DeepMind 用模拟患者 RL 训练 Gemini 精神科问诊
New Google DeepMind paper shows that training a model with RL on simulated patients made it a better psychiatric intake interviewer, as judged by psychiatrists.
Google DeepMind 用模拟患者对 Gemini 3.1 Flash-Lite 做多轮 RL 训练，在 60 组配对病例中有 76.7% 被精神科医生盲评判定优于同提示词的基座模型。在 256 个对抗性病例中，急性风险未升级率从 30.5% 降至 12.9%，即约每 8 例仍有 1 例危机未被上报。32 名精神科医生扮演患者，另 15 名医生盲评对话记录。
正文
New Google DeepMind paper shows that training a model with RL on simulated patients made it a better psychiatric intake interviewer, as judged by psychiatrists.
training on simulated psychiatric intakes cut missed escalations of acute risk from 30.5% to 12.9% in adversarial cases.
In psychiatry, the intake interview is a main diagnostic tool, and key facts often surface only through trust and careful questions. Most conversational medical AI evaluations have left psychiatric cases out.
They trained Gemini 3.1 Flash-Lite with multi-turn RL against simulated patients who reveal key facts only after rapport and targeted questions. Then 32 psychiatrists role-played patients, while 15 others graded the transcripts blind.
The trained agent beat its identically prompted base model in 76.7% of 60 matched case pairs. On 256 adversarial cases, it failed to escalate acute risks 12.9% of the time, down from 30.5%.
If you build health agents, train them on simulated conversations, but keep clinicians supervising, since about 1 in 8 adversarial crises went unescalated.
– arxiv. org/abs/2610.40476
Title: "Training language models in simulation for psychiatric intake"