Which Experiences Are Influential for RL Agents? Efficiently Estimating The Influence of Experiences
This addresses a computational bottleneck for RL practitioners by enabling efficient identification and correction of negative influences in experience replay, though it is incremental as it builds on existing influence estimation concepts.
The paper tackles the problem of efficiently estimating which experiences in a replay buffer influence reinforcement learning agents' performance, presenting PIToD as a method that significantly improves agents' performance by amending them based on these estimates.
In reinforcement learning (RL) with experience replay, experiences stored in a replay buffer influence the RL agent's performance. Information about how these experiences influence the agent's performance is valuable for various purposes, such as identifying experiences that negatively influence underperforming agents. One method for estimating the influence of experiences is the leave-one-out (LOO) method. However, this method is usually computationally prohibitive. In this paper, we present Policy Iteration with Turn-over Dropout (PIToD), which efficiently estimates the influence of experiences. We evaluate how correctly PIToD estimates the influence of experiences and its efficiency compared to LOO. We then apply PIToD to amend underperforming RL agents, i.e., we use PIToD to estimate negatively influential experiences for the RL agents and to delete the influence of these experiences. We show that RL agents' performance is significantly improved via amendments with PIToD.