t-Soft Update of Target Network for Deep Reinforcement Learning

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.

arXiv comment11 pages, 7 figures

References in corpus (12)

• Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
• Addressing Function Approximation Error in Actor-Critic Methods
• Searching for Activation Functions
• Layer Normalization
• Hindsight Experience Replay
• Student-t Processes as Alternatives to Gaussian Processes
• ChainerRL: A Deep Reinforcement Learning Library
• rlpyt: A Research Code Base for Deep Reinforcement Learning in PyTorch
• LaProp: Separating Momentum and Adaptivity in Adam
• Adaptive and Multiple Time-scale Eligibility Traces for Online Deep Reinforcement Learning
• Proximal Policy Optimization with Relative Pearson Divergence
• TAdam: A Robust Stochastic Gradient Optimizer

Cited by in corpus (11)

• AdaTerm: Adaptive T-Distribution Estimated Robust Moments for Noise-Robust Stochastic Gradient Optimization
• Optimistic Reinforcement Learning by Forward Kullback-Leibler Divergence Optimization
• L2C2: Locally Lipschitz Continuous Constraint towards Stable and Smooth Reinforcement Learning
• Physics-Guided Reinforcement Learning System for Realistic Vehicle Active Suspension Control
• Proximal Policy Optimization with Relative Pearson Divergence
• Proximal Policy Optimization with Adaptive Threshold for Symmetric Relative Density Ratio
• DRL-Based Dynamic Channel Access and SCLAR Maximization for Networks Under Jamming
• Optimization Algorithm for Feedback and Feedforward Policies towards Robot Control Robust to Sensing Failures
• Design of Restricted Normalizing Flow towards Arbitrary Stochastic Policy with Computational Efficiency
• Consolidated Adaptive T-soft Update for Deep Reinforcement Learning
• Cross Modality 3D Navigation Using Reinforcement Learning and Neural Style Transfer