EvoClinician: A Self-Evolving Agent for Multi-Turn Medical Diagnosis via Test-Time Evolutionary Learning
🤖 AI Summary
This work addresses the limitation of existing medical AI systems, which typically rely on static, complete patient records and fail to emulate the dynamic, iterative diagnostic process clinicians employ through sequential questioning and examinations. To bridge this gap, the authors introduce the Med-Inquire benchmark and EvoClinician, an intelligent agent featuring a novel test-time evolutionary learning mechanism. EvoClinician operates via a “diagnose–score–evolve” loop, dynamically refining its diagnostic strategy through interaction. It employs a multi-agent architecture comprising an Actor for diagnosis, a Process Grader to evaluate action utility, and an Evolver that iteratively updates the strategy using prompts and memory. Evaluated in a simulation environment built from real clinical cases, EvoClinician significantly outperforms continual learning and other self-evolution baselines in both diagnostic accuracy and resource efficiency, overcoming the constraints of conventional static models.
Application Category
📝 Abstract
Prevailing medical AI operates on an unrealistic''one-shot''model, diagnosing from a complete patient file. However, real-world diagnosis is an iterative inquiry where Clinicians sequentially ask questions and order tests to strategically gather information while managing cost and time. To address this, we first propose Med-Inquire, a new benchmark designed to evaluate an agent's ability to perform multi-turn diagnosis. Built upon a dataset of real-world clinical cases, Med-Inquire simulates the diagnostic process by hiding a complete patient file behind specialized Patient and Examination agents. They force the agent to proactively ask questions and order tests to gather information piece by piece. To tackle the challenges posed by Med-Inquire, we then introduce EvoClinician, a self-evolving agent that learns efficient diagnostic strategies at test time. Its core is a''Diagnose-Grade-Evolve''loop: an Actor agent attempts a diagnosis; a Process Grader agent performs credit assignment by evaluating each action for both clinical yield and resource efficiency; finally, an Evolver agent uses this feedback to update the Actor's strategy by evolving its prompt and memory. Our experiments show EvoClinician outperforms continual learning baselines and other self-evolving agents like memory agents. The code is available at https://github.com/yf-he/EvoClinician
Problem
Research questions and friction points this paper is trying to address.
multi-turn medical diagnosis
test-time learning
clinical decision-making
information gathering
diagnostic reasoning
Innovation
Methods, ideas, or system contributions that make the work stand out.
test-time evolution
multi-turn diagnosis
self-evolving agent
diagnostic reasoning
medical AI
🔎 Similar Papers
No similar papers found.
