Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Reasoning Like a Doctor: Improving Medical Dialogue Systems via Diagnostic Reasoning Process Alignment

Kaishuai Xu, Yi Cheng, Wenjun Hou, Qiaoyu Tan, Wenjie Li

TL;DR

This work proposes a novel framework, Emulation, designed to generate an appropriate response that relies on abductive and deductive diagnostic reasoning analyses and aligns with clinician preferences through thought process modeling, and furnishes clear explanations for the generated responses.

Abstract

Medical dialogue systems have attracted significant attention for their potential to act as medical assistants. Enabling these medical systems to emulate clinicians' diagnostic reasoning process has been the long-standing research focus. Previous studies rudimentarily realized the simulation of clinicians' diagnostic process by fine-tuning language models on high-quality dialogue datasets. Nonetheless, they overly focus on the outcomes of the clinician's reasoning process while ignoring their internal thought processes and alignment with clinician preferences. Our work aims to build a medical dialogue system that aligns with clinicians' diagnostic reasoning processes. We propose a novel framework, Emulation, designed to generate an appropriate response that relies on abductive and deductive diagnostic reasoning analyses and aligns with clinician preferences through thought process modeling. Experimental results on two datasets confirm the efficacy of Emulation. Crucially, our framework furnishes clear explanations for the generated responses, enhancing its transparency in medical consultations.

Reasoning Like a Doctor: Improving Medical Dialogue Systems via Diagnostic Reasoning Process Alignment

TL;DR

This work proposes a novel framework, Emulation, designed to generate an appropriate response that relies on abductive and deductive diagnostic reasoning analyses and aligns with clinician preferences through thought process modeling, and furnishes clear explanations for the generated responses.

Abstract

Medical dialogue systems have attracted significant attention for their potential to act as medical assistants. Enabling these medical systems to emulate clinicians' diagnostic reasoning process has been the long-standing research focus. Previous studies rudimentarily realized the simulation of clinicians' diagnostic process by fine-tuning language models on high-quality dialogue datasets. Nonetheless, they overly focus on the outcomes of the clinician's reasoning process while ignoring their internal thought processes and alignment with clinician preferences. Our work aims to build a medical dialogue system that aligns with clinicians' diagnostic reasoning processes. We propose a novel framework, Emulation, designed to generate an appropriate response that relies on abductive and deductive diagnostic reasoning analyses and aligns with clinician preferences through thought process modeling. Experimental results on two datasets confirm the efficacy of Emulation. Crucially, our framework furnishes clear explanations for the generated responses, enhancing its transparency in medical consultations.
Paper Structure (38 sections, 5 equations, 14 figures, 7 tables)

This paper contains 38 sections, 5 equations, 14 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Preliminary
  3. Problem Formulation
  4. Diagnostic Reasoning
  5. Disease Annotation
  6. Method
  7. Abductive Reasoner
  8. Disease Retrieval.
  9. Diagnosis Refinement.
  10. Deductive Reasoner
  11. Thought Alignment
  12. Disease Priority Alignment
  13. Thought Process Alignment
  14. Thought Process Corpus
  15. Experiments
  16. ...and 23 more sections

Figures (14)

  • Figure 1: An example of a medical dialogue. IBS is the abbreviation for irritable bowel syndrome.
  • Figure 2: Illustration of our Framework Emulation. The symbols of a snowflake or a flame represent the module mainly operating in a prompt-based way or having undergone fine-tuning, respectively.
  • Figure 3: Human evaluation results of baseline methods. Red is for Win, yellow for Tie, and green for Lose.
  • Figure 4: The top-5 diagnosis results with or without disease priority alignment.
  • Figure 5: The top-5 diagnosis results of Emulation at different conversation turns.
  • ...and 9 more figures