Extracting Expert's Goals by What-if Interpretable Modeling
This work addresses the challenge of applying reinforcement learning in real-world domains like healthcare where reward specification and exploration are limited, offering an interpretable solution for clinicians.
The paper tackled the problem of recovering clinicians' rewards in healthcare settings where specifying rewards and exploring treatments is difficult, by using what-if reasoning and generalized additive models (GAMs) to outperform baselines in simulations and real-world hospital data, with explanations matching clinical guidelines unlike linear models.
Although reinforcement learning (RL) has tremendous success in many fields, applying RL to real-world settings such as healthcare is challenging when the reward is hard to specify and no exploration is allowed. In this work, we focus on recovering clinicians' rewards in treating patients. We incorporate the what-if reasoning to explain the clinician's treatments based on their potential future outcomes. We use generalized additive models (GAMs) - a class of accurate, interpretable models - to recover the reward. In both simulation and a real-world hospital dataset, we show our model outperforms baselines. Finally, our model's explanations match several clinical guidelines when treating patients while we found the commonly-used linear model often contradicts them.