DySTreSS: Dynamically Scaled Temperature in Self-Supervised Contrastive Learning
Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation 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.
Keywords: Self-supervised, Contrastive Learning
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Abstract: In contemporary self-supervised contrastive algorithms like SimCLR, MoCo, etc.,the task of balancing attraction between two semantically similar samples and repulsion between two samples from different classes is primarily affected by the presence of hard negative samples. While the InfoNCE loss has been shown to impose penalties based on hardness, the temperature hyper-parameter is the key to regulating the penalties and the trade-off between uniformity and tolerance. In this work, we focus our attention on improving the performance of InfoNCE loss in Self-supervised learning by proposing a novel cosine-similarity dependent temperature scaling function to effectively optimize the distribution of the samples in the feature space. We also provide mathematical analyses to support the construction of such a dynamically scaled temperature function. Experimental evidence shows that the proposed framework outperforms the contrastive loss-based SSL algorithms. Our code is available at https://www.github.com/subanon/dystress.
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: 554
Loading