New LLM Framework CARing Enhances Medical Diagnosis Prediction

Researchers have developed CARing, a novel framework designed to improve the accuracy of large language models (LLMs) in predicting diagnoses from clinical data. CARing addresses two key challenges: the tendency for LLMs to focus on a few correct diagnoses and the fragmentation of medical codes into less meaningful tokens. The framework uses compositional Semantic IDs (SIDs) to represent diagnoses and employs a coverage reward mechanism in reinforcement learning to ensure a broader range of potential diagnoses are considered. Tested on the MIMIC-III and MIMIC-IV datasets, CARing demonstrated superior performance in weighted F1 scores and top-k recall compared to existing baselines. AI

IMPACT This research could lead to more accurate and comprehensive diagnostic tools for healthcare professionals, improving patient care through better utilization of clinical data.

RANK_REASON The cluster contains a research paper detailing a new framework for LLM-based medical diagnosis prediction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• Hugging Face
• ICD codes
• Large language models (LLMs)
• MIMIC-III
• MIMIC-IV
• Semantic IDs (SIDs)

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →