Chain-of-Though (CoT) prompting strategies for medical error detection and correction

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Wu, Zhaolong, Hasan, Abul, Wu, Jinge, Kim, Yunsoo, Cheung, Jason P. Y., Zhang, Teng, Wu, Honghan
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866914833688952832
author Wu, Zhaolong
Hasan, Abul
Wu, Jinge
Kim, Yunsoo
Cheung, Jason P. Y.
Zhang, Teng
Wu, Honghan
author_facet Wu, Zhaolong
Hasan, Abul
Wu, Jinge
Kim, Yunsoo
Cheung, Jason P. Y.
Zhang, Teng
Wu, Honghan
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2406_09103
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Chain-of-Though (CoT) prompting strategies for medical error detection and correction
Wu, Zhaolong
Hasan, Abul
Wu, Jinge
Kim, Yunsoo
Cheung, Jason P. Y.
Zhang, Teng
Wu, Honghan
Computation and Language
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.
title Chain-of-Though (CoT) prompting strategies for medical error detection and correction
topic Computation and Language
url https://arxiv.org/abs/2406.09103