Interactive Explanations: Diagnosis and Repair of Reinforcement Learning Based Agent Behaviors
This addresses the challenge of making RL agents more transparent and user-controllable, which is incremental but practical for improving human-AI interaction in specific domains like gaming.
The paper tackles the problem of aligning reinforcement learning agent behaviors with user preferences by introducing an interactive explanation method using natural language templates, enabling two-way communication for diagnosis and repair, and demonstrates its effectiveness in a Super Mario Bros. clone.
Reinforcement learning techniques successfully generate convincing agent behaviors, but it is still difficult to tailor the behavior to align with a user's specific preferences. What is missing is a communication method for the system to explain the behavior and for the user to repair it. In this paper, we present a novel interaction method that uses interactive explanations using templates of natural language as a communication method. The main advantage of this interaction method is that it enables a two-way communication channel between users and the agent; the bot can explain its thinking procedure to the users, and the users can communicate their behavior preferences to the bot using the same interactive explanations. In this manner, the thinking procedure of the bot is transparent, and users can provide corrections to the bot that include a suggested action to take, a goal to achieve, and the reasons behind these decisions. We tested our proposed method in a clone of the video game named \textit{Super Mario Bros.}, and the results demonstrate that our interactive explanation approach is effective at diagnosing and repairing bot behaviors.