EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Think Like a Doctor: Conversational Diagnosis through the Exploration of Diagnostic Knowledge Graphs

Jeongmoon Won, Seungwon Kook, Yohan Jo

Transactions of the Association for Computational Linguistics · 05.11.2026 · 14 · 2794-2818

DOI: 10.1162/tacl.a.826 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Transactions of the Association for Computational Linguistics, online erschienen am 05.11.2026. Verfasst von Jeongmoon Won und 2 weiteren Autor:innen. Die Arbeit nennt 39 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Conversational diagnosis requires multiturn history-taking, where an agent asks clarifying questions to refine differential diagnoses under incomplete information. Existing approaches often rely on the parametric knowledge of a model or the unrealistic assumption that patients provide rich and concrete information. To address these limitations, we propose KGCQ, a conversational diagnosis system that explores a diagnostic knowledge graph to reason in three steps: (i) generating diagnostic hypotheses from the dialogue context, (ii) extracting a subgraph for diagnostic evidence, and (iii) verifying hypotheses through clarifying questions, which are repeated until a final diagnosis is reached. Since evaluating the system requires a realistic patient simulator that responds to the system’s questions, we adopt PatientSim, a persona-driven patient simulator, together with patient profiles from MIMIC-IV. We further adapt it with low-specificity symptom reporting to reflect how real-world patients describe symptoms vaguely during early clinical encounters. Experiments show improved diagnostic accuracy and efficiency over strong baselines, and physician evaluations support the realism of our simulator and the clinical utility of the generated clarifying questions. Our code is available at https://github.com/holi-lab/KGCQ.

Zitieren

APA

Won, J., Kook, S., & Jo, Y. (2026). Think Like a Doctor: Conversational Diagnosis through the Exploration of Diagnostic Knowledge Graphs. Transactions of the Association for Computational Linguistics, 14, 2794-2818. https://doi.org/10.1162/tacl.a.826

Vancouver

Won J, Kook S, Jo Y. Think Like a Doctor: Conversational Diagnosis through the Exploration of Diagnostic Knowledge Graphs. Transactions of the Association for Computational Linguistics. 2026;14:2794-2818. doi:10.1162/tacl.a.826

BibTeX
@article{Won2026,
  author = {Won, Jeongmoon and Kook, Seungwon and Jo, Yohan},
  title = {Think Like a Doctor: Conversational Diagnosis through the Exploration of Diagnostic Knowledge Graphs},
  journal = {Transactions of the Association for Computational Linguistics},
  year = {2026},
  volume = {14},
  pages = {2794--2818},
  doi = {10.1162/tacl.a.826}
}
Kostenlos starten →