Intrinsic Dynamics-Driven Generalizable Scene Representations for Vision-Oriented Decision-Making Applications

TL;DR

This paper introduces DSR, a novel intrinsic dynamics-driven representation learning method for visual reinforcement learning that enhances scene representation and decision-making, especially in complex tasks like autonomous driving.

Contribution

The paper proposes a new sequence model-based approach, DSR, that leverages state-transition dynamics to improve generalizable scene representations in visual RL tasks.

Findings

Achieved 78.9% performance improvement over baseline in DMControl tasks.

Outperformed existing methods in real-world autonomous driving simulations.

Validated superior generalization ability through qualitative analysis.

Abstract

How to improve the ability of scene representation is a key issue in vision-oriented decision-making applications, and current approaches usually learn task-relevant state representations within visual reinforcement learning to address this problem. While prior work typically introduces one-step behavioral similarity metrics with elements (e.g., rewards and actions) to extract task-relevant state information from observations, they often ignore the inherent dynamics relationships among the elements that are essential for learning accurate representations, which further impedes the discrimination of short-term similar task/behavior information in long-term dynamics transitions. To alleviate this problem, we propose an intrinsic dynamics-driven representation learning method with sequence models in visual reinforcement learning, namely DSR. Concretely, DSR optimizes the parameterized encoder by the state-transition dynamics of the underlying system, which prompts the latent encoding information to satisfy the state-transition process and then the state space and the noise space can be distinguished. In the implementation and to further improve the representation ability of DSR on encoding similar tasks, sequential elements' frequency domain and multi-step prediction are adopted for sequentially modeling the inherent dynamics. Finally, experimental results show that DSR has achieved significant performance improvements in the visual Distracting DMControl control tasks, especially with an average of 78.9\% over the backbone baseline. Further results indicate that it also achieves the best performances in real-world autonomous driving applications on the CARLA simulator. Moreover, qualitative analysis results validate that our method possesses the superior ability to learn generalizable scene representations on visual tasks. The source code is available at https://github.com/DMU-XMU/DSR.