Clinical Reasoning
Clinical reasoning, the cognitive process of diagnosing and treating patients, is a focus of intense research leveraging large language models (LLMs). Current efforts concentrate on improving LLMs' ability to handle complex, multi-step diagnoses by integrating knowledge graphs, refining tokenization methods for clinical text, and developing interactive question-answering frameworks that mimic physician-patient interactions. This research aims to enhance the reliability and explainability of AI-driven clinical decision support, ultimately improving diagnostic accuracy and patient care, particularly in resource-constrained settings.
Papers
medIKAL: Integrating Knowledge Graphs as Assistants of LLMs for Enhanced Clinical Diagnosis on EMRs
Mingyi Jia, Junwen Duan, Yan Song, Jianxin Wang
Infusing clinical knowledge into tokenisers for language models
Abul Hasan, Jinge Wu, Quang Ngoc Nguyen, Salomé Andres, Imane Guellil, Huayu Zhang, Arlene Casey, Beatrice Alex, Bruce Guthrie, Honghan Wu