Understanding and Mitigating the Limitations of Prioritized Experience Replay
This work addresses theoretical gaps in prioritized sampling for reinforcement learning, offering incremental improvements for researchers and practitioners in control and autonomous systems.
The paper tackled the lack of theoretical understanding of Prioritized Experience Replay (ER) by analyzing its limitations, such as outdated priorities and insufficient coverage, and proposed a model-based stochastic gradient Langevin dynamics sampling method to mitigate these issues, showing improved performance in discrete and continuous control problems and an autonomous driving application.
Prioritized Experience Replay (ER) has been empirically shown to improve sample efficiency across many domains and attracted great attention; however, there is little theoretical understanding of why such prioritized sampling helps and its limitations. In this work, we take a deep look at the prioritized ER. In a supervised learning setting, we show the equivalence between the error-based prioritized sampling method for mean squared error and uniform sampling for cubic power loss. We then provide theoretical insight into why it improves convergence rate upon uniform sampling during early learning. Based on the insight, we further point out two limitations of the prioritized ER method: 1) outdated priorities and 2) insufficient coverage of the sample space. To mitigate the limitations, we propose our model-based stochastic gradient Langevin dynamics sampling method. We show that our method does provide states distributed close to an ideal prioritized sampling distribution estimated by the brute-force method, which does not suffer from the two limitations. We conduct experiments on both discrete and continuous control problems to show our approach's efficacy and examine the practical implication of our method in an autonomous driving application.