English

Reinforcement Learning with Generalizable Gaussian Splatting

Computer Vision and Pattern Recognition 2024-08-07 v3 Artificial Intelligence Machine Learning

Abstract

An excellent representation is crucial for reinforcement learning (RL) performance, especially in vision-based reinforcement learning tasks. The quality of the environment representation directly influences the achievement of the learning task. Previous vision-based RL typically uses explicit or implicit ways to represent environments, such as images, points, voxels, and neural radiance fields. However, these representations contain several drawbacks. They cannot either describe complex local geometries or generalize well to unseen scenes, or require precise foreground masks. Moreover, these implicit neural representations are akin to a ``black box", significantly hindering interpretability. 3D Gaussian Splatting (3DGS), with its explicit scene representation and differentiable rendering nature, is considered a revolutionary change for reconstruction and representation methods. In this paper, we propose a novel Generalizable Gaussian Splatting framework to be the representation of RL tasks, called GSRL. Through validation in the RoboMimic environment, our method achieves better results than other baselines in multiple tasks, improving the performance by 10%, 44%, and 15% compared with baselines on the hardest task. This work is the first attempt to leverage generalizable 3DGS as a representation for RL.

Keywords

Cite

@article{arxiv.2404.07950,
  title  = {Reinforcement Learning with Generalizable Gaussian Splatting},
  author = {Jiaxu Wang and Qiang Zhang and Jingkai Sun and Jiahang Cao and Gang Han and Wen Zhao and Weining Zhang and Yecheng Shao and Yijie Guo and Renjing Xu},
  journal= {arXiv preprint arXiv:2404.07950},
  year   = {2024}
}

Comments

7 pages,2 figures

R2 v1 2026-06-28T15:51:36.203Z