Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

LLM-Guided Evolution for Medical Decision Pipelines

About

Adapting large language models (LLMs) to clinical workflows often requires costly fine-tuning or manual prompt and pipeline engineering. We study LLM-guided MAP-Elites evolution as an inference-time alternative for discovering medical decision strategies and provide an implementation repository at https://github.com/univanxx/llm_guided_evo_medical. We formulate urgency triage, interactive consultation, and medical image classification as evolutionary searches over executable artifacts optimized by task-specific fitness functions. Across all three settings, evolution improves over manually designed baselines under practical constraints. In triage, evolved programs increase Semigran accuracy from $77.3\%$ to $87.1\%$ and emergency recall from $0.60$ to $0.97$, while improving safety-weighted held-out MIMIC-ESI performance. In interactive consultation, evolved policies improve the accuracy--cost frontier across Llama-3, Qwen-3.5, and Gemma-4 and transfer to held-out iCRAFTMD. In PneumoniaMNIST, prompt-only evolution improves frozen MedGemma VLMs while preserving strict JSON outputs. Qualitative analysis shows that the gains come from interpretable program-level mechanisms, calibrated triage boundaries, targeted evidence acquisition, selective commitment, and finding-oriented visual decision rules, rather than superficial prompt rewording alone.

Ivan Sviridov, Artem Oskin, Ivan Panin, Iaroslav Bespalov, Dmitry Dylov, Ivan Oseledets, Aleksandr Nesterov• 2026

Related benchmarks

TaskDatasetResultRank
Clinical triage evaluationMEDIQ
Accuracy89.5
20
Pneumonia ClassificationPneumoniaMNIST v1 (test)
Accuracy84.46
16
Emergency Severity Index PredictionMIMIC-ESI (test)
Exact Accuracy62
11
Medical Question AnsweringMediQ (test)
Accuracy75.6
10
Medical TriageSemigran (test)
Accuracy87.1
10
Urgency TriageSemigran (test)
Accuracy87.1
6
Medical Triage PredictionLevine vignettes 2023 (held-out)
Accuracy (%)75.1
3
Showing 7 of 7 rows

Other info

Follow for update