Rethinking Latent Redundancy in Behavior Cloning: An Information Bottleneck Approach for Robot Manipulation

Shuanghao Bai, Wanqi Zhou, Pengxiang Ding, Wei Zhao, Donglin Wang, Badong Chen
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:2560-2580, 2025.

Abstract

Behavior Cloning (BC) is a widely adopted visual imitation learning method in robot manipulation. Current BC approaches often enhance generalization by leveraging large datasets and incorporating additional visual and textual modalities to capture more diverse information. However, these methods overlook whether the learned representations contain redundant information and lack a solid theoretical foundation to guide the learning process. To address these limitations, we adopt an information-theoretic perspective and introduce mutual information to quantify and mitigate redundancy in latent representations. Building on this, we incorporate the Information Bottleneck (IB) principle into BC, which extends the idea of reducing redundancy by providing a structured framework for compressing irrelevant information while preserving task-relevant features. This work presents the first comprehensive study on redundancy in latent representations across various methods, backbones, and experimental settings, while extending the generalizability of the IB to BC. Extensive experiments and analyses on the CortexBench and LIBERO benchmarks show consistent performance improvements with IB across various settings, underscoring the importance of reducing input data redundancy and highlighting its practical value for real-world applications.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-bai25e, title = {Rethinking Latent Redundancy in Behavior Cloning: An Information Bottleneck Approach for Robot Manipulation}, author = {Bai, Shuanghao and Zhou, Wanqi and Ding, Pengxiang and Zhao, Wei and Wang, Donglin and Chen, Badong}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {2560--2580}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/bai25e/bai25e.pdf}, url = {https://proceedings.mlr.press/v267/bai25e.html}, abstract = {Behavior Cloning (BC) is a widely adopted visual imitation learning method in robot manipulation. Current BC approaches often enhance generalization by leveraging large datasets and incorporating additional visual and textual modalities to capture more diverse information. However, these methods overlook whether the learned representations contain redundant information and lack a solid theoretical foundation to guide the learning process. To address these limitations, we adopt an information-theoretic perspective and introduce mutual information to quantify and mitigate redundancy in latent representations. Building on this, we incorporate the Information Bottleneck (IB) principle into BC, which extends the idea of reducing redundancy by providing a structured framework for compressing irrelevant information while preserving task-relevant features. This work presents the first comprehensive study on redundancy in latent representations across various methods, backbones, and experimental settings, while extending the generalizability of the IB to BC. Extensive experiments and analyses on the CortexBench and LIBERO benchmarks show consistent performance improvements with IB across various settings, underscoring the importance of reducing input data redundancy and highlighting its practical value for real-world applications.} }
Endnote
%0 Conference Paper %T Rethinking Latent Redundancy in Behavior Cloning: An Information Bottleneck Approach for Robot Manipulation %A Shuanghao Bai %A Wanqi Zhou %A Pengxiang Ding %A Wei Zhao %A Donglin Wang %A Badong Chen %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-bai25e %I PMLR %P 2560--2580 %U https://proceedings.mlr.press/v267/bai25e.html %V 267 %X Behavior Cloning (BC) is a widely adopted visual imitation learning method in robot manipulation. Current BC approaches often enhance generalization by leveraging large datasets and incorporating additional visual and textual modalities to capture more diverse information. However, these methods overlook whether the learned representations contain redundant information and lack a solid theoretical foundation to guide the learning process. To address these limitations, we adopt an information-theoretic perspective and introduce mutual information to quantify and mitigate redundancy in latent representations. Building on this, we incorporate the Information Bottleneck (IB) principle into BC, which extends the idea of reducing redundancy by providing a structured framework for compressing irrelevant information while preserving task-relevant features. This work presents the first comprehensive study on redundancy in latent representations across various methods, backbones, and experimental settings, while extending the generalizability of the IB to BC. Extensive experiments and analyses on the CortexBench and LIBERO benchmarks show consistent performance improvements with IB across various settings, underscoring the importance of reducing input data redundancy and highlighting its practical value for real-world applications.
APA
Bai, S., Zhou, W., Ding, P., Zhao, W., Wang, D. & Chen, B.. (2025). Rethinking Latent Redundancy in Behavior Cloning: An Information Bottleneck Approach for Robot Manipulation. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:2560-2580 Available from https://proceedings.mlr.press/v267/bai25e.html.

Related Material