New Method Cuts LLM Use in ICD-10 Code Prediction via Supervised Candidate Generation
Researchers propose a supervised candidate generation approach with selective LLM invocation to predict ICD-10 codes from short clinical diagnosis text, reducing calls to large language models.
Read the full story at arXiv cs.CL ↗
Timeline · 1 report
- 2026-11-18 05:00 · arXiv cs.CL
Call-Efficient ICD-10 Code Prediction via Supervised Candidate Generation and Selective LLM Invocation