BarlowRL: Barlow Twins for Data-Efficient Reinforcement Learning

Omer Veysel Cagatan, Baris Akgun
Proceedings of the 15th Asian Conference on Machine Learning, PMLR 222:201-216, 2024.

Abstract

This paper introduces BarlowRL, a data-efficient reinforcement learning agent that combines the Barlow Twins self-supervised learning framework with DER (Data-Efficient Rainbow) algorithm. BarlowRL outperforms both DER and its contrastive counterpart CURL on the Atari 100k benchmark. BarlowRL avoids dimensional collapse by enforcing information spread to the whole space. This helps RL algorithms to utilize uniformly spread state representation that eventually results in a remarkable performance. The integration of Barlow Twins with DER enhances data efficiency and achieves superior performance in the RL tasks. BarlowRL demonstrates the potential of incorporating self-supervised learning techniques to improve RL algorithms.

Cite this Paper

BibTeX
@InProceedings{pmlr-v222-cagatan24a, title = {{BarlowRL}: {B}arlow Twins for Data-Efficient Reinforcement Learning}, author = {Cagatan, Omer Veysel and Akgun, Baris}, booktitle = {Proceedings of the 15th Asian Conference on Machine Learning}, pages = {201--216}, year = {2024}, editor = {Yanıkoğlu, Berrin and Buntine, Wray}, volume = {222}, series = {Proceedings of Machine Learning Research}, month = {11--14 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v222/cagatan24a/cagatan24a.pdf}, url = {https://proceedings.mlr.press/v222/cagatan24a.html}, abstract = {This paper introduces BarlowRL, a data-efficient reinforcement learning agent that combines the Barlow Twins self-supervised learning framework with DER (Data-Efficient Rainbow) algorithm. BarlowRL outperforms both DER and its contrastive counterpart CURL on the Atari 100k benchmark. BarlowRL avoids dimensional collapse by enforcing information spread to the whole space. This helps RL algorithms to utilize uniformly spread state representation that eventually results in a remarkable performance. The integration of Barlow Twins with DER enhances data efficiency and achieves superior performance in the RL tasks. BarlowRL demonstrates the potential of incorporating self-supervised learning techniques to improve RL algorithms.} }
Endnote
%0 Conference Paper %T BarlowRL: Barlow Twins for Data-Efficient Reinforcement Learning %A Omer Veysel Cagatan %A Baris Akgun %B Proceedings of the 15th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Berrin Yanıkoğlu %E Wray Buntine %F pmlr-v222-cagatan24a %I PMLR %P 201--216 %U https://proceedings.mlr.press/v222/cagatan24a.html %V 222 %X This paper introduces BarlowRL, a data-efficient reinforcement learning agent that combines the Barlow Twins self-supervised learning framework with DER (Data-Efficient Rainbow) algorithm. BarlowRL outperforms both DER and its contrastive counterpart CURL on the Atari 100k benchmark. BarlowRL avoids dimensional collapse by enforcing information spread to the whole space. This helps RL algorithms to utilize uniformly spread state representation that eventually results in a remarkable performance. The integration of Barlow Twins with DER enhances data efficiency and achieves superior performance in the RL tasks. BarlowRL demonstrates the potential of incorporating self-supervised learning techniques to improve RL algorithms.
APA
Cagatan, O.V. & Akgun, B.. (2024). BarlowRL: Barlow Twins for Data-Efficient Reinforcement Learning. Proceedings of the 15th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 222:201-216 Available from https://proceedings.mlr.press/v222/cagatan24a.html.

Related Material