MidMed: Towards Mixed-Type Dialogues for Medical Consultation
Xiaoming Shi, Zeming Liu, Chuan Wang, Haitao Leng, Kui Xue, Xiaofan Zhang, Shaoting Zhang
Abstract
Most medical dialogue systems assume that patients have clear goals (seeking a diagnosis, medicine querying, etc.) before medical consultation. However, in many real situations, due to the lack of medical knowledge, it is usually difficult for patients to determine clear goals with all necessary slots. In this paper, we identify this challenge as how to construct medical consultation dialogue systems to help patients clarify their goals. For further study, we create a novel human-to-human mixed-type medical consultation dialogue corpus, termed MidMed, covering four dialogue types: task-oriented dialogue for diagnosis, recommendation, QA, and chitchat. MidMed covers four departments (otorhinolaryngology, ophthalmology, skin, and digestive system), with 8,309 dialogues. Furthermore, we build benchmarking baselines on MidMed and propose an instruction-guiding medical dialogue generation framework, termed InsMed, to handle mixed-type dialogues. Experimental results show the effectiveness of InsMed.
BibTeX
@inproceedings{shi-etal-2023-midmed,
title = "{M}id{M}ed: Towards Mixed-Type Dialogues for Medical Consultation",
author = "Shi, Xiaoming and
Liu, Zeming and
Wang, Chuan and
Leng, Haitao and
Xue, Kui and
Zhang, Xiaofan and
Zhang, Shaoting",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.453/",
doi = "10.18653/v1/2023.acl-long.453",
pages = "8145--8157"
}