Learning State Representations via Retracing in Reinforcement Learning
arXiv:2111.12600
Abstract
We propose learning via retracing, a novel self-supervised approach for learning the state representation (and the associated dynamics model) for reinforcement learning tasks. In addition to the predictive (reconstruction) supervision in the forward direction, we propose to include "retraced" transitions for representation / model learning, by enforcing the cycle-consistency constraint between the original and retraced states, hence improve upon the sample efficiency of learning. Moreover, learning via retracing explicitly propagates information about future transitions backward for inferring previous states, thus facilitates stronger representation learning for the downstream reinforcement learning tasks. We introduce Cycle-Consistency World Model (CCWM), a concrete model-based instantiation of learning via retracing. Additionally we propose a novel adaptive "truncation" mechanism for counteracting the negative impacts brought by "irreversible" transitions such that learning via retracing can be maximally effective. Through extensive empirical studies on visual-based continuous control benchmarks, we demonstrate that CCWM achieves state-of-the-art performance in terms of sample efficiency and asymptotic performance, whilst exhibiting behaviours that are indicative of stronger representation learning.
References in corpus (11)
- Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
- Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
- High-Dimensional Continuous Control Using Generalized Advantage Estimation
- StarCraft II: A New Challenge for Reinforcement Learning
- Reinforcement Learning with Augmented Data
- Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixels
- Learning Invariant Representations for Reinforcement Learning without Reconstruction
- Mastering Visual Continuous Control: Improved Data-Augmented Reinforcement Learning
- Learning and Querying Fast Generative Models for Reinforcement Learning
- PlayVirtual: Augmenting Cycle-Consistent Virtual Trajectories for Reinforcement Learning
- Weakly-Supervised Reinforcement Learning for Controllable Behavior