ViTaS: Visual Tactile Soft Fusion Contrastive Learning for Reinforcement Learning
Keywords: visuo-tactile representation learning, reinforcement learning, contrastive learning
Abstract: Tactile information plays a crucial role in human manipulation tasks and has recently garnered increasing attention in robotic manipulation. However, existing approaches struggle to effectively integrate visual and tactile information, resulting in suboptimal performance. In this paper, we present **ViTaS**, a simple yet effective framework that incorporates both visual and tactile information to guide an agent's behavior. We introduce _Soft Fusion Contrastive Learning_, an advanced version of conventional contrastive learning method, to enhance the fusion of these two modalities, and adopt a CVAE module to utilize complementary information within visuo-tactile representation. We conduct comprehensive experiments, including $\mathbf{9}$ tasks in simulation environment, across $\mathbf{5}$ different benchmarks, to compare ViTaS with existing baselines. The results demonstrate that ViTaS achieves state-of-the-art performance, with an average improvement of $\mathbf{51}$%. Furthermore, our method significantly enhances sample efficiency while maintaining minimal parameters, underscoring the effectiveness of our approach. The code will be released upon acceptance.
Supplementary Material: zip
Primary Area: reinforcement learning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 1144
Loading