PDiT: Interleaving Perception and Decision-making Transformers for Deep Reinforcement Learning
This work addresses the challenge of improving deep RL networks for researchers and practitioners, offering a generally applicable design that is incremental over existing transformer-based methods.
The paper tackles the problem of designing better deep networks for reinforcement learning by proposing PDiT, an interleaving transformer architecture that separates perception and decision-making, achieving superior performance across various RL settings.
Designing better deep networks and better reinforcement learning (RL) algorithms are both important for deep RL. This work studies the former. Specifically, the Perception and Decision-making Interleaving Transformer (PDiT) network is proposed, which cascades two Transformers in a very natural way: the perceiving one focuses on \emph{the environmental perception} by processing the observation at the patch level, whereas the deciding one pays attention to \emph{the decision-making} by conditioning on the history of the desired returns, the perceiver's outputs, and the actions. Such a network design is generally applicable to a lot of deep RL settings, e.g., both the online and offline RL algorithms under environments with either image observations, proprioception observations, or hybrid image-language observations. Extensive experiments show that PDiT can not only achieve superior performance than strong baselines in different settings but also extract explainable feature representations. Our code is available at \url{https://github.com/maohangyu/PDiT}.