Double-blind OSCE: simulation-trained agent wins 76.7% of psychiatrist pairs
SRSchmidgall · x · 2026-10-31
In a double-blind OSCE with psychiatrists role-playing patients, the simulation-trained agent beat its base model in 76.7% of pairs: clinical examination 3.04 vs 2.32, differential diagnosis accuracy 68.3% vs 33.3%, and communication 82.0% vs 65.2%.
Figures show the full pipeline (text consult → questionnaires → psychiatrist evaluation), with the trained model leading on management-plan appropriateness, avoiding inappropriate treatments, and patient-centered communication (adherence 82% vs 65%).
More from Research
- NVIDIA Nemotron 3.5 scales LLM self-verification, +9-10 points on agentic benchmarks — sarahcat21 · 2026-10-31
- Unofficial OpenTargets MCP Server Brings Gene-Drug-Disease Data to LLMs — modelcontextprotocol · 2026-10-31
- Researchers find 5 personality structures in LLM weight space, Big Five-like — luke_drago_ · 2026-10-31
- NVS benchmarks are misaligned: proper pose alignment puts 3D Gaussians back at SOTA — kwangmoo_yi · 2026-10-31
- arXiv moderator: submission surge causes months-long holds, new moderators wanted — tdietterich · 2026-10-31
- arXiv policy checks reject class projects, proposals, surveys and position papers — tdietterich · 2026-10-31