Learning Reinforced Agents with Counterfactual Simulation for Medical Automatic Diagnosis.
Junfan Lin, Ziliang Chen, Xiaodan Liang, Keze Wang, Liang Lin · arXiv (Cornell University) · 2020
Medical automatic diagnosis (MAD) aims to learn an agent that mimics the behavior of human doctors, i.e. inquiring symptoms and informing diseases. Due to medical ethics concerns, it is impractical to directly apply reinforcement learning techniques to MAD, e.g., training a reinforced agent with human patients. Developing a patient simulator by using the collected patient-doctor dialogue records has been proposed as a promising workaround to MAD. However, most of these existing works overlook the causal relationship between patient symptoms and diseases. For example, these simulators simply generate the not-sure response to the symptom inquiry if the symptom was not observed in the dialogue record. Consequently, the MAD agent is usually trained without exploiting the counterfactual reasoning beyond the factual observations. To address this problem, this paper presents a propensity-based patient simulator (PBPS), which is capable of facilitating the training of MAD agents by generating informative counterfactual answers along with the disease diagnosis. Specifically, our PBPS estimates the propensity score of each record with the patient-doctor dialogue reasoning, and can thus generate the counterfactual answers by searching across records. That is, the unrecorded symptom for one patient can be found in the records of other patients according to the propensity score matching. The informative and causal-aware responses from PBPS are beneficial for modeling diagnostic confidence. To this end, we also propose a progressive assurance agent~(P2A) trained with PBPS, which includes two separate yet cooperative branches accounting for the execution of symptom-inquiry and disease-diagnosis actions, respectively.