Picture for Jinwen Wang

Jinwen Wang

A Reliable Representation with Bidirectional Transition Model for Visual Reinforcement Learning Generalization

Add code
Dec 04, 2023
Xiaobo Hu, Youfang Lin, Yue Liu, Jinwen Wang, Shuo Wang, Hehe Fan, Kai Lv

Figure 1 for A Reliable Representation with Bidirectional Transition Model for Visual Reinforcement Learning Generalization
Figure 2 for A Reliable Representation with Bidirectional Transition Model for Visual Reinforcement Learning Generalization
Figure 3 for A Reliable Representation with Bidirectional Transition Model for Visual Reinforcement Learning Generalization
Figure 4 for A Reliable Representation with Bidirectional Transition Model for Visual Reinforcement Learning Generalization

Visual reinforcement learning has proven effective in solving control tasks with high-dimensional observations. However, extracting reliable and generalizable representations from vision-based observations remains a central challenge. Inspired by the human thought process, when the representation extracted from the observation can predict the future and trace history, the representation is reliable and accurate in comprehending the environment. Based on this concept, we introduce a Bidirectional Transition (BiT) model, which leverages the ability to bidirectionally predict environmental transitions both forward and backward to extract reliable representations. Our model demonstrates competitive generalization performance and sample efficiency on two settings of the DeepMind Control suite. Additionally, we utilize robotic manipulation and CARLA simulators to demonstrate the wide applicability of our method.

Viaarxiv icon
Github IconAccess Paper or Ask Questions
Twitter IconFacebook IconLinkedin IconWhatsapp IconMessenger Icon