MEDSAGE: Enhancing Robustness of Medical Dialogue Summarization to ASR Errors with LLM-generated Synthetic Dialogues

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Binici, Kuluhan, Kashyap, Abhinav Ramesh, Schlegel, Viktor, Liu, Andy T., Dwivedi, Vijay Prakash, Nguyen, Thanh-Tung, Gao, Xiaoxue, Chen, Nancy F., Winkler, Stefan
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917886964006912
author Binici, Kuluhan
Kashyap, Abhinav Ramesh
Schlegel, Viktor
Liu, Andy T.
Dwivedi, Vijay Prakash
Nguyen, Thanh-Tung
Gao, Xiaoxue
Chen, Nancy F.
Winkler, Stefan
author_facet Binici, Kuluhan
Kashyap, Abhinav Ramesh
Schlegel, Viktor
Liu, Andy T.
Dwivedi, Vijay Prakash
Nguyen, Thanh-Tung
Gao, Xiaoxue
Chen, Nancy F.
Winkler, Stefan
contents Automatic Speech Recognition (ASR) systems are pivotal in transcribing speech into text, yet the errors they introduce can significantly degrade the performance of downstream tasks like summarization. This issue is particularly pronounced in clinical dialogue summarization, a low-resource domain where supervised data for fine-tuning is scarce, necessitating the use of ASR models as black-box solutions. Employing conventional data augmentation for enhancing the noise robustness of summarization models is not feasible either due to the unavailability of sufficient medical dialogue audio recordings and corresponding ASR transcripts. To address this challenge, we propose MEDSAGE, an approach for generating synthetic samples for data augmentation using Large Language Models (LLMs). Specifically, we leverage the in-context learning capabilities of LLMs and instruct them to generate ASR-like errors based on a few available medical dialogue examples with audio recordings. Experimental results show that LLMs can effectively model ASR noise, and incorporating this noisy data into the training process significantly improves the robustness and accuracy of medical dialogue summarization systems. This approach addresses the challenges of noisy ASR outputs in critical applications, offering a robust solution to enhance the reliability of clinical dialogue summarization.
format Preprint
id arxiv_https___arxiv_org_abs_2408_14418
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MEDSAGE: Enhancing Robustness of Medical Dialogue Summarization to ASR Errors with LLM-generated Synthetic Dialogues
Binici, Kuluhan
Kashyap, Abhinav Ramesh
Schlegel, Viktor
Liu, Andy T.
Dwivedi, Vijay Prakash
Nguyen, Thanh-Tung
Gao, Xiaoxue
Chen, Nancy F.
Winkler, Stefan
Computation and Language
Artificial Intelligence
Automatic Speech Recognition (ASR) systems are pivotal in transcribing speech into text, yet the errors they introduce can significantly degrade the performance of downstream tasks like summarization. This issue is particularly pronounced in clinical dialogue summarization, a low-resource domain where supervised data for fine-tuning is scarce, necessitating the use of ASR models as black-box solutions. Employing conventional data augmentation for enhancing the noise robustness of summarization models is not feasible either due to the unavailability of sufficient medical dialogue audio recordings and corresponding ASR transcripts. To address this challenge, we propose MEDSAGE, an approach for generating synthetic samples for data augmentation using Large Language Models (LLMs). Specifically, we leverage the in-context learning capabilities of LLMs and instruct them to generate ASR-like errors based on a few available medical dialogue examples with audio recordings. Experimental results show that LLMs can effectively model ASR noise, and incorporating this noisy data into the training process significantly improves the robustness and accuracy of medical dialogue summarization systems. This approach addresses the challenges of noisy ASR outputs in critical applications, offering a robust solution to enhance the reliability of clinical dialogue summarization.
title MEDSAGE: Enhancing Robustness of Medical Dialogue Summarization to ASR Errors with LLM-generated Synthetic Dialogues
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2408.14418