Benchmark, Toolbox, and Reflection-based Method for Clinical Agent
-
Updated
Nov 6, 2024 - Python
Benchmark, Toolbox, and Reflection-based Method for Clinical Agent
[EMNLP 2026]* GraphMed-LT: Patient-Specific Graph Memory with Latent Clinical Thought Refinement for Multi-Turn Medical Conversations
To associate your repository with the clinical-agent topic, visit your repo's landing page and select "manage topics."