EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Task-Relevant Representation Decoupling for Visual Reinforcement Learning Generalization

Jinwen Wang, Youfang Lin, Xiaobo Hu, Qian Xu, Shuo Wang, Zhuo Chen, Kai Lv

ACM Transactions on Multimedia Computing, Communications, and Applications · 31.08.2026 · 22 (8) · 1-24

DOI: 10.1145/3820367 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACM Transactions on Multimedia Computing, Communications, and Applications, online erschienen am 21.07.2026. Verfasst von Jinwen Wang und 6 weiteren Autor:innen. Die Arbeit nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Fundamental Research Funds for the Central Universities, Beijing Natural Science Foundation, National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Visual Reinforcement Learning (VRL) has achieved considerable success in solving control tasks. However, generalizing learned policies to new environments remains a major challenge, as agents often overfit to task-irrelevant features in the training environment. To solve this problem, we introduce the concept of decoupling observations into task-relevant and task-irrelevant representations. Building on this idea, we propose a self-supervised T ask- R elevant R epresentation D ecoupling (T2RD) algorithm for VRL. This algorithm consists of three components: task-relevant representation consistency , cross-reconstruction , and cross-dynamic prediction . The first two components achieve the decoupling of content and style features, but the resulting content representations are not necessarily task-relevant. To further refine task-relevant features from content representations, we design the third component that introduces dynamic prediction. T2RD achieves State-of-the-Art (SOTA) generalization performance and sample efficiency in the DeepMind Control Suite and Robotic Manipulation tasks.

Förderung

  • Fundamental Research Funds for the Central Universities
  • Beijing Natural Science Foundation
  • National Natural Science Foundation of China

Zitieren

APA

Wang, J., Lin, Y., Hu, X., Xu, Q., Wang, S., Chen, Z., & Lv, K. (2026). Task-Relevant Representation Decoupling for Visual Reinforcement Learning Generalization. ACM Transactions on Multimedia Computing, Communications, and Applications, 22(8), 1-24. https://doi.org/10.1145/3820367

Vancouver

Wang J, Lin Y, Hu X, Xu Q, Wang S, Chen Z, et al. Task-Relevant Representation Decoupling for Visual Reinforcement Learning Generalization. ACM Transactions on Multimedia Computing, Communications, and Applications. 2026;22(8):1-24. doi:10.1145/3820367

BibTeX
@article{Wang2026,
  author = {Wang, Jinwen and Lin, Youfang and Hu, Xiaobo and Xu, Qian and Wang, Shuo and Chen, Zhuo and Lv, Kai},
  title = {Task-Relevant Representation Decoupling for Visual Reinforcement Learning Generalization},
  journal = {ACM Transactions on Multimedia Computing, Communications, and Applications},
  year = {2026},
  volume = {22},
  number = {8},
  pages = {1--24},
  doi = {10.1145/3820367}
}
Kostenlos starten →