PACT: Learning Diverse Diagnostic Strategies via Privileged Synthesis and Branch Consensus
2026-06-08 • Computation and Language
Computation and LanguageArtificial Intelligence
AI summaryⓘ
The authors created a new training method called PACT to help AI models perform medical diagnoses better by using different ways of reasoning without mixing them up. They designed a system called DPS that uses full medical records to check the quality of doctor-patient conversations but only lets the AI doctor see what a real patient would. PACT trains separate parts of the model for each reasoning style and regularly combines them to improve overall performance. They also made a Chinese medical diagnosis test to see how well the system works. Their experiments show that PACT outperforms other medical AI methods in both diagnosis accuracy and the consultation process.
Large language modelsMedical diagnosisMulti-paradigm reasoningDialogue synthesisElectronic medical recordsLoRA (Low-Rank Adaptation)Consensus trainingChinese medical benchmarkInteractive consultationTask adaptation
Authors
Gen Li, Yuanze Hu, Zhichao Yang, Qingchen Yu, Jianwei Lv, Yue Guo, Yujing Liu, Faguo Wu, Hongwei Zheng, Xiandong Li, Bo Yuan, Yifan Sun, Zhaoxin Fan
Abstract
Clinical diagnosis requires flexible use of multiple reasoning paradigms under incomplete patient information. Existing LLM-based medical agents show strong medical reasoning ability, but single-paradigm or naively mixed dialogue supervision makes these paradigms difficult to learn without interference. We propose \textbf{PACT} (Periodic Anchor Consensus Training), a framework that couples supervised multi-paradigm dialogue synthesis with consensus-based Branch training. At the data level, \textbf{DPS} (Doctor-Patient-Supervisor) uses complete electronic medical records (EMRs) for quality control while keeping the doctor agent restricted to patient-visible information. This produces validated dialogues under four diagnostic reasoning paradigms without leaking hidden clinical answers. At the training level, PACT trains one paradigm-specific LoRA Branch per paradigm and periodically aggregates Branches into a shared Anchor through sign consensus. We further construct a dynamic multi-turn Chinese medical diagnosis benchmark for interactive consultation. Experiments show that PACT achieves state-of-the-art performance among compared proprietary, medical-specialized, and task-adapted baselines on diagnostic outcome and consultation-process metrics.