliot and Tso, Geoffrey and Kannan, Anitha", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.12", doi = "10.18653/v1/2024.clinicalnlp-1.12", pages = "122--161", abstract = "Large language models (LLMs) have emerged as valuable tools for many natural language understanding tasks. In safety-critical applications such as healthcare, the utility of these models is governed by their ability to generate factually accurate and complete outputs. In this work, we present dialog-enabled resolving agents (DERA). DERA is a paradigm made possible by the increased conversational abilities of LLMs. It provides a simple, interpretable forum for models to communicate feedback and iteratively improve output. We frame our dialog as a discussion between two agent types {--} a Researcher, who processes information and identifies crucial problem components, and a Decider, who has the autonomy to integrate the Researcher{'}s information and makes judgments on the final output.We test DERA against three clinically-focused tasks, with GPT-4 serving as our LLM. DERA shows significant improvement over the base GPT-4 performance in both human expert preference evaluations and quantitative metrics for medical conversation summarization and care plan generation. In a new finding, we also show that GPT-4{'}s performance (70{\%}) on an open-ended version of the MedQA question-answering (QA) dataset (Jin 2021; USMLE) is well above the passing level (60{\%}), with DERA showing similar performance. We will release the open-ended MedQA dataset.", }

Large language models (LLMs) have emerged as valuable tools for many natural language understanding tasks. In safety-critical applications such as healthcare, the utility of these models is governed by their ability to generate factually accurate and complete outputs. In this work, we present dialog-enabled resolving agents (DERA). DERA is a paradigm made possible by the increased conversational abilities of LLMs. It provides a simple, interpretable forum for models to communicate feedback and iteratively improve output. We frame our dialog as a discussion between two agent types {--} a Researcher, who processes information and identifies crucial problem components, and a Decider, who has the autonomy to integrate the Researcher{'}s information and makes judgments on the final output.We test DERA against three clinically-focused tasks, with GPT-4 serving as our LLM. DERA shows significant improvement over the base GPT-4 performance in both human expert preference evaluations and quantitative metrics for medical conversation summarization and care plan generation. In a new finding, we also show that GPT-4{'}s performance (70{\%}) on an open-ended version of the MedQA question-answering (QA) dataset (Jin 2021; USMLE) is well above the passing level (60{\%}), with DERA showing similar performance. We will release the open-ended MedQA dataset.

DERA: Enhancing Large Language Model Completions with Dialog-Enabled Resolving Agents

clinicalnlp-1.12

Poster

2303.17071

https://huggingface.co/papers/2303.17071

https://aclanthology.org/2024.clinicalnlp-1.13.bib

https://aclanthology.org/2024.clinicalnlp-1.13/

@inproceedings{lilli-etal-2024-llamamts, title = "{L}lama{MTS}: Optimizing Metastasis Detection with Llama Instruction Tuning and {BERT}-Based Ensemble in {I}talian Clinical Reports", author = "Lilli, Livia and Patarnello, Stefano and Masciocchi, Carlotta and Masiello, Valeria and Marazzi, Fabio and Luca, Tagliaferri and Capocchiano, Nikola", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.13", doi = "10.18653/v1/2024.clinicalnlp-1.13", pages = "162--171", abstract = "Information extraction from Electronic Health Records (EHRs) is a crucial task in healthcare, and the lack of resources and language specificity pose significant challenges. This study addresses the limited availability of Italian Natural Language Processing (NLP) tools for clinical applications and the computational demand of large language models (LLMs) for training. We present LlamaMTS, an instruction-tuned Llama for the Italian language, leveraging the LoRA technique. It is ensembled with a BERT-based model to classify EHRs based on the presence or absence of metastasis in patients affected by Breast cancer. Through our evaluation analysis, we discovered that LlamaMTS exhibits superior performance compared to both zero-shot LLMs and other Italian BERT-based models specifically fine-tuned on the same metastatic task. LlamaMTS demonstrates promising results in resource-constrained environments, offering a practical solution for information extraction from Italian EHRs in oncology, potentially improving patient care and outcomes.", }

Information extraction from Electronic Health Records (EHRs) is a crucial task in healthcare, and the lack of resources and language specificity pose significant challenges. This study addresses the limited availability of Italian Natural Language Processing (NLP) tools for clinical applications and the computational demand of large language models (LLMs) for training. We present LlamaMTS, an instruction-tuned Llama for the Italian language, leveraging the LoRA technique. It is ensembled with a BERT-based model to classify EHRs based on the presence or absence of metastasis in patients affected by Breast cancer. Through our evaluation analysis, we discovered that LlamaMTS exhibits superior performance compared to both zero-shot LLMs and other Italian BERT-based models specifically fine-tuned on the same metastatic task. LlamaMTS demonstrates promising results in resource-constrained environments, offering a practical solution for information extraction from Italian EHRs in oncology, potentially improving patient care and outcomes.

LlamaMTS: Optimizing Metastasis Detection with Llama Instruction Tuning and BERT-Based Ensemble in Italian Clinical Reports

clinicalnlp-1.13

Poster

https://aclanthology.org/2024.clinicalnlp-1.14.bib

https://aclanthology.org/2024.clinicalnlp-1.14/

@inproceedings{boulanger-etal-2024-using, title = "Using Structured Health Information for Controlled Generation of Clinical Cases in {F}rench", author = {Boulanger, Hugo and Hiebel, Nicolas and Ferret, Olivier and Fort, Kar{\"e}n and N{\'e}v{\'e}ol, Aur{\'e}lie}, editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.14", doi = "10.18653/v1/2024.clinicalnlp-1.14", pages = "172--184", abstract = "Text generation opens up new prospects for overcoming the lack of open corpora in fields such as healthcare, where data sharing is bound by confidentiality. In this study, we compare the performance of encoder-decoder and decoder-only language models for the controlled generation of clinical cases in French. To do so, we fine-tuned several pre-trained models on French clinical cases for each architecture and generate clinical cases conditioned by patient demographic information (gender and age) and clinical features.Our results suggest that encoder-decoder models are easier to control than decoder-only models, but more costly to train.", }

Text generation opens up new prospects for overcoming the lack of open corpora in fields such as healthcare, where data sharing is bound by confidentiality. In this study, we compare the performance of encoder-decoder and decoder-only language models for the controlled generation of clinical cases in French. To do so, we fine-tuned several pre-trained models on French clinical cases for each architecture and generate clinical cases conditioned by patient demographic information (gender and age) and clinical features.Our results suggest that encoder-decoder models are easier to control than decoder-only models, but more costly to train.

Using Structured Health Information for Controlled Generation of Clinical Cases in French

clinicalnlp-1.14

Poster

https://aclanthology.org/2024.clinicalnlp-1.15.bib

https://aclanthology.org/2024.clinicalnlp-1.15/

@inproceedings{amara-etal-2024-large, title = "Large Language Models Provide Human-Level Medical Text Snippet Labeling", author = "Amara, Ibtihel and Yu, Haiyang and Zhang, Fan and Liu, Yuchen and Li, Benny and Liu, Chang and Kartha, Rupesh and Goel, Akshay", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.15", doi = "10.18653/v1/2024.clinicalnlp-1.15", pages = "185--195", abstract = "This study evaluates the proficiency of Large Language Models (LLMs) in accurately labeling clinical document excerpts. Our focus is on the assignment of potential or confirmed diagnoses and medical procedures to snippets of medical text sourced from unstructured clinical patient records. We explore how the performance of LLMs compare against human annotators in classifying these excerpts. Employing a few-shot, chain-of-thought prompting approach with the MIMIC-III dataset, Med-PaLM 2 showcases annotation accuracy comparable to human annotators, achieving a notable precision rate of approximately 92{\%} relative to the gold standard labels established by human experts.", }

This study evaluates the proficiency of Large Language Models (LLMs) in accurately labeling clinical document excerpts. Our focus is on the assignment of potential or confirmed diagnoses and medical procedures to snippets of medical text sourced from unstructured clinical patient records. We explore how the performance of LLMs compare against human annotators in classifying these excerpts. Employing a few-shot, chain-of-thought prompting approach with the MIMIC-III dataset, Med-PaLM 2 showcases annotation accuracy comparable to human annotators, achieving a notable precision rate of approximately 92{\%} relative to the gold standard labels established by human experts.

Large Language Models Provide Human-Level Medical Text Snippet Labeling

clinicalnlp-1.15

Poster

https://aclanthology.org/2024.clinicalnlp-1.16.bib

https://aclanthology.org/2024.clinicalnlp-1.16/

@inproceedings{gunal-etal-2024-conversational, title = "Conversational Topic Recommendation in Counseling and Psychotherapy with Decision Transformer and Large Language Models", author = "Gunal, Aylin and Lin, Baihan and Bouneffouf, Djallel", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.16", doi = "10.18653/v1/2024.clinicalnlp-1.16", pages = "196--201", abstract = "Given the increasing demand for mental health assistance, artificial intelligence (AI), particularly large language models (LLMs), may be valuable for integration into automated clinical support systems. In this work, we leverage a decision transformer architecture for topic recommendation in counseling conversations between patients and mental health professionals. The architecture is utilized for offline reinforcement learning, and we extract states (dialogue turn embeddings), actions (conversation topics), and rewards (scores measuring the alignment between patient and therapist) from previous turns within a conversation to train a decision transformer model. We demonstrate an improvement over baseline reinforcement learning methods, and propose a novel system of utilizing our model{'}s output as synthetic labels for fine-tuning a large language model for the same task. Although our implementation based on LLaMA-2 7B has mixed results, future work can undoubtedly build on the design.", }

Given the increasing demand for mental health assistance, artificial intelligence (AI), particularly large language models (LLMs), may be valuable for integration into automated clinical support systems. In this work, we leverage a decision transformer architecture for topic recommendation in counseling conversations between patients and mental health professionals. The architecture is utilized for offline reinforcement learning, and we extract states (dialogue turn embeddings), actions (conversation topics), and rewards (scores measuring the alignment between patient and therapist) from previous turns within a conversation to train a decision transformer model. We demonstrate an improvement over baseline reinforcement learning methods, and propose a novel system of utilizing our model{'}s output as synthetic labels for fine-tuning a large language model for the same task. Although our implementation based on LLaMA-2 7B has mixed results, future work can undoubtedly build on the design.

Conversational Topic Recommendation in Counseling and Psychotherapy with Decision Transformer and Large Language Models

clinicalnlp-1.16

Poster

2405.05060

https://aclanthology.org/2024.clinicalnlp-1.17.bib

https://aclanthology.org/2024.clinicalnlp-1.17/

@inproceedings{mustafa-etal-2024-leveraging, title = "Leveraging {W}ikidata for Biomedical Entity Linking in a Low-Resource Setting: A Case Study for {G}erman", author = "Mustafa, Faizan E and Dima, Corina and Ochoa, Juan and Staab, Steffen", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.17", doi = "10.18653/v1/2024.clinicalnlp-1.17", pages = "202--207", abstract = "Biomedical Entity Linking (BEL) is a challenging task for low-resource languages, due to the lack of appropriate resources: datasets, knowledge bases (KBs), and pre-trained models. In this paper, we propose an approach to create a biomedical knowledge base for German BEL using UMLS information from Wikidata, that provides good coverage and can be easily extended to further languages. As a further contribution, we adapt several existing approaches for use in the German BEL setup, and report on their results. The chosen methods include a sparse model using character n-grams, a multilingual biomedical entity linker, and two general-purpose text retrieval models. Our results show that a language-specific KB that provides good coverage leads to most improvement in entity linking performance, irrespective of the used model. The finetuned German BEL model, newly created UMLS$_{Wikidata}$ KB as well as the code to reproduce our results are publicly available.", }

Biomedical Entity Linking (BEL) is a challenging task for low-resource languages, due to the lack of appropriate resources: datasets, knowledge bases (KBs), and pre-trained models. In this paper, we propose an approach to create a biomedical knowledge base for German BEL using UMLS information from Wikidata, that provides good coverage and can be easily extended to further languages. As a further contribution, we adapt several existing approaches for use in the German BEL setup, and report on their results. The chosen methods include a sparse model using character n-grams, a multilingual biomedical entity linker, and two general-purpose text retrieval models. Our results show that a language-specific KB that provides good coverage leads to most improvement in entity linking performance, irrespective of the used model. The finetuned German BEL model, newly created UMLS$_{Wikidata}$ KB as well as the code to reproduce our results are publicly available.

Leveraging Wikidata for Biomedical Entity Linking in a Low-Resource Setting: A Case Study for German

clinicalnlp-1.17

Poster

https://aclanthology.org/2024.clinicalnlp-1.18.bib

https://aclanthology.org/2024.clinicalnlp-1.18/

@inproceedings{rohr-etal-2024-revisiting, title = "Revisiting Clinical Outcome Prediction for {MIMIC}-{IV}", author = {R{\"o}hr, Tom and Figueroa, Alexei and Papaioannou, Jens-Michalis and Fallon, Conor and Bressem, Keno and Nejdl, Wolfgang and L{\"o}ser, Alexander}, editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.18", doi = "10.18653/v1/2024.clinicalnlp-1.18", pages = "208--217", abstract = "Clinical Decision Support Systems assist medical professionals in providing optimal care for patients.A prominent data source used for creating tasks for such systems is the Medical Information Mart for Intensive Care (MIMIC).MIMIC contains electronic health records (EHR) gathered in a tertiary hospital in the United States.The majority of past work is based on the third version of MIMIC, although the fourth is the most recent version.This new version, not only introduces more data into MIMIC, but also increases the variety of patients.While MIMIC-III is limited to intensive care units, MIMIC-IV also offers EHRs from the emergency department.In this work, we investigate how to adapt previous work to update clinical outcome prediction for MIMIC-IV.We revisit several established tasks, including prediction of diagnoses, procedures, length-of-stay, and also introduce a novel task: patient routing prediction.Furthermore, we quantitatively and qualitatively evaluate all tasks on several bio-medical transformer encoder models.Finally, we provide narratives for future research directions in the clinical outcome prediction domain. We make our source code publicly available to reproduce our experiments, data, and tasks.", }

Clinical Decision Support Systems assist medical professionals in providing optimal care for patients.A prominent data source used for creating tasks for such systems is the Medical Information Mart for Intensive Care (MIMIC).MIMIC contains electronic health records (EHR) gathered in a tertiary hospital in the United States.The majority of past work is based on the third version of MIMIC, although the fourth is the most recent version.This new version, not only introduces more data into MIMIC, but also increases the variety of patients.While MIMIC-III is limited to intensive care units, MIMIC-IV also offers EHRs from the emergency department.In this work, we investigate how to adapt previous work to update clinical outcome prediction for MIMIC-IV.We revisit several established tasks, including prediction of diagnoses, procedures, length-of-stay, and also introduce a novel task: patient routing prediction.Furthermore, we quantitatively and qualitatively evaluate all tasks on several bio-medical transformer encoder models.Finally, we provide narratives for future research directions in the clinical outcome prediction domain. We make our source code publicly available to reproduce our experiments, data, and tasks.

Revisiting Clinical Outcome Prediction for MIMIC-IV

clinicalnlp-1.18

Poster

https://aclanthology.org/2024.clinicalnlp-1.19.bib

https://aclanthology.org/2024.clinicalnlp-1.19/

@inproceedings{sayin-etal-2024-llms, title = "Can {LLM}s Correct Physicians, Yet? Investigating Effective Interaction Methods in the Medical Domain", author = "Sayin, Burcu and Minervini, Pasquale and Staiano, Jacopo and Passerini, Andrea", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.19", doi = "10.18653/v1/2024.clinicalnlp-1.19", pages = "218--237", abstract = "We explore the potential of Large Language Models (LLMs) to assist and potentially correct physicians in medical decision-making tasks. We evaluate several LLMs, including Meditron, Llama2, and Mistral, to analyze the ability of these models to interact effectively with physicians across different scenarios. We consider questions from PubMedQA and several tasks, ranging from binary (yes/no) responses to long answer generation, where the answer of the model is produced after an interaction with a physician. Our findings suggest that prompt design significantly influences the downstream accuracy of LLMs and that LLMs can provide valuable feedback to physicians, challenging incorrect diagnoses and contributing to more accurate decision-making. For example, when the physician is accurate 38{\%} of the time, Mistral can produce the correct answer, improving accuracy up to 74{\%} depending on the prompt being used, while Llama2 and Meditron models exhibit greater sensitivity to prompt choice. Our analysis also uncovers the challenges of ensuring that LLM-generated suggestions are pertinent and useful, emphasizing the need for further research in this area.", }

We explore the potential of Large Language Models (LLMs) to assist and potentially correct physicians in medical decision-making tasks. We evaluate several LLMs, including Meditron, Llama2, and Mistral, to analyze the ability of these models to interact effectively with physicians across different scenarios. We consider questions from PubMedQA and several tasks, ranging from binary (yes/no) responses to long answer generation, where the answer of the model is produced after an interaction with a physician. Our findings suggest that prompt design significantly influences the downstream accuracy of LLMs and that LLMs can provide valuable feedback to physicians, challenging incorrect diagnoses and contributing to more accurate decision-making. For example, when the physician is accurate 38{\%} of the time, Mistral can produce the correct answer, improving accuracy up to 74{\%} depending on the prompt being used, while Llama2 and Meditron models exhibit greater sensitivity to prompt choice. Our analysis also uncovers the challenges of ensuring that LLM-generated suggestions are pertinent and useful, emphasizing the need for further research in this area.

Can LLMs Correct Physicians, Yet? Investigating Effective Interaction Methods in the Medical Domain

clinicalnlp-1.19

Poster

2403.20288

https://aclanthology.org/2024.clinicalnlp-1.20.bib

https://aclanthology.org/2024.clinicalnlp-1.20/

@inproceedings{cong-etal-2024-leveraging, title = "Leveraging pre-trained large language models for aphasia detection in {E}nglish and {C}hinese speakers", author = "Cong, Yan and Lee, Jiyeon and LaCroix, Arianna", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.20", doi = "10.18653/v1/2024.clinicalnlp-1.20", pages = "238--245", abstract = "We explore the utility of pre-trained Large Language Models (LLMs) in detecting the presence, subtypes, and severity of aphasia across English and Mandarin Chinese speakers. Our investigation suggests that even without fine-tuning or domain-specific training, pre-trained LLMs can offer some insights on language disorders, regardless of speakers{'} first language. Our analysis also reveals noticeable differences between English and Chinese LLMs. While the English LLMs exhibit near-chance level accuracy in subtyping aphasia, the Chinese counterparts demonstrate less than satisfactory performance in distinguishing between individuals with and without aphasia. This research advocates for the importance of linguistically tailored and specified approaches in leveraging LLMs for clinical applications, especially in the context of multilingual populations.", }

We explore the utility of pre-trained Large Language Models (LLMs) in detecting the presence, subtypes, and severity of aphasia across English and Mandarin Chinese speakers. Our investigation suggests that even without fine-tuning or domain-specific training, pre-trained LLMs can offer some insights on language disorders, regardless of speakers{'} first language. Our analysis also reveals noticeable differences between English and Chinese LLMs. While the English LLMs exhibit near-chance level accuracy in subtyping aphasia, the Chinese counterparts demonstrate less than satisfactory performance in distinguishing between individuals with and without aphasia. This research advocates for the importance of linguistically tailored and specified approaches in leveraging LLMs for clinical applications, especially in the context of multilingual populations.

Leveraging pre-trained large language models for aphasia detection in English and Chinese speakers

clinicalnlp-1.20

Poster

https://aclanthology.org/2024.clinicalnlp-1.21.bib

https://aclanthology.org/2024.clinicalnlp-1.21/

@inproceedings{ha-etal-2024-fusion, title = "Fusion of Domain-Adapted Vision and Language Models for Medical Visual Question Answering", author = "Ha, Cuong and Asaadi, Shima and Karn, Sanjeev Kumar and Farri, Oladimeji and Heimann, Tobias and Runkler, Thomas", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.21", doi = "10.18653/v1/2024.clinicalnlp-1.21", pages = "246--257", abstract = "Vision-language models, while effective in general domains and showing strong performance in diverse multi-modal applications like visual question-answering (VQA), struggle to maintain the same level of effectiveness in more specialized domains, e.g., medical. We propose a medical vision-language model that integrates large vision and language models adapted for the medical domain. This model goes through three stages of parameter-efficient training using three separate biomedical and radiology multi-modal visual and text datasets. The proposed model achieves state-of-the-art performance on the SLAKE 1.0 medical VQA (MedVQA) dataset with an overall accuracy of 87.5{\%} and demonstrates strong performance on another MedVQA dataset, VQA-RAD, achieving an overall accuracy of 73.2{\%}.", }

Vision-language models, while effective in general domains and showing strong performance in diverse multi-modal applications like visual question-answering (VQA), struggle to maintain the same level of effectiveness in more specialized domains, e.g., medical. We propose a medical vision-language model that integrates large vision and language models adapted for the medical domain. This model goes through three stages of parameter-efficient training using three separate biomedical and radiology multi-modal visual and text datasets. The proposed model achieves state-of-the-art performance on the SLAKE 1.0 medical VQA (MedVQA) dataset with an overall accuracy of 87.5{\%} and demonstrates strong performance on another MedVQA dataset, VQA-RAD, achieving an overall accuracy of 73.2{\%}.

Fusion of Domain-Adapted Vision and Language Models for Medical Visual Question Answering

clinicalnlp-1.21

Poster

2404.16192

https://aclanthology.org/2024.clinicalnlp-1.22.bib

https://aclanthology.org/2024.clinicalnlp-1.22/

@inproceedings{krishnamoorthy-etal-2024-llm, title = "{LLM}-Based Section Identifiers Excel on Open Source but Stumble in Real World Applications", author = "Krishnamoorthy, Saranya and Singh, Ayush and Tafreshi, Shabnam", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.22", doi = "10.18653/v1/2024.clinicalnlp-1.22", pages = "258--270", abstract = "Electronic health records (EHR) even though a boon for healthcare practitioners, are grow- ing convoluted and longer every day. Sifting around these lengthy EHRs is taxing and be- comes a cumbersome part of physician-patient interaction. Several approaches have been pro- posed to help alleviate this prevalent issue ei- ther via summarization or sectioning, however, only a few approaches have truly been helpful in the past. With the rise of automated methods, machine learning (ML) has shown promise in solving the task of identifying relevant sections in EHR. However, most ML methods rely on labeled data which is difficult to get in health- care. Large language models (LLMs) on the other hand, have performed impressive feats in natural language processing (NLP), that too in a zero-shot manner, i.e. without any labeled data. To that end, we propose using LLMs to identify relevant section headers. We find that GPT-4 can effectively solve the task on both zero and few-shot settings as well as segment dramatically better than state-of-the-art meth- ods. Additionally, we also annotate a much harder real world dataset and find that GPT-4 struggles to perform well, alluding to further research and harder benchmarks.", }

Electronic health records (EHR) even though a boon for healthcare practitioners, are grow- ing convoluted and longer every day. Sifting around these lengthy EHRs is taxing and be- comes a cumbersome part of physician-patient interaction. Several approaches have been pro- posed to help alleviate this prevalent issue ei- ther via summarization or sectioning, however, only a few approaches have truly been helpful in the past. With the rise of automated methods, machine learning (ML) has shown promise in solving the task of identifying relevant sections in EHR. However, most ML methods rely on labeled data which is difficult to get in health- care. Large language models (LLMs) on the other hand, have performed impressive feats in natural language processing (NLP), that too in a zero-shot manner, i.e. without any labeled data. To that end, we propose using LLMs to identify relevant section headers. We find that GPT-4 can effectively solve the task on both zero and few-shot settings as well as segment dramatically better than state-of-the-art meth- ods. Additionally, we also annotate a much harder real world dataset and find that GPT-4 struggles to perform well, alluding to further research and harder benchmarks.

LLM-Based Section Identifiers Excel on Open Source but Stumble in Real World Applications

clinicalnlp-1.22

Poster

2404.16294

https://aclanthology.org/2024.clinicalnlp-1.23.bib

https://aclanthology.org/2024.clinicalnlp-1.23/

@inproceedings{cai-etal-2024-adapting, title = "Adapting {A}bstract {M}eaning {R}epresentation Parsing to the Clinical Narrative {--} the {SPRING} {THYME} parser", author = "Cai, Jon and Wright-Bettner, Kristin and Palmer, Martha and Savova, Guergana and Martin, James", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.23", doi = "10.18653/v1/2024.clinicalnlp-1.23", pages = "271--282", abstract = "This paper is dedicated to the design and evaluation of the first AMR parser tailored for clinical notes. Our objective was to facilitate the precise transformation of the clinical notes into structured AMR expressions, thereby enhancing the interpretability and usability of clinical text data at scale. Leveraging the colon cancer dataset from the Temporal Histories of Your Medical Events (THYME) corpus, we adapted a state-of-the-art AMR parser utilizing continuous training. Our approach incorporates data augmentation techniques to enhance the accuracy of AMR structure predictions. Notably, through this learning strategy, our parser achieved an impressive F1 score of 88{\%} on the THYME corpus{'}s colon cancer dataset. Moreover, our research delved into the efficacy of data required for domain adaptation within the realm of clinical notes, presenting domain adaptation data requirements for AMR parsing. This exploration not only underscores the parser{'}s robust performance but also highlights its potential in facilitating a deeper understanding of clinical narratives through structured semantic representations.", }

This paper is dedicated to the design and evaluation of the first AMR parser tailored for clinical notes. Our objective was to facilitate the precise transformation of the clinical notes into structured AMR expressions, thereby enhancing the interpretability and usability of clinical text data at scale. Leveraging the colon cancer dataset from the Temporal Histories of Your Medical Events (THYME) corpus, we adapted a state-of-the-art AMR parser utilizing continuous training. Our approach incorporates data augmentation techniques to enhance the accuracy of AMR structure predictions. Notably, through this learning strategy, our parser achieved an impressive F1 score of 88{\%} on the THYME corpus{'}s colon cancer dataset. Moreover, our research delved into the efficacy of data required for domain adaptation within the realm of clinical notes, presenting domain adaptation data requirements for AMR parsing. This exploration not only underscores the parser{'}s robust performance but also highlights its potential in facilitating a deeper understanding of clinical narratives through structured semantic representations.

Adapting Abstract Meaning Representation Parsing to the Clinical Narrative – the SPRING THYME parser

clinicalnlp-1.23

Poster

https://aclanthology.org/2024.clinicalnlp-1.24.bib

https://aclanthology.org/2024.clinicalnlp-1.24/

@inproceedings{kapadnis-etal-2024-serpent, title = "{SERPENT}-{VLM} : Self-Refining Radiology Report Generation Using Vision Language Models", author = "Kapadnis, Manav and Patnaik, Sohan and Nandy, Abhilash and Ray, Sourjyadip and Goyal, Pawan and Sheet, Debdoot", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.24", doi = "10.18653/v1/2024.clinicalnlp-1.24", pages = "283--291", abstract = "Radiology Report Generation (R2Gen) demonstrates how Multi-modal Large Language Models (MLLMs) can automate the creation of accurate and coherent radiological reports. Existing methods often hallucinate details in text-based reports that don{'}t accurately reflect the image content. To mitigate this, we introduce a novel strategy, SERPENT-VLM (SElf Refining Radiology RePort GENeraTion using Vision Language Models), which improves the R2Gen task by integrating a self-refining mechanism into the MLLM framework. We employ a unique self-supervised loss that leverages similarity between pooled image representations and the contextual representations of the generated radiological text, alongside the standard Causal Language Modeling objective, to refine image-text representations. This allows the model to scrutinize and align the generated text through dynamic interaction between a given image and the generated text, therefore reducing hallucination and continuously enhancing nuanced report generation. SERPENT-VLM outperforms existing baselines such as LlaVA-Med, BiomedGPT, etc., achieving SoTA performance on the IU X-ray and Radiology Objects in COntext (ROCO) datasets, and also proves to be robust against noisy images. A qualitative case study emphasizes the significant advancements towards more sophisticated MLLM frameworks for R2Gen, opening paths for further research into self-supervised refinement in the medical imaging domain.", }

Radiology Report Generation (R2Gen) demonstrates how Multi-modal Large Language Models (MLLMs) can automate the creation of accurate and coherent radiological reports. Existing methods often hallucinate details in text-based reports that don{'}t accurately reflect the image content. To mitigate this, we introduce a novel strategy, SERPENT-VLM (SElf Refining Radiology RePort GENeraTion using Vision Language Models), which improves the R2Gen task by integrating a self-refining mechanism into the MLLM framework. We employ a unique self-supervised loss that leverages similarity between pooled image representations and the contextual representations of the generated radiological text, alongside the standard Causal Language Modeling objective, to refine image-text representations. This allows the model to scrutinize and align the generated text through dynamic interaction between a given image and the generated text, therefore reducing hallucination and continuously enhancing nuanced report generation. SERPENT-VLM outperforms existing baselines such as LlaVA-Med, BiomedGPT, etc., achieving SoTA performance on the IU X-ray and Radiology Objects in COntext (ROCO) datasets, and also proves to be robust against noisy images. A qualitative case study emphasizes the significant advancements towards more sophisticated MLLM frameworks for R2Gen, opening paths for further research into self-supervised refinement in the medical imaging domain.

SERPENT-VLM : Self-Refining Radiology Report Generation Using Vision Language Models

clinicalnlp-1.24

Poster

2404.17912

https://aclanthology.org/2024.clinicalnlp-1.25.bib

https://aclanthology.org/2024.clinicalnlp-1.25/

@inproceedings{lim-etal-2024-erd, title = "{ERD}: A Framework for Improving {LLM} Reasoning for Cognitive Distortion Classification", author = "Lim, Sehee and Kim, Yejin and Choi, Chi-Hyun and Sohn, Jy-yong and Kim, Byung-Hoon", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.25", doi = "10.18653/v1/2024.clinicalnlp-1.25", pages = "292--300", abstract = "Improving the accessibility of psychotherapy with the aid of Large Language Models (LLMs) is garnering a significant attention in recent years. Recognizing cognitive distortions from the interviewee{'}s utterances can be an essential part of psychotherapy, especially for cognitive behavioral therapy. In this paper, we propose ERD, which improves LLM-based cognitive distortion classification performance with the aid of additional modules of (1) extracting the parts related to cognitive distortion, and (2) debating the reasoning steps by multiple agents. Our experimental results on a public dataset show that ERD improves the multi-class F1 score as well as binary specificity score. Regarding the latter score, it turns out that our method is effective in debiasing the baseline method which has high false positive rate, especially when the summary of multi-agent debate is provided to LLMs.", }

Improving the accessibility of psychotherapy with the aid of Large Language Models (LLMs) is garnering a significant attention in recent years. Recognizing cognitive distortions from the interviewee{'}s utterances can be an essential part of psychotherapy, especially for cognitive behavioral therapy. In this paper, we propose ERD, which improves LLM-based cognitive distortion classification performance with the aid of additional modules of (1) extracting the parts related to cognitive distortion, and (2) debating the reasoning steps by multiple agents. Our experimental results on a public dataset show that ERD improves the multi-class F1 score as well as binary specificity score. Regarding the latter score, it turns out that our method is effective in debiasing the baseline method which has high false positive rate, especially when the summary of multi-agent debate is provided to LLMs.

ERD: A Framework for Improving LLM Reasoning for Cognitive Distortion Classification

clinicalnlp-1.25

Poster

2403.14255

https://aclanthology.org/2024.clinicalnlp-1.26.bib

https://aclanthology.org/2024.clinicalnlp-1.26/

@inproceedings{hazan-etal-2024-leveraging, title = "Leveraging Prompt-Learning for Structured Information Extraction from Crohn{'}s Disease Radiology Reports in a Low-Resource Language", author = "Hazan, Liam and Gavrielov, Naama and Reichart, Roi and Hagopian, Talar and Greer, Mary-Louise and Cytter-Kuint, Ruth and Focht, Gili and Turner, Dan and Freiman, Moti", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.26", doi = "10.18653/v1/2024.clinicalnlp-1.26", pages = "301--309", abstract = "Automatic conversion of free-text radiology reports into structured data using Natural Language Processing (NLP) techniques is crucial for analyzing diseases on a large scale. While effective for tasks in widely spoken languages like English, generative large language models (LLMs) typically underperform with less common languages and can pose potential risks to patient privacy. Fine-tuning local NLP models is hindered by the skewed nature of real-world medical datasets, where rare findings represent a significant data imbalance. We introduce SMP-BERT, a novel prompt learning method that leverages the structured nature of reports to overcome these challenges. In our studies involving a substantial collection of Crohn{'}s disease radiology reports in Hebrew (over 8,000 patients and 10,000 reports), SMP-BERT greatly surpassed traditional fine-tuning methods in performance, notably in detecting infrequent conditions (AUC: 0.99 vs 0.94, F1: 0.84 vs 0.34). SMP-BERT empowers more accurate AI diagnostics available for low-resource languages.", }

Automatic conversion of free-text radiology reports into structured data using Natural Language Processing (NLP) techniques is crucial for analyzing diseases on a large scale. While effective for tasks in widely spoken languages like English, generative large language models (LLMs) typically underperform with less common languages and can pose potential risks to patient privacy. Fine-tuning local NLP models is hindered by the skewed nature of real-world medical datasets, where rare findings represent a significant data imbalance. We introduce SMP-BERT, a novel prompt learning method that leverages the structured nature of reports to overcome these challenges. In our studies involving a substantial collection of Crohn{'}s disease radiology reports in Hebrew (over 8,000 patients and 10,000 reports), SMP-BERT greatly surpassed traditional fine-tuning methods in performance, notably in detecting infrequent conditions (AUC: 0.99 vs 0.94, F1: 0.84 vs 0.34). SMP-BERT empowers more accurate AI diagnostics available for low-resource languages.

Leveraging Prompt-Learning for Structured Information Extraction from Crohn's Disease Radiology Reports in a Low-Resource Language

clinicalnlp-1.26

Poster

2405.01682

https://aclanthology.org/2024.clinicalnlp-1.27.bib

https://aclanthology.org/2024.clinicalnlp-1.27/

@inproceedings{liu-etal-2024-context, title = "Context Aggregation with Topic-focused Summarization for Personalized Medical Dialogue Generation", author = "Liu, Zhengyuan and Salleh, Siti and Krishnaswamy, Pavitra and Chen, Nancy", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.27", doi = "10.18653/v1/2024.clinicalnlp-1.27", pages = "310--321", abstract = "In the realm of dialogue systems, generated responses often lack personalization. This is particularly true in the medical domain, where research is limited by scarce available domain-specific data and the complexities of modeling medical context and persona information. In this work, we investigate the potential of harnessing large language models for personalized medical dialogue generation. In particular, to better aggregate the long conversational context, we adopt topic-focused summarization to distill core information from the dialogue history, and use such information to guide the conversation flow and generated content. Drawing inspiration from real-world telehealth conversations, we outline a comprehensive pipeline encompassing data processing, profile construction, and domain adaptation. This work not only highlights our technical approach but also shares distilled insights from the data preparation and model construction phases.", }

In the realm of dialogue systems, generated responses often lack personalization. This is particularly true in the medical domain, where research is limited by scarce available domain-specific data and the complexities of modeling medical context and persona information. In this work, we investigate the potential of harnessing large language models for personalized medical dialogue generation. In particular, to better aggregate the long conversational context, we adopt topic-focused summarization to distill core information from the dialogue history, and use such information to guide the conversation flow and generated content. Drawing inspiration from real-world telehealth conversations, we outline a comprehensive pipeline encompassing data processing, profile construction, and domain adaptation. This work not only highlights our technical approach but also shares distilled insights from the data preparation and model construction phases.

Context Aggregation with Topic-focused Summarization for Personalized Medical Dialogue Generation

clinicalnlp-1.27

Poster

https://aclanthology.org/2024.clinicalnlp-1.28.bib

https://aclanthology.org/2024.clinicalnlp-1.28/

@inproceedings{milintsevich-etal-2024-evaluating, title = "Evaluating Lexicon Incorporation for Depression Symptom Estimation", author = {Milintsevich, Kirill and Dias, Ga{\"e}l and Sirts, Kairit}, editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.28", doi = "10.18653/v1/2024.clinicalnlp-1.28", pages = "322--328", abstract = "This paper explores the impact of incorporating sentiment, emotion, and domain-specific lexicons into a transformer-based model for depression symptom estimation. Lexicon information is added by marking the words in the input transcripts of patient-therapist conversations as well as in social media posts. Overall results show that the introduction of external knowledge within pre-trained language models can be beneficial for prediction performance, while different lexicons show distinct behaviours depending on the targeted task. Additionally, new state-of-the-art results are obtained for the estimation of depression level over patient-therapist interviews.", }

This paper explores the impact of incorporating sentiment, emotion, and domain-specific lexicons into a transformer-based model for depression symptom estimation. Lexicon information is added by marking the words in the input transcripts of patient-therapist conversations as well as in social media posts. Overall results show that the introduction of external knowledge within pre-trained language models can be beneficial for prediction performance, while different lexicons show distinct behaviours depending on the targeted task. Additionally, new state-of-the-art results are obtained for the estimation of depression level over patient-therapist interviews.

Evaluating Lexicon Incorporation for Depression Symptom Estimation

clinicalnlp-1.28

Poster

2404.19359

https://aclanthology.org/2024.clinicalnlp-1.29.bib

https://aclanthology.org/2024.clinicalnlp-1.29/

@inproceedings{sugihara-etal-2024-semi, title = "Semi-automatic Construction of a Word Complexity Lexicon for {J}apanese Medical Terminology", author = "Sugihara, Soichiro and Kajiwara, Tomoyuki and Ninomiya, Takashi and Wakamiya, Shoko and Aramaki, Eiji", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.29", doi = "10.18653/v1/2024.clinicalnlp-1.29", pages = "329--333", abstract = "We construct a word complexity lexicon for medical terms in Japanese.To facilitate communication between medical practitioners and patients, medical text simplification is being studied.Medical text simplification is a natural language processing task that paraphrases complex technical terms into expressions that patients can understand.However, in contrast to English, where this task is being actively studied, there are insufficient language resources in Japanese.As a first step in advancing research on medical text simplification in Japanese, we annotate the 370,000 words from a large-scale medical terminology lexicon with a five-point scale of complexity for patients.", }

We construct a word complexity lexicon for medical terms in Japanese.To facilitate communication between medical practitioners and patients, medical text simplification is being studied.Medical text simplification is a natural language processing task that paraphrases complex technical terms into expressions that patients can understand.However, in contrast to English, where this task is being actively studied, there are insufficient language resources in Japanese.As a first step in advancing research on medical text simplification in Japanese, we annotate the 370,000 words from a large-scale medical terminology lexicon with a five-point scale of complexity for patients.

Semi-automatic Construction of a Word Complexity Lexicon for Japanese Medical Terminology

clinicalnlp-1.29

Poster

https://aclanthology.org/2024.clinicalnlp-1.30.bib

https://aclanthology.org/2024.clinicalnlp-1.30/

@inproceedings{kim-etal-2024-team, title = "{TEAM} {MIPAL} at {MEDIQA}-{M}3{G} 2024: Large {VQA} Models for Dermatological Diagnosis", author = "Kim, Hyeonjin and Kim, Min and Jang, Jae and Yoo, KiYoon and Kwak, Nojun", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.30", doi = "10.18653/v1/2024.clinicalnlp-1.30", pages = "334--338", abstract = "This paper describes the methods used for the NAACL 2024 workshop MEDIQA-M3G shared task for generating medical answers from image and query data for skin diseases. MedVInT-Decoder, LLaVA, and LLaVA-Med are chosen as base models. Finetuned with the task dataset on the dermatological domain, MedVInT-Decoder achieved a BLEU score of 3.82 during competition, while LLaVA and LLaVA-Med reached 6.98 and 4.62 afterward, respectively.", }

This paper describes the methods used for the NAACL 2024 workshop MEDIQA-M3G shared task for generating medical answers from image and query data for skin diseases. MedVInT-Decoder, LLaVA, and LLaVA-Med are chosen as base models. Finetuned with the task dataset on the dermatological domain, MedVInT-Decoder achieved a BLEU score of 3.82 during competition, while LLaVA and LLaVA-Med reached 6.98 and 4.62 afterward, respectively.

TEAM MIPAL at MEDIQA-M3G 2024: Large VQA Models for Dermatological Diagnosis

clinicalnlp-1.30

Poster

https://aclanthology.org/2024.clinicalnlp-1.31.bib

https://aclanthology.org/2024.clinicalnlp-1.31/

@inproceedings{saeed-2024-medifact, title = "{M}edi{F}act at {MEDIQA}-{M}3{G} 2024: Medical Question Answering in Dermatology with Multimodal Learning", author = "Saeed, Nadia", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.31", doi = "10.18653/v1/2024.clinicalnlp-1.31", pages = "339--345", abstract = "The MEDIQA-M3G 2024 challenge necessitates novel solutions for Multilingual {\&} Multimodal Medical Answer Generation in dermatology (wai Yim et al., 2024a). This paper addresses the limitations of traditional methods by proposing a weakly supervised learning approach for open-ended medical question-answering (QA). Our system leverages readily available MEDIQA-M3G images via a VGG16-CNN-SVM model, enabling multilingual (English, Chinese, Spanish) learning of informative skin condition representations. Using pre-trained QA models, we further bridge the gap between visual and textual information through multimodal fusion. This approach tackles complex, open-ended questions even without predefined answer choices. We empower the generation of comprehensive answers by feeding the ViT-CLIP model with multiple responses alongside images. This work advances medical QA research, paving the way for clinical decision support systems and ultimately improving healthcare delivery.", }

The MEDIQA-M3G 2024 challenge necessitates novel solutions for Multilingual {\&} Multimodal Medical Answer Generation in dermatology (wai Yim et al., 2024a). This paper addresses the limitations of traditional methods by proposing a weakly supervised learning approach for open-ended medical question-answering (QA). Our system leverages readily available MEDIQA-M3G images via a VGG16-CNN-SVM model, enabling multilingual (English, Chinese, Spanish) learning of informative skin condition representations. Using pre-trained QA models, we further bridge the gap between visual and textual information through multimodal fusion. This approach tackles complex, open-ended questions even without predefined answer choices. We empower the generation of comprehensive answers by feeding the ViT-CLIP model with multiple responses alongside images. This work advances medical QA research, paving the way for clinical decision support systems and ultimately improving healthcare delivery.

MediFact at MEDIQA-M3G 2024: Medical Question Answering in Dermatology with Multimodal Learning

clinicalnlp-1.31

Poster

2405.01583

https://aclanthology.org/2024.clinicalnlp-1.32.bib

https://aclanthology.org/2024.clinicalnlp-1.32/

@inproceedings{saeed-2024-medifact-mediqa, title = "{M}edi{F}act at {MEDIQA}-{CORR} 2024: Why {AI} Needs a Human Touch", author = "Saeed, Nadia", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.32", doi = "10.18653/v1/2024.clinicalnlp-1.32", pages = "346--352", abstract = "Accurate representation of medical information is crucial for patient safety, yet artificial intelligence (AI) systems, such as Large Language Models (LLMs), encounter challenges in error-free clinical text interpretation. This paper presents a novel approach submitted to the MEDIQA-CORR 2024 shared task k (Ben Abacha et al., 2024a), focusing on the automatic correction of single-word errors in clinical notes. Unlike LLMs that rely on extensive generic data, our method emphasizes extracting contextually relevant information from available clinical text data. Leveraging an ensemble of extractive and abstractive question-answering approaches, we construct a supervised learning framework with domain-specific feature engineering. Our methodology incorporates domain expertise to enhance error correction accuracy. By integrating domain expertise and prioritizing meaningful information extraction, our approach underscores the significance of a human-centric strategy in adapting AI for healthcare.", }

Accurate representation of medical information is crucial for patient safety, yet artificial intelligence (AI) systems, such as Large Language Models (LLMs), encounter challenges in error-free clinical text interpretation. This paper presents a novel approach submitted to the MEDIQA-CORR 2024 shared task k (Ben Abacha et al., 2024a), focusing on the automatic correction of single-word errors in clinical notes. Unlike LLMs that rely on extensive generic data, our method emphasizes extracting contextually relevant information from available clinical text data. Leveraging an ensemble of extractive and abstractive question-answering approaches, we construct a supervised learning framework with domain-specific feature engineering. Our methodology incorporates domain expertise to enhance error correction accuracy. By integrating domain expertise and prioritizing meaningful information extraction, our approach underscores the significance of a human-centric strategy in adapting AI for healthcare.

MediFact at MEDIQA-CORR 2024: Why AI Needs a Human Touch

clinicalnlp-1.32

Poster

2404.17999

https://aclanthology.org/2024.clinicalnlp-1.33.bib

https://aclanthology.org/2024.clinicalnlp-1.33/

@inproceedings{wu-etal-2024-knowlab, title = "{K}now{L}ab{\_}{AIM}ed at {MEDIQA}-{CORR} 2024: Chain-of-Though ({C}o{T}) prompting strategies for medical error detection and correction", author = "Wu, Zhaolong and Hasan, Abul and Wu, Jinge and Kim, Yunsoo and Cheung, Jason and Zhang, Teng and Wu, Honghan", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.33", doi = "10.18653/v1/2024.clinicalnlp-1.33", pages = "353--359", abstract = "This paper describes our submission to the MEDIQA-CORR 2024 shared task for automatically detecting and correcting medical errors in clinical notes. We report results for three methods of few-shot In-Context Learning (ICL) augmented with Chain-of-Thought (CoT) and reason prompts using a large language model (LLM). In the first method, we manually analyse a subset of train and validation dataset to infer three CoT prompts by examining error types in the clinical notes. In the second method, we utilise the training dataset to prompt the LLM to deduce reasons about their correctness or incorrectness. The constructed CoTs and reasons are then augmented with ICL examples to solve the tasks of error detection, span identification, and error correction. Finally, we combine the two methods using a rule-based ensemble method. Across the three sub-tasks, our ensemble method achieves a ranking of 3rd for both sub-task 1 and 2, while securing 7th place in sub-task 3 among all submissions.", }

This paper describes our submission to the MEDIQA-CORR 2024 shared task for automatically detecting and correcting medical errors in clinical notes. We report results for three methods of few-shot In-Context Learning (ICL) augmented with Chain-of-Thought (CoT) and reason prompts using a large language model (LLM). In the first method, we manually analyse a subset of train and validation dataset to infer three CoT prompts by examining error types in the clinical notes. In the second method, we utilise the training dataset to prompt the LLM to deduce reasons about their correctness or incorrectness. The constructed CoTs and reasons are then augmented with ICL examples to solve the tasks of error detection, span identification, and error correction. Finally, we combine the two methods using a rule-based ensemble method. Across the three sub-tasks, our ensemble method achieves a ranking of 3rd for both sub-task 1 and 2, while securing 7th place in sub-task 3 among all submissions.

KnowLab_AIMed at MEDIQA-CORR 2024: Chain-of-Though (CoT) prompting strategies for medical error detection and correction

clinicalnlp-1.33

Poster

https://aclanthology.org/2024.clinicalnlp-1.34.bib

https://aclanthology.org/2024.clinicalnlp-1.34/

@inproceedings{gundabathula-kolar-2024-promptmind, title = "{P}rompt{M}ind Team at {EHRSQL}-2024: Improving Reliability of {SQL} Generation using Ensemble {LLM}s", author = "Gundabathula, Satya and Kolar, Sriram", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.34", doi = "10.18653/v1/2024.clinicalnlp-1.34", pages = "360--366", abstract = "This paper presents our approach to the EHRSQL-2024 shared task, which aims to develop a reliable Text-to-SQL system for electronic health records. We propose two approaches that leverage large language models (LLMs) for prompting and fine-tuning to generate EHRSQL queries. In both techniques, we concentrate on bridging the gap between the real-world knowledge on which LLMs are trained and the domain-specific knowledge required for the task. The paper provides the results of each approach individually, demonstrating that they achieve high execution accuracy. Additionally, we show that an ensemble approach further enhances generation reliability by reducing errors. This approach secured us 2nd place in the shared task competition. The methodologies outlined in this paper are designed to be transferable to domain-specific Text-to-SQL problems that emphasize both accuracy and reliability.", }

This paper presents our approach to the EHRSQL-2024 shared task, which aims to develop a reliable Text-to-SQL system for electronic health records. We propose two approaches that leverage large language models (LLMs) for prompting and fine-tuning to generate EHRSQL queries. In both techniques, we concentrate on bridging the gap between the real-world knowledge on which LLMs are trained and the domain-specific knowledge required for the task. The paper provides the results of each approach individually, demonstrating that they achieve high execution accuracy. Additionally, we show that an ensemble approach further enhances generation reliability by reducing errors. This approach secured us 2nd place in the shared task competition. The methodologies outlined in this paper are designed to be transferable to domain-specific Text-to-SQL problems that emphasize both accuracy and reliability.

PromptMind Team at EHRSQL-2024: Improving Reliability of SQL Generation using Ensemble LLMs

clinicalnlp-1.34

Poster

2405.08839

https://aclanthology.org/2024.clinicalnlp-1.35.bib

https://aclanthology.org/2024.clinicalnlp-1.35/

@inproceedings{gundabathula-kolar-2024-promptmind-team, title = "{P}rompt{M}ind Team at {MEDIQA}-{CORR} 2024: Improving Clinical Text Correction with Error Categorization and {LLM} Ensembles", author = "Gundabathula, Satya and Kolar, Sriram", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.35", doi = "10.18653/v1/2024.clinicalnlp-1.35", pages = "367--373", abstract = "This paper describes our approach to the MEDIQA-CORR shared task, which involves error detection and correction in clinical notes curated by medical professionals. This task involves handling three subtasks: detecting the presence of errors, identifying the specific sentence containing the error, and correcting it. Through our work, we aim to assess the capabilities of Large Language Models (LLMs) trained on a vast corpora of internet data that contain both factual and unreliable information. We propose to comprehensively address all subtasks together, and suggest employing a unique prompt-based in-context learning strategy. We will evaluate its efficacy in this specialized task demanding a combination of general reasoning and medical knowledge. In medical systems where prediction errors can have grave consequences, we propose leveraging self-consistency and ensemble methods to enhance error correction and error detection performance.", }

This paper describes our approach to the MEDIQA-CORR shared task, which involves error detection and correction in clinical notes curated by medical professionals. This task involves handling three subtasks: detecting the presence of errors, identifying the specific sentence containing the error, and correcting it. Through our work, we aim to assess the capabilities of Large Language Models (LLMs) trained on a vast corpora of internet data that contain both factual and unreliable information. We propose to comprehensively address all subtasks together, and suggest employing a unique prompt-based in-context learning strategy. We will evaluate its efficacy in this specialized task demanding a combination of general reasoning and medical knowledge. In medical systems where prediction errors can have grave consequences, we propose leveraging self-consistency and ensemble methods to enhance error correction and error detection performance.

PromptMind Team at MEDIQA-CORR 2024: Improving Clinical Text Correction with Error Categorization and LLM Ensembles

clinicalnlp-1.35

Poster

2405.08373

https://aclanthology.org/2024.clinicalnlp-1.36.bib

https://aclanthology.org/2024.clinicalnlp-1.36/

@inproceedings{jadhav-etal-2024-maven, title = "Maven at {MEDIQA}-{CORR} 2024: Leveraging {RAG} and Medical {LLM} for Error Detection and Correction in Medical Notes", author = "Jadhav, Suramya and Shanbhag, Abhay and Joshi, Sumedh and Date, Atharva and Sonawane, Sheetal", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.36", doi = "10.18653/v1/2024.clinicalnlp-1.36", pages = "374--381", abstract = "Addressing the critical challenge of identifying and rectifying medical errors in clinical notes, we present a novel approach tailored for the MEDIQA-CORR task @ NAACL-ClinicalNLP 2024, which comprises three subtasks: binary classification, span identification, and natural language generation for error detection and correction. Binary classification involves detecting whether the text contains a medical error; span identification entails identifying the text span associated with any detected error; and natural language generation focuses on providing a free text correction if a medical error exists. Our proposed architecture leverages Named Entity Recognition (NER) for identifying disease-related terms, Retrieval-Augmented Generation (RAG) for contextual understanding from external datasets, and a quantized and fine-tuned Palmyra model for error correction. Our model achieved a global rank of 5 with an aggregate score of 0.73298, calculated as the mean of ROUGE-1-F, BERTScore, and BLEURT scores.", }

Addressing the critical challenge of identifying and rectifying medical errors in clinical notes, we present a novel approach tailored for the MEDIQA-CORR task @ NAACL-ClinicalNLP 2024, which comprises three subtasks: binary classification, span identification, and natural language generation for error detection and correction. Binary classification involves detecting whether the text contains a medical error; span identification entails identifying the text span associated with any detected error; and natural language generation focuses on providing a free text correction if a medical error exists. Our proposed architecture leverages Named Entity Recognition (NER) for identifying disease-related terms, Retrieval-Augmented Generation (RAG) for contextual understanding from external datasets, and a quantized and fine-tuned Palmyra model for error correction. Our model achieved a global rank of 5 with an aggregate score of 0.73298, calculated as the mean of ROUGE-1-F, BERTScore, and BLEURT scores.

Maven at MEDIQA-CORR 2024: Leveraging RAG and Medical LLM for Error Detection and Correction in Medical Notes

clinicalnlp-1.36

Poster

https://aclanthology.org/2024.clinicalnlp-1.37.bib

https://aclanthology.org/2024.clinicalnlp-1.37/

@inproceedings{haddadan-etal-2024-lailab, title = "{LAIL}ab at Chemotimelines 2024: Finetuning sequence-to-sequence language models for temporal relation extraction towards cancer patient undergoing chemotherapy treatment", author = "Haddadan, Shohreh and Le, Tuan-Dung and Duong, Thanh and Thieu, Thanh", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.37", doi = "10.18653/v1/2024.clinicalnlp-1.37", pages = "382--393", abstract = "In this paper, we report our effort to tackle the challenge of extracting chemotimelines from EHR notes across a dataset of three cancer types. We focus on the two subtasks: 1) detection and classification of temporal relations given the annotated chemotherapy events and time expressions and 2) directly extracting patient chemotherapy timelines from EHR notes. We address both subtasks using Large Language Models. Our best-performing methods in both subtasks use Flan-T5, an instruction-tuned language model. Our proposed system achieves the highest average score in both subtasks. Our results underscore the effectiveness of finetuning general-domain large language models in domain-specific and unseen tasks.", }

In this paper, we report our effort to tackle the challenge of extracting chemotimelines from EHR notes across a dataset of three cancer types. We focus on the two subtasks: 1) detection and classification of temporal relations given the annotated chemotherapy events and time expressions and 2) directly extracting patient chemotherapy timelines from EHR notes. We address both subtasks using Large Language Models. Our best-performing methods in both subtasks use Flan-T5, an instruction-tuned language model. Our proposed system achieves the highest average score in both subtasks. Our results underscore the effectiveness of finetuning general-domain large language models in domain-specific and unseen tasks.

LAILab at Chemotimelines 2024: Finetuning sequence-to-sequence language models for temporal relation extraction towards cancer patient undergoing chemotherapy treatment

clinicalnlp-1.37

Poster

https://aclanthology.org/2024.clinicalnlp-1.38.bib

https://aclanthology.org/2024.clinicalnlp-1.38/

@inproceedings{sharma-etal-2024-lexicans, title = "Lexicans at Chemotimelines 2024: Chemotimeline Chronicles - Leveraging Large Language Models ({LLM}s) for Temporal Relations Extraction in Oncological Electronic Health Records", author = "Sharma, Vishakha and Fernandez, Andres and Ioanovici, Andrei and Talby, David and Buijs, Frederik", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.38", doi = "10.18653/v1/2024.clinicalnlp-1.38", pages = "394--405", abstract = "Automatic generation of chemotherapy treatment timelines from electronic health records (EHRs) notes not only streamlines clinical workflows but also promotes better coordination and improvements in cancer treatment and quality of care. This paper describes the submission to the Chemotimelines 2024 shared task that aims to automatically build a chemotherapy treatment timeline for each patient using their complete set of EHR notes, spanning various sources such as primary care provider, oncology, discharge summaries, emergency department, pathology, radiology, and more. We report results from two large language models (LLMs), namely Llama 2 and Mistral 7B, applied to the shared task data using zero-shot prompting.", }

Automatic generation of chemotherapy treatment timelines from electronic health records (EHRs) notes not only streamlines clinical workflows but also promotes better coordination and improvements in cancer treatment and quality of care. This paper describes the submission to the Chemotimelines 2024 shared task that aims to automatically build a chemotherapy treatment timeline for each patient using their complete set of EHR notes, spanning various sources such as primary care provider, oncology, discharge summaries, emergency department, pathology, radiology, and more. We report results from two large language models (LLMs), namely Llama 2 and Mistral 7B, applied to the shared task data using zero-shot prompting.

Lexicans at Chemotimelines 2024: Chemotimeline Chronicles - Leveraging Large Language Models (LLMs) for Temporal Relations Extraction in Oncological Electronic Health Records

clinicalnlp-1.38

Poster

https://aclanthology.org/2024.clinicalnlp-1.39.bib

https://aclanthology.org/2024.clinicalnlp-1.39/

@inproceedings{bannour-etal-2024-team, title = "Team {NLP}eers at Chemotimelines 2024: Evaluation of two timeline extraction methods, can generative {LLM} do it all or is smaller model fine-tuning still relevant ?", author = "Bannour, Nesrine and Andrew, Judith Jeyafreeda and Vincent, Marc", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.39", doi = "10.18653/v1/2024.clinicalnlp-1.39", pages = "406--416", abstract = "This paper presents our two deep learning-based approaches to participate in subtask 1 of the Chemotimelines 2024 Shared task. The first uses a fine-tuning strategy on a relatively small general domain Masked Language Model (MLM) model, with additional normalization steps obtained using a simple Large Language Model (LLM) prompting technique. The second is an LLM-based approach combining advanced automated prompt search with few-shot in-context learning using the DSPy framework.Our results confirm the continued relevance of the smaller MLM fine-tuned model. It also suggests that the automated few-shot LLM approach can perform close to the fine-tuning-based method without extra LLM normalization and be advantageous under scarce data access conditions. We finally hint at the possibility to choose between lower training examples or lower computing resources requirements when considering both methods.", }

This paper presents our two deep learning-based approaches to participate in subtask 1 of the Chemotimelines 2024 Shared task. The first uses a fine-tuning strategy on a relatively small general domain Masked Language Model (MLM) model, with additional normalization steps obtained using a simple Large Language Model (LLM) prompting technique. The second is an LLM-based approach combining advanced automated prompt search with few-shot in-context learning using the DSPy framework.Our results confirm the continued relevance of the smaller MLM fine-tuned model. It also suggests that the automated few-shot LLM approach can perform close to the fine-tuning-based method without extra LLM normalization and be advantageous under scarce data access conditions. We finally hint at the possibility to choose between lower training examples or lower computing resources requirements when considering both methods.

Team NLPeers at Chemotimelines 2024: Evaluation of two timeline extraction methods, can generative LLM do it all or is smaller model fine-tuning still relevant ?

clinicalnlp-1.39

Poster

https://aclanthology.org/2024.clinicalnlp-1.40.bib

https://aclanthology.org/2024.clinicalnlp-1.40/

@inproceedings{tan-etal-2024-kclab, title = "{KCL}ab at Chemotimelines 2024: End-to-end system for chemotherapy timeline extraction {--} Subtask2", author = "Tan, Yukun and Dede, Merve and Chen, Ken", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.40", doi = "10.18653/v1/2024.clinicalnlp-1.40", pages = "417--421", abstract = "This paper presents our participation in the Chemotimelines 2024 subtask2, focusing on the development of an end-to-end system for chemotherapy timeline extraction. We initially adopt a basic framework from subtask2, utilizing Apache cTAKES for entity recognition and a BERT-based model for classifying the temporal relationship between chemotherapy events and associated times. Subsequently, we enhance this pipeline through two key directions: first, by expanding the exploration of the system, achieved by extending the search dictionary of cTAKES with the UMLS database; second, by reducing false positives through preprocessing of clinical notes and implementing filters to reduce the potential errors from the BERT-based model. To validate the effectiveness of our framework, we conduct extensive experiments using clinical notes from breast, ovarian, and melanoma cancer cases. Our results demonstrate improvements over the previous approach.", }

This paper presents our participation in the Chemotimelines 2024 subtask2, focusing on the development of an end-to-end system for chemotherapy timeline extraction. We initially adopt a basic framework from subtask2, utilizing Apache cTAKES for entity recognition and a BERT-based model for classifying the temporal relationship between chemotherapy events and associated times. Subsequently, we enhance this pipeline through two key directions: first, by expanding the exploration of the system, achieved by extending the search dictionary of cTAKES with the UMLS database; second, by reducing false positives through preprocessing of clinical notes and implementing filters to reduce the potential errors from the BERT-based model. To validate the effectiveness of our framework, we conduct extensive experiments using clinical notes from breast, ovarian, and melanoma cancer cases. Our results demonstrate improvements over the previous approach.

KCLab at Chemotimelines 2024: End-to-end system for chemotherapy timeline extraction – Subtask2

clinicalnlp-1.40

Poster

https://aclanthology.org/2024.clinicalnlp-1.41.bib

https://aclanthology.org/2024.clinicalnlp-1.41/

@inproceedings{joy-etal-2024-project, title = "Project {PRIMUS} at {EHRSQL} 2024 : Text-to-{SQL} Generation using Large Language Model for {EHR} Analysis", author = "Joy, Sourav and Ahmed, Rohan and Saha, Argha and Habil, Minhaj and Das, Utsho and Bhowmik, Partha", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.41", doi = "10.18653/v1/2024.clinicalnlp-1.41", pages = "422--427", abstract = "This paper explores the application of the sqlcoders model, a pre-trained neural network, for automatic SQL query generation from natural language questions. We focus on the model{'}s internal functionality and demonstrate its effectiveness on a domain-specific validation dataset provided by EHRSQL. The sqlcoders model, based on transformers with attention mechanisms, has been trained on paired examples of natural language questions and corresponding SQL queries. It takes advantage of a carefully crafted prompt that incorporates the database schema alongside the question to guide the model towards the desired output format.", }

This paper explores the application of the sqlcoders model, a pre-trained neural network, for automatic SQL query generation from natural language questions. We focus on the model{'}s internal functionality and demonstrate its effectiveness on a domain-specific validation dataset provided by EHRSQL. The sqlcoders model, based on transformers with attention mechanisms, has been trained on paired examples of natural language questions and corresponding SQL queries. It takes advantage of a carefully crafted prompt that incorporates the database schema alongside the question to guide the model towards the desired output format.

Project PRIMUS at EHRSQL 2024 : Text-to-SQL Generation using Large Language Model for EHR Analysis

clinicalnlp-1.41

Poster

https://aclanthology.org/2024.clinicalnlp-1.42.bib

https://aclanthology.org/2024.clinicalnlp-1.42/

@inproceedings{zhang-etal-2024-nyulangone, title = "{NYUL}angone at Chemotimelines 2024: Utilizing Open-Weights Large Language Models for Chemotherapy Event Extraction", author = "Zhang, Jeff and Aphinyanaphongs, Yin and Cardillo, Anthony", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.42", doi = "10.18653/v1/2024.clinicalnlp-1.42", pages = "428--430", abstract = "The extraction of chemotherapy treatment timelines from clinical narratives poses significant challenges due to the complexity of medical language and patient-specific treatment regimens. This paper describes the NYULangone team{'}s approach to Subtask 2 of the Chemotimelines 2024 shared task, focusing on leveraging a locally hosted Large Language Model (LLM), Mixtral 8x7B (Mistral AI, France), to interpret and extract relevant events from clinical notes without relying on domain-specific training data. Despite facing challenges due to the task{'}s complexity and the current capacity of open-source AI, our methodology highlights the future potential of local foundational LLMs in specialized domains like biomedical data processing.", }

The extraction of chemotherapy treatment timelines from clinical narratives poses significant challenges due to the complexity of medical language and patient-specific treatment regimens. This paper describes the NYULangone team{'}s approach to Subtask 2 of the Chemotimelines 2024 shared task, focusing on leveraging a locally hosted Large Language Model (LLM), Mixtral 8x7B (Mistral AI, France), to interpret and extract relevant events from clinical notes without relying on domain-specific training data. Despite facing challenges due to the task{'}s complexity and the current capacity of open-source AI, our methodology highlights the future potential of local foundational LLMs in specialized domains like biomedical data processing.

NYULangone at Chemotimelines 2024: Utilizing Open-Weights Large Language Models for Chemotherapy Event Extraction

clinicalnlp-1.42

Poster

https://aclanthology.org/2024.clinicalnlp-1.43.bib

https://aclanthology.org/2024.clinicalnlp-1.43/

@inproceedings{somov-etal-2024-airi, title = "{AIRI} {NLP} Team at {EHRSQL} 2024 Shared Task: T5 and Logistic Regression to the Rescue", author = "Somov, Oleg and Dontsov, Alexey and Tutubalina, Elena", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.43", doi = "10.18653/v1/2024.clinicalnlp-1.43", pages = "431--438", abstract = "This paper presents a system developed for the Clinical NLP 2024 Shared Task, focusing on reliable text-to-SQL modeling on Electronic Health Records (EHRs). The goal is to create a model that accurately generates SQL queries for answerable questions while avoiding incorrect responses and handling unanswerable queries. Our approach comprises three main components: a query correspondence model, a text-to-SQL model, and an SQL verifier.For the query correspondence model, we trained a logistic regression model using hand-crafted features to distinguish between answerable and unanswerable queries. As for the text-to-SQL model, we utilized T5-3B as a pretrained language model, further fine-tuned on pairs of natural language questions and corresponding SQL queries. Finally, we applied the SQL verifier to inspect the resulting SQL queries.During the evaluation stage of the shared task, our system achieved an accuracy of 68.9 {\%} (metric version without penalty), positioning it at the fifth-place ranking. While our approach did not surpass solutions based on large language models (LMMs) like ChatGPT, it demonstrates the promising potential of domain-specific specialized models that are more resource-efficient. The code is publicly available at https://github.com/runnerup96/EHRSQL-text2sql-solution.", }

This paper presents a system developed for the Clinical NLP 2024 Shared Task, focusing on reliable text-to-SQL modeling on Electronic Health Records (EHRs). The goal is to create a model that accurately generates SQL queries for answerable questions while avoiding incorrect responses and handling unanswerable queries. Our approach comprises three main components: a query correspondence model, a text-to-SQL model, and an SQL verifier.For the query correspondence model, we trained a logistic regression model using hand-crafted features to distinguish between answerable and unanswerable queries. As for the text-to-SQL model, we utilized T5-3B as a pretrained language model, further fine-tuned on pairs of natural language questions and corresponding SQL queries. Finally, we applied the SQL verifier to inspect the resulting SQL queries.During the evaluation stage of the shared task, our system achieved an accuracy of 68.9 {\%} (metric version without penalty), positioning it at the fifth-place ranking. While our approach did not surpass solutions based on large language models (LMMs) like ChatGPT, it demonstrates the promising potential of domain-specific specialized models that are more resource-efficient. The code is publicly available at https://github.com/runnerup96/EHRSQL-text2sql-solution.

AIRI NLP Team at EHRSQL 2024 Shared Task: T5 and Logistic Regression to the Rescue

clinicalnlp-1.43

Poster

https://aclanthology.org/2024.clinicalnlp-1.44.bib

https://aclanthology.org/2024.clinicalnlp-1.44/

@inproceedings{bauer-etal-2024-ikim, title = "{IKIM} at {MEDIQA}-{M}3{G} 2024: Multilingual Visual Question-Answering for Dermatology through {VLM} Fine-tuning and {LLM} Translations", author = "Bauer, Marie and Seibold, Constantin and Kleesiek, Jens and Dada, Amin", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.44", doi = "10.18653/v1/2024.clinicalnlp-1.44", pages = "439--447", abstract = "This paper presents our solution to the MEDIQA-M3G Challenge at NAACL-ClinicalNLP 2024. We participated in all three languages, ranking first in Chinese and Spanish and third in English. Our approach utilizes LLaVA-med, an open-source, medical vision-language model (VLM) for visual question-answering in Chinese, and Mixtral-8x7B-instruct, a Large Language Model (LLM) for a subsequent translation into English and Spanish. In addition to our final method, we experiment with alternative approaches: Training three different models for each language instead of translating the results from one model, using different combinations and numbers of input images, and additional training on publicly available data that was not part of the original challenge training set.", }

This paper presents our solution to the MEDIQA-M3G Challenge at NAACL-ClinicalNLP 2024. We participated in all three languages, ranking first in Chinese and Spanish and third in English. Our approach utilizes LLaVA-med, an open-source, medical vision-language model (VLM) for visual question-answering in Chinese, and Mixtral-8x7B-instruct, a Large Language Model (LLM) for a subsequent translation into English and Spanish. In addition to our final method, we experiment with alternative approaches: Training three different models for each language instead of translating the results from one model, using different combinations and numbers of input images, and additional training on publicly available data that was not part of the original challenge training set.

IKIM at MEDIQA-M3G 2024: Multilingual Visual Question-Answering for Dermatology through VLM Fine-tuning and LLM Translations

clinicalnlp-1.44

Poster

https://aclanthology.org/2024.clinicalnlp-1.45.bib

https://aclanthology.org/2024.clinicalnlp-1.45/

@inproceedings{garcia-lithgow-serrano-2024-neui, title = "{NEUI} at {MEDIQA}-{M}3{G} 2024: Medical {VQA} through consensus", author = "Garc{\'\i}a, Ricardo and Lithgow-Serrano, Oscar", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.45", doi = "10.18653/v1/2024.clinicalnlp-1.45", pages = "448--460", abstract = "This document describes our solution to the MEDIQA-M3G: Multilingual {\&} Multimodal Medical Answer Generation. To build our solution, we leveraged two pre-trained models, a Visual Language Model (VLM) and a Large Language Model (LLM). We fine-tuned both models using the MEDIQA-M3G and MEDIQA-CORR training datasets, respectively. In the first stage, the VLM provides singular responses for each pair of image {\&} text inputs in a case. In the second stage, the LLM consolidates the VLM responses using it as context among the original text input. By changing the original English case content field in the context component of the second stage to the one in Spanish, we adapt the pipeline to generate submissions in English and Spanish. We performed an ablation study to explore the impact of the different models{'} capabilities, such as multimodality and reasoning, on the MEDIQA-M3G task. Our approach favored privacy and feasibility by adopting open-source and self-hosted small models and ranked 4th in English and 2nd in Spanish.", }

This document describes our solution to the MEDIQA-M3G: Multilingual {\&} Multimodal Medical Answer Generation. To build our solution, we leveraged two pre-trained models, a Visual Language Model (VLM) and a Large Language Model (LLM). We fine-tuned both models using the MEDIQA-M3G and MEDIQA-CORR training datasets, respectively. In the first stage, the VLM provides singular responses for each pair of image {\&} text inputs in a case. In the second stage, the LLM consolidates the VLM responses using it as context among the original text input. By changing the original English case content field in the context component of the second stage to the one in Spanish, we adapt the pipeline to generate submissions in English and Spanish. We performed an ablation study to explore the impact of the different models{'} capabilities, such as multimodality and reasoning, on the MEDIQA-M3G task. Our approach favored privacy and feasibility by adopting open-source and self-hosted small models and ranked 4th in English and 2nd in Spanish.

NEUI at MEDIQA-M3G 2024: Medical VQA through consensus

clinicalnlp-1.45

Poster

https://aclanthology.org/2024.clinicalnlp-1.46.bib

https://aclanthology.org/2024.clinicalnlp-1.46/

@inproceedings{pajaro-etal-2024-verbanexai, title = "{V}erba{N}ex{AI} at {MEDIQA}-{CORR}: Efficacy of {GRU} with {B}io{W}ord{V}ec and {C}linical{BERT} in Error Correction in Clinical Notes", author = "Pajaro, Juan and Puertas, Edwin and Villate, David and Estrada, Laura and Tinjaca, Laura", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.46", doi = "10.18653/v1/2024.clinicalnlp-1.46", pages = "461--469", abstract = "The automatic identification of medical errors in clinical notes is crucial for improving the quality of healthcare services.LLMs emerge as a powerful artificial intelligence tool for automating this task. However, LLMs present vulnerabilities, high costs, and sometimes a lack of transparency. This article addresses the detection of medical errors through the fine-tuning approach, conducting a comprehensive comparison between various models and exploring in depth the components of the machine learning pipeline. The results obtained with the fine-tuned ClinicalBert and Gated recurrent units (Gru) models show an accuracy of 0.56 and 0.55, respectively. This approach not only mitigates the problems associated with the use of LLMs but also demonstrates how exhaustive iteration in critical phases of the pipeline, especially in feature selection, can facilitate the automation of clinical record analysis.", }

The automatic identification of medical errors in clinical notes is crucial for improving the quality of healthcare services.LLMs emerge as a powerful artificial intelligence tool for automating this task. However, LLMs present vulnerabilities, high costs, and sometimes a lack of transparency. This article addresses the detection of medical errors through the fine-tuning approach, conducting a comprehensive comparison between various models and exploring in depth the components of the machine learning pipeline. The results obtained with the fine-tuned ClinicalBert and Gated recurrent units (Gru) models show an accuracy of 0.56 and 0.55, respectively. This approach not only mitigates the problems associated with the use of LLMs but also demonstrates how exhaustive iteration in critical phases of the pipeline, especially in feature selection, can facilitate the automation of clinical record analysis.

VerbaNexAI at MEDIQA-CORR: Efficacy of GRU with BioWordVec and ClinicalBERT in Error Correction in Clinical Notes

clinicalnlp-1.46

Poster

https://aclanthology.org/2024.clinicalnlp-1.47.bib

https://aclanthology.org/2024.clinicalnlp-1.47/

@inproceedings{valiev-tutubalina-2024-hse, title = "{HSE} {NLP} Team at {MEDIQA}-{CORR} 2024 Task: In-Prompt Ensemble with Entities and Knowledge Graph for Medical Error Correction", author = "Valiev, Airat and Tutubalina, Elena", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.47", doi = "10.18653/v1/2024.clinicalnlp-1.47", pages = "470--482", abstract = "This paper presents our LLM-based system designed for the MEDIQA-CORR @ NAACL-ClinicalNLP 2024 Shared Task 3, focusing on medical error detection and correction in medical records. Our approach consists of three key components: entity extraction, prompt engineering, and ensemble. First, we automatically extract biomedical entities such as therapies, diagnoses, and biological species. Next, we explore few-shot learning techniques and incorporate graph information from the MeSH database for the identified entities. Finally, we investigate two methods for ensembling: (i) combining the predictions of three previous LLMs using an AND strategy within a prompt and (ii) integrating the previous predictions into the prompt as separate {`}expert{'} solutions, accompanied by trust scores representing their performance. The latter system ranked second with a BERTScore score of 0.8059 and third with an aggregated score of 0.7806 out of the 15 teams{'} solutions in the shared task.", }

This paper presents our LLM-based system designed for the MEDIQA-CORR @ NAACL-ClinicalNLP 2024 Shared Task 3, focusing on medical error detection and correction in medical records. Our approach consists of three key components: entity extraction, prompt engineering, and ensemble. First, we automatically extract biomedical entities such as therapies, diagnoses, and biological species. Next, we explore few-shot learning techniques and incorporate graph information from the MeSH database for the identified entities. Finally, we investigate two methods for ensembling: (i) combining the predictions of three previous LLMs using an AND strategy within a prompt and (ii) integrating the previous predictions into the prompt as separate {`}expert{'} solutions, accompanied by trust scores representing their performance. The latter system ranked second with a BERTScore score of 0.8059 and third with an aggregated score of 0.7806 out of the 15 teams{'} solutions in the shared task.

HSE NLP Team at MEDIQA-CORR 2024 Task: In-Prompt Ensemble with Entities and Knowledge Graph for Medical Error Correction

clinicalnlp-1.47

Poster

https://aclanthology.org/2024.clinicalnlp-1.48.bib

https://aclanthology.org/2024.clinicalnlp-1.48/

@inproceedings{wang-etal-2024-wonder, title = "Wonder at Chemotimelines 2024: {M}ed{T}imeline: An End-to-End {NLP} System for Timeline Extraction from Clinical Narratives", author = "Wang, Liwei and Lu, Qiuhao and Li, Rui and Fu, Sunyang and Liu, Hongfang", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.48", doi = "10.18653/v1/2024.clinicalnlp-1.48", pages = "483--487", abstract = "Extracting timeline information from clinical narratives is critical for cancer research and practice using electronic health records (EHRs). In this study, we apply MedTimeline, our end-to-end hybrid NLP system combining large language model, deep learning with knowledge engineering, to the ChemoTimeLine challenge subtasks. Our experiment results in 0.83, 0.90, 0.84, and 0.53, 0.63, 0.39, respectively, for subtask1 and subtask2 in breast, melanoma and ovarian cancer.", }

Extracting timeline information from clinical narratives is critical for cancer research and practice using electronic health records (EHRs). In this study, we apply MedTimeline, our end-to-end hybrid NLP system combining large language model, deep learning with knowledge engineering, to the ChemoTimeLine challenge subtasks. Our experiment results in 0.83, 0.90, 0.84, and 0.53, 0.63, 0.39, respectively, for subtask1 and subtask2 in breast, melanoma and ovarian cancer.

Wonder at Chemotimelines 2024: MedTimeline: An End-to-End NLP System for Timeline Extraction from Clinical Narratives

clinicalnlp-1.48

Poster

https://aclanthology.org/2024.clinicalnlp-1.49.bib

https://aclanthology.org/2024.clinicalnlp-1.49/

@inproceedings{gema-etal-2024-edinburgh-clinical, title = "{E}dinburgh Clinical {NLP} at {MEDIQA}-{CORR} 2024: Guiding Large Language Models with Hints", author = "Gema, Aryo and Lee, Chaeeun and Minervini, Pasquale and Daines, Luke and Simpson, T. and Alex, Beatrice", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.49", doi = "10.18653/v1/2024.clinicalnlp-1.49", pages = "488--501", abstract = "The MEDIQA-CORR 2024 shared task aims to assess the ability of Large Language Models (LLMs) to identify and correct medical errors in clinical notes. In this study, we evaluate the capability of general LLMs, specifically GPT-3.5 and GPT-4, to identify and correct medical errors with multiple prompting strategies. Recognising the limitation of LLMs in generating accurate corrections only via prompting strategies, we propose incorporating error-span predictions from a smaller, fine-tuned model in two ways: 1) by presenting it as a hint in the prompt and 2) by framing it as multiple-choice questions from which the LLM can choose the best correction. We found that our proposed prompting strategies significantly improve the LLM{'}s ability to generate corrections. Our best-performing solution with 8-shot + CoT + hints ranked sixth in the shared task leaderboard. Additionally, our comprehensive analyses show the impact of the location of the error sentence, the prompted role, and the position of the multiple-choice option on the accuracy of the LLM. This prompts further questions about the readiness of LLM to be implemented in real-world clinical settings.", }

The MEDIQA-CORR 2024 shared task aims to assess the ability of Large Language Models (LLMs) to identify and correct medical errors in clinical notes. In this study, we evaluate the capability of general LLMs, specifically GPT-3.5 and GPT-4, to identify and correct medical errors with multiple prompting strategies. Recognising the limitation of LLMs in generating accurate corrections only via prompting strategies, we propose incorporating error-span predictions from a smaller, fine-tuned model in two ways: 1) by presenting it as a hint in the prompt and 2) by framing it as multiple-choice questions from which the LLM can choose the best correction. We found that our proposed prompting strategies significantly improve the LLM{'}s ability to generate corrections. Our best-performing solution with 8-shot + CoT + hints ranked sixth in the shared task leaderboard. Additionally, our comprehensive analyses show the impact of the location of the error sentence, the prompted role, and the position of the multiple-choice option on the accuracy of the LLM. This prompts further questions about the readiness of LLM to be implemented in real-world clinical settings.

Edinburgh Clinical NLP at MEDIQA-CORR 2024: Guiding Large Language Models with Hints

clinicalnlp-1.49

Poster

2405.18028

https://aclanthology.org/2024.clinicalnlp-1.50.bib

https://aclanthology.org/2024.clinicalnlp-1.50/

@inproceedings{vashisht-etal-2024-umass, title = "{UM}ass-{B}io{NLP} at {MEDIQA}-{M}3{G} 2024: {D}erm{P}rompt - A Systematic Exploration of Prompt Engineering with {GPT}-4{V} for Dermatological Diagnosis", author = "Vashisht, Parth and Lodha, Abhilasha and Maddipatla, Mukta and Yao, Zonghai and Mitra, Avijit and Yang, Zhichao and Kwon, Sunjae and Wang, Junda and Yu, Hong", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.50", doi = "10.18653/v1/2024.clinicalnlp-1.50", pages = "502--525", abstract = "This paper presents our team{'}s participation in the MEDIQA-ClinicalNLP 2024 shared task B. We present a novel approach to diagnosing clinical dermatology cases by integrating large multimodal models, specifically leveraging the capabilities of GPT-4V under a retriever and a re-ranker framework. Our investigation reveals that GPT-4V, when used as a retrieval agent, can accurately retrieve the correct skin condition 85{\%} of the time using dermatological images and brief patient histories. Additionally, we empirically show that Naive Chain-of-Thought (CoT) works well for retrieval while Medical Guidelines Grounded CoT is required for accurate dermatological diagnosis. Further, we introduce a Multi-Agent Conversation (MAC) framework and show it{'}s superior performance and potential over the best CoT strategy. The experiments suggest that using naive CoT for retrieval and multi-agent conversation for critique-based diagnosis, GPT-4V can lead to an early and accurate diagnosis of dermatological conditions. The implications of this work extend to improving diagnostic workflows, supporting dermatological education, and enhancing patient care by providing a scalable, accessible, and accurate diagnostic tool.", }

This paper presents our team{'}s participation in the MEDIQA-ClinicalNLP 2024 shared task B. We present a novel approach to diagnosing clinical dermatology cases by integrating large multimodal models, specifically leveraging the capabilities of GPT-4V under a retriever and a re-ranker framework. Our investigation reveals that GPT-4V, when used as a retrieval agent, can accurately retrieve the correct skin condition 85{\%} of the time using dermatological images and brief patient histories. Additionally, we empirically show that Naive Chain-of-Thought (CoT) works well for retrieval while Medical Guidelines Grounded CoT is required for accurate dermatological diagnosis. Further, we introduce a Multi-Agent Conversation (MAC) framework and show it{'}s superior performance and potential over the best CoT strategy. The experiments suggest that using naive CoT for retrieval and multi-agent conversation for critique-based diagnosis, GPT-4V can lead to an early and accurate diagnosis of dermatological conditions. The implications of this work extend to improving diagnostic workflows, supporting dermatological education, and enhancing patient care by providing a scalable, accessible, and accurate diagnostic tool.

UMass-BioNLP at MEDIQA-M3G 2024: DermPrompt - A Systematic Exploration of Prompt Engineering with GPT-4V for Dermatological Diagnosis

clinicalnlp-1.50

Poster

https://aclanthology.org/2024.clinicalnlp-1.51.bib

https://aclanthology.org/2024.clinicalnlp-1.51/

@inproceedings{hwang-etal-2024-ku, title = "{KU}-{DMIS} at {MEDIQA}-{CORR} 2024: Exploring the Reasoning Capabilities of Small Language Models in Medical Error Correction", author = "Hwang, Hyeon and Lee, Taewhoo and Kim, Hyunjae and Kang, Jaewoo", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.51", doi = "10.18653/v1/2024.clinicalnlp-1.51", pages = "526--536", abstract = "Recent advancements in large language models (LM) like OpenAI{'}s GPT-4 have shown promise in healthcare, particularly in medical question answering and clinical applications. However, their deployment raises privacy concerns and their size limits use in resource-constrained environments.Smaller open-source LMs have emerged as alternatives, but their reliability in medicine remains underexplored.This study evaluates small LMs in the medical field using the MEDIQA-CORR 2024 task, which assesses the ability of models to identify and correct errors in clinical notes. Initially, zero-shot inference and simple fine-tuning of small models resulted in poor performance. When fine-tuning with chain-of-thought (CoT) reasoning using synthetic data generated by GPT-4, their performance significantly improved. Meerkat-7B, a small LM trained with medical CoT reasoning, demonstrated notable performance gains. Our model outperforms other small non-commercial LMs and some larger models, achieving a 73.36 aggregate score on MEDIQA-CORR 2024.", }

Recent advancements in large language models (LM) like OpenAI{'}s GPT-4 have shown promise in healthcare, particularly in medical question answering and clinical applications. However, their deployment raises privacy concerns and their size limits use in resource-constrained environments.Smaller open-source LMs have emerged as alternatives, but their reliability in medicine remains underexplored.This study evaluates small LMs in the medical field using the MEDIQA-CORR 2024 task, which assesses the ability of models to identify and correct errors in clinical notes. Initially, zero-shot inference and simple fine-tuning of small models resulted in poor performance. When fine-tuning with chain-of-thought (CoT) reasoning using synthetic data generated by GPT-4, their performance significantly improved. Meerkat-7B, a small LM trained with medical CoT reasoning, demonstrated notable performance gains. Our model outperforms other small non-commercial LMs and some larger models, achieving a 73.36 aggregate score on MEDIQA-CORR 2024.

KU-DMIS at MEDIQA-CORR 2024: Exploring the Reasoning Capabilities of Small Language Models in Medical Error Correction

clinicalnlp-1.51

Poster

https://aclanthology.org/2024.clinicalnlp-1.52.bib

https://aclanthology.org/2024.clinicalnlp-1.52/

@inproceedings{alzghoul-etal-2024-cld, title = "{CLD}-{MEC} at {MEDIQA}- {CORR} 2024 Task: {GPT}-4 Multi-Stage Clinical Chain of Thought Prompting for Medical Errors Detection and Correction", author = "Alzghoul, Renad and Ayaabdelhaq, Ayaabdelhaq and Tabaza, Abdulrahman and Altamimi, Ahmad", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.52", doi = "10.18653/v1/2024.clinicalnlp-1.52", pages = "537--556", abstract = "This paper demonstrates CLD-MEC team submission to the MEDIQA-CORR 2024 shared task for identifying and correcting medical errors from clinical notes. We developed a framework to track two main types of medical errors: diagnostics and medical management-related errors. The tracking framework is implied utilizing a GPT-4 multi-stage prompting-based pipeline that ends with the three downstream tasks: classification of medical error existence (Task 1), identification of error location (Task 2), and correction error (Task 3). Throughout the pipeline, we employed clinical Chain of Thought (CoT) and Chain-of-Verification (CoVe) techniques to mitigate the hallucination and enforce the clinical context learning. The model performance is acceptable, given it is based on zero-shot learning. In addition, we developed a RAG system injected with clinical practice guidelines as an external knowledge datastore. Our RAG is based on the Bio{\_}ClinicalBERT as a vector embedding model. However, our RAG system failed to get the desired results. We proposed recommendations to be investigated in future research work to overcome the limitations of our approach.", }

This paper demonstrates CLD-MEC team submission to the MEDIQA-CORR 2024 shared task for identifying and correcting medical errors from clinical notes. We developed a framework to track two main types of medical errors: diagnostics and medical management-related errors. The tracking framework is implied utilizing a GPT-4 multi-stage prompting-based pipeline that ends with the three downstream tasks: classification of medical error existence (Task 1), identification of error location (Task 2), and correction error (Task 3). Throughout the pipeline, we employed clinical Chain of Thought (CoT) and Chain-of-Verification (CoVe) techniques to mitigate the hallucination and enforce the clinical context learning. The model performance is acceptable, given it is based on zero-shot learning. In addition, we developed a RAG system injected with clinical practice guidelines as an external knowledge datastore. Our RAG is based on the Bio{\_}ClinicalBERT as a vector embedding model. However, our RAG system failed to get the desired results. We proposed recommendations to be investigated in future research work to overcome the limitations of our approach.

CLD-MEC at MEDIQA- CORR 2024 Task: GPT-4 Multi-Stage Clinical Chain of Thought Prompting for Medical Errors Detection and Correction

clinicalnlp-1.52

Poster

https://aclanthology.org/2024.clinicalnlp-1.53.bib

https://aclanthology.org/2024.clinicalnlp-1.53/

@inproceedings{yao-etal-2024-overview, title = "Overview of the 2024 Shared Task on Chemotherapy Treatment Timeline Extraction", author = "Yao, Jiarui and Hochheiser, Harry and Yoon, WonJin and Goldner, Eli and Savova, Guergana", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.53", doi = "10.18653/v1/2024.clinicalnlp-1.53", pages = "557--569", abstract = "The 2024 Shared Task on Chemotherapy Treatment Timeline Extraction aims to advance the state of the art of clinical event timeline extraction from the Electronic Health Records (EHRs). Specifically, this edition focuses on chemotherapy event timelines from EHRs of patients with breast, ovarian and skin cancers. These patient-level timelines present a novel challenge which involves tasks such as the extraction of relevant events, time expressions and temporal relations from each document and then summarizing over the documents. De-identified EHRs for 57,530 patients with breast and ovarian cancer spanning 2004-2020, and approximately 15,946 patients with melanoma spanning 2010-2020 were made available to participants after executing a Data Use Agreement. A subset of patients is annotated for gold entities, time expressions, temporal relations and patient-level timelines. The rest is considered unlabeled data. In Subtask1, gold chemotherapy event mentions and time expressions are provided (along with the EHR notes). Participants are asked to build the patient-level timelines using gold annotations as input. Thus, the subtask seeks to explore the topics of temporal relations extraction and timeline creation if event and time expression input is perfect. In Subtask2, which is the realistic real-world setting, only EHR notes are provided. Thus, the subtask aims at developing an end-to-end system for chemotherapy treatment timeline extraction from patient{'}s EHR notes. There were 18 submissions for Subtask 1 and 9 submissions for Subtask 2. The organizers provided a baseline system. The teams employed a variety of methods including Logistic Regression, TF-IDF, n-grams, transformer models, zero-shot prompting with Large Language Models (LLMs), and instruction tuning. The gap in performance between prompting LLMs and finetuning smaller-sized LMs indicates that for a challenging task such as patient-level chemotherapy timeline extraction, more sophisticated LLMs or prompting techniques are necessary in order to achieve optimal results as finetuing smaller-sized LMs outperforms by a wide margin.", }

The 2024 Shared Task on Chemotherapy Treatment Timeline Extraction aims to advance the state of the art of clinical event timeline extraction from the Electronic Health Records (EHRs). Specifically, this edition focuses on chemotherapy event timelines from EHRs of patients with breast, ovarian and skin cancers. These patient-level timelines present a novel challenge which involves tasks such as the extraction of relevant events, time expressions and temporal relations from each document and then summarizing over the documents. De-identified EHRs for 57,530 patients with breast and ovarian cancer spanning 2004-2020, and approximately 15,946 patients with melanoma spanning 2010-2020 were made available to participants after executing a Data Use Agreement. A subset of patients is annotated for gold entities, time expressions, temporal relations and patient-level timelines. The rest is considered unlabeled data. In Subtask1, gold chemotherapy event mentions and time expressions are provided (along with the EHR notes). Participants are asked to build the patient-level timelines using gold annotations as input. Thus, the subtask seeks to explore the topics of temporal relations extraction and timeline creation if event and time expression input is perfect. In Subtask2, which is the realistic real-world setting, only EHR notes are provided. Thus, the subtask aims at developing an end-to-end system for chemotherapy treatment timeline extraction from patient{'}s EHR notes. There were 18 submissions for Subtask 1 and 9 submissions for Subtask 2. The organizers provided a baseline system. The teams employed a variety of methods including Logistic Regression, TF-IDF, n-grams, transformer models, zero-shot prompting with Large Language Models (LLMs), and instruction tuning. The gap in performance between prompting LLMs and finetuning smaller-sized LMs indicates that for a challenging task such as patient-level chemotherapy timeline extraction, more sophisticated LLMs or prompting techniques are necessary in order to achieve optimal results as finetuing smaller-sized LMs outperforms by a wide margin.

Overview of the 2024 Shared Task on Chemotherapy Treatment Timeline Extraction

clinicalnlp-1.53

Poster

https://aclanthology.org/2024.clinicalnlp-1.54.bib

https://aclanthology.org/2024.clinicalnlp-1.54/

@inproceedings{corbeil-2024-iryonlp, title = "{I}ryo{NLP} at {MEDIQA}-{CORR} 2024: Tackling the Medical Error Detection {\&} Correction Task on the Shoulders of Medical Agents", author = "Corbeil, Jean-Philippe", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.54", doi = "10.18653/v1/2024.clinicalnlp-1.54", pages = "570--580", abstract = "In natural language processing applied to the clinical domain, utilizing large language models has emerged as a promising avenue for error detection and correction on clinical notes, a knowledge-intensive task for which annotated data is scarce. This paper presents MedReAct{'}N{'}MedReFlex, which leverages a suite of four LLM-based medical agents. The MedReAct agent initiates the process by observing, analyzing, and taking action, generating trajectories to guide the search to target a potential error in the clinical notes. Subsequently, the MedEval agent employs five evaluators to assess the targeted error and the proposed correction. In cases where MedReAct{'}s actions prove insufficient, the MedReFlex agent intervenes, engaging in reflective analysis and proposing alternative strategies. Finally, the MedFinalParser agent formats the final output, preserving the original style while ensuring the integrity of the error correction process. One core component of our method is our RAG pipeline based on our ClinicalCorp corpora. Among other well-known sources containing clinical guidelines and information, we preprocess and release the open-source MedWiki dataset for clinical RAG application. Our results demonstrate the central role of our RAG approach with ClinicalCorp leveraged through the MedReAct{'}N{'}MedReFlex framework. It achieved the ninth rank on the MEDIQA-CORR 2024 final leaderboard.", }

In natural language processing applied to the clinical domain, utilizing large language models has emerged as a promising avenue for error detection and correction on clinical notes, a knowledge-intensive task for which annotated data is scarce. This paper presents MedReAct{'}N{'}MedReFlex, which leverages a suite of four LLM-based medical agents. The MedReAct agent initiates the process by observing, analyzing, and taking action, generating trajectories to guide the search to target a potential error in the clinical notes. Subsequently, the MedEval agent employs five evaluators to assess the targeted error and the proposed correction. In cases where MedReAct{'}s actions prove insufficient, the MedReFlex agent intervenes, engaging in reflective analysis and proposing alternative strategies. Finally, the MedFinalParser agent formats the final output, preserving the original style while ensuring the integrity of the error correction process. One core component of our method is our RAG pipeline based on our ClinicalCorp corpora. Among other well-known sources containing clinical guidelines and information, we preprocess and release the open-source MedWiki dataset for clinical RAG application. Our results demonstrate the central role of our RAG approach with ClinicalCorp leveraged through the MedReAct{'}N{'}MedReFlex framework. It achieved the ninth rank on the MEDIQA-CORR 2024 final leaderboard.

IryoNLP at MEDIQA-CORR 2024: Tackling the Medical Error Detection & Correction Task on the Shoulders of Medical Agents

clinicalnlp-1.54

Poster

2404.15488

https://aclanthology.org/2024.clinicalnlp-1.55.bib

https://aclanthology.org/2024.clinicalnlp-1.55/

@inproceedings{yim-etal-2024-overview, title = "Overview of the {MEDIQA}-{M}3{G} 2024 Shared Task on Multilingual Multimodal Medical Answer Generation", author = "Yim, Wen-wai and Ben Abacha, Asma and Fu, Yujuan and Sun, Zhaoyi and Xia, Fei and Yetisgen, Meliha and Krallinger, Martin", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.55", doi = "10.18653/v1/2024.clinicalnlp-1.55", pages = "581--589", abstract = "Remote patient care provides opportunities for expanding medical access, saving healthcare costs, and offering on-demand convenient services. In the MEDIQA-M3G 2024 Shared Task, researchers explored solutions for the specific task of dermatological consumer health visual question answering, where user generated queries and images are used as input and a free-text answer response is generated as output. In this novel challenge, eight teams with a total of 48 submissions were evaluated across three language test sets. In this work, we provide a summary of the dataset, as well as results and approaches. We hope that the insights learned here will inspire future research directions that can lead to technology that deburdens clinical workload and improves care.", }

Remote patient care provides opportunities for expanding medical access, saving healthcare costs, and offering on-demand convenient services. In the MEDIQA-M3G 2024 Shared Task, researchers explored solutions for the specific task of dermatological consumer health visual question answering, where user generated queries and images are used as input and a free-text answer response is generated as output. In this novel challenge, eight teams with a total of 48 submissions were evaluated across three language test sets. In this work, we provide a summary of the dataset, as well as results and approaches. We hope that the insights learned here will inspire future research directions that can lead to technology that deburdens clinical workload and improves care.

Overview of the MEDIQA-M3G 2024 Shared Task on Multilingual Multimodal Medical Answer Generation

clinicalnlp-1.55

Poster

https://aclanthology.org/2024.clinicalnlp-1.56.bib

https://aclanthology.org/2024.clinicalnlp-1.56/

@inproceedings{rajwal-etal-2024-em, title = "{EM}{\_}{M}ixers at {MEDIQA}-{CORR} 2024: Knowledge-Enhanced Few-Shot In-Context Learning for Medical Error Detection and Correction", author = "Rajwal, Swati and Agichtein, Eugene and Sarker, Abeed", editor = "Naumann, Tristan and Ben Abacha, Asma and Bethard, Steven and Roberts, Kirk and Bitterman, Danielle", booktitle = "Proceedings of the 6th Clinical Natural Language Processing Workshop", month = jun, year = "2024", address = "Mexico City, Mexico", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.clinicalnlp-1.56", doi = "10.18653/v1/2024.clinicalnlp-1.56", pages = "590--595", abstract = "This paper describes our submission to MEDIQA-CORR 2024 shared task for automatic identification and correction of medical errors in a given clinical text. We report results from two approaches: the first uses a few-shot in-context learning (ICL) with a Large Language Model (LLM) and the second approach extends the idea by using a knowledge-enhanced few-shot ICL approach. We used Azure OpenAI GPT-4 API as the LLM and Wikipedia as the external knowledge source. We report evaluation metrics (accuracy, ROUGE, BERTScore, BLEURT) across both approaches for validation and test datasets. Of the two approaches implemented, our experimental results show that the knowledge-enhanced few-shot ICL approach with GPT-4 performed better with error flag (subtask A) and error sentence detection (subtask B) with accuracies of 68{\%} and 64{\%}, respectively on the test dataset. These results positioned us fourth in subtask A and second in subtask B, respectively in the shared task.", }

This paper describes our submission to MEDIQA-CORR 2024 shared task for automatic identification and correction of medical errors in a given clinical text. We report results from two approaches: the first uses a few-shot in-context learning (ICL) with a Large Language Model (LLM) and the second approach extends the idea by using a knowledge-enhanced few-shot ICL approach. We used Azure OpenAI GPT-4 API as the LLM and Wikipedia as the external knowledge source. We report evaluation metrics (accuracy, ROUGE, BERTScore, BLEURT) across both approaches for validation and test datasets. Of the two approaches implemented, our experimental results show that the knowledge-enhanced few-shot ICL approach with GPT-4 performed better with error flag (subtask A) and error sentence detection (subtask B) with accuracies of 68{\%} and 64{\%}, respectively on the test dataset. These results positioned us fourth in subtask A and second in subtask B, respectively in the shared task.

EM_Mixers at MEDIQA-CORR 2024: Knowledge-Enhanced Few-Shot In-Context Learning for Medical Error Detection and Correction

clinicalnlp-1.56

Poster