t-Soft Update of Target Network for Deep Reinforcement Learning

TL;DR

This paper introduces a t-soft update rule for target networks in deep reinforcement learning, inspired by student-t distribution, to improve robustness and learning efficiency over traditional exponential moving average methods.

Contribution

The paper proposes a novel t-soft update method for target networks in DRL, which adaptively excludes extreme updates and enhances learning stability compared to conventional methods.

Findings

Outperforms traditional methods in PyBullet robotics simulations.

Automatically excludes extreme updates due to heavy-tailed property.

Maintains learning speed while improving robustness.

Abstract

This paper proposes a new robust update rule of target network for deep reinforcement learning (DRL), to replace the conventional update rule, given as an exponential moving average. The target network is for smoothly generating the reference signals for a main network in DRL, thereby reducing learning variance. The problem with its conventional update rule is the fact that all the parameters are smoothly copied with the same speed from the main network, even when some of them are trying to update toward the wrong directions. This behavior increases the risk of generating the wrong reference signals. Although slowing down the overall update speed is a naive way to mitigate wrong updates, it would decrease learning speed. To robustly update the parameters while keeping learning speed, a t-soft update method, which is inspired by student-t distribution, is derived with reference to the analogy between the exponential moving average and the normal distribution. Through the analysis of the derived t-soft update, we show that it takes over the properties of the student-t distribution. Specifically, with a heavy-tailed property of the student-t distribution, the t-soft update automatically excludes extreme updates that differ from past experiences. In addition, when the updates are similar to the past experiences, it can mitigate the learning delay by increasing the amount of updates. In PyBullet robotics simulations for DRL, an online actor-critic algorithm with the t-soft update outperformed the conventional methods in terms of the obtained return and/or its variance. From the training process by the t-soft update, we found that the t-soft update is globally consistent with the standard soft update, and the update rates are locally adjusted for acceleration or suppression.