Fidelity-Induced Interpretable Policy Extraction for Reinforcement Learning
This addresses the issue of opaque decision-making in DRL agents for users needing trust and scrutiny, though it is incremental as it builds on existing interpretable policy extraction methods.
The paper tackled the problem of interpretable policy extraction in deep reinforcement learning by proposing Fidelity-Induced Policy Extraction (FIPE), which improved consistency with the agent's behavior and outperformed baselines in StarCraft II experiments.
Deep Reinforcement Learning (DRL) has achieved remarkable success in sequential decision-making problems. However, existing DRL agents make decisions in an opaque fashion, hindering the user from establishing trust and scrutinizing weaknesses of the agents. While recent research has developed Interpretable Policy Extraction (IPE) methods for explaining how an agent takes actions, their explanations are often inconsistent with the agent's behavior and thus, frequently fail to explain. To tackle this issue, we propose a novel method, Fidelity-Induced Policy Extraction (FIPE). Specifically, we start by analyzing the optimization mechanism of existing IPE methods, elaborating on the issue of ignoring consistency while increasing cumulative rewards. We then design a fidelity-induced mechanism by integrate a fidelity measurement into the reinforcement learning feedback. We conduct experiments in the complex control environment of StarCraft II, an arena typically avoided by current IPE methods. The experiment results demonstrate that FIPE outperforms the baselines in terms of interaction performance and consistency, meanwhile easy to understand.