Recurrent Value Functions

Abstract

Despite recent successes in Reinforcement Learning, value-based methods often suffer from high variance hindering performance. In this paper, we illustrate this in a continuous control setting where state of the art methods perform poorly whenever sensor noise is introduced. To overcome this issue, we introduce Recurrent Value Functions (RVFs) as an alternative to estimate the value function of a state. We propose to estimate the value function of the current state using the value function of past states visited along the trajectory. Due to the nature of their formulation, RVFs have a natural way of learning an emphasis function that selectively emphasizes important states. First, we establish RVF's asymptotic convergence properties in tabular settings. We then demonstrate their robustness on a partially observable domain and continuous control tasks. Finally, we provide a qualitative interpretation of the learned emphasis function.

References in corpus (10)

• Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
• Playing Atari with Deep Reinforcement Learning
• High-Dimensional Continuous Control Using Generalized Advantage Estimation
• StarCraft II: A New Challenge for Reinforcement Learning
• Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation
• A Dissection of Overfitting and Generalization in Continuous Reinforcement Learning
• Taming the Noise in Reinforcement Learning via Soft Updates
• Emphatic Temporal-Difference Learning
• Q-learning with Nearest Neighbors
• Temporal Regularization in Markov Decision Process

Cited by in corpus (1)

• Preferential Temporal Difference Learning