Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

On Feature Decorrelation in Self-Supervised Learning

May 02, 2021
Tianyu Hua, Wenxiao Wang, Zihui Xue, Yue Wang, Sucheng Ren, Hang Zhao

Share this with someone who'll enjoy it:

In self-supervised representation learning, a common idea behind most of the state-of-the-art approaches is to enforce the robustness of the representations to predefined augmentations. A potential issue of this idea is the existence of completely collapsed solutions (i.e., constant features), which are typically avoided implicitly by carefully chosen implementation details. In this work, we study a relatively concise framework containing the most common components from recent approaches. We verify the existence of complete collapse and discover another reachable collapse pattern that is usually overlooked, namely dimensional collapse. We connect dimensional collapse with strong correlations between axes and consider such connection as a strong motivation for feature decorrelation (i.e., standardizing the covariance matrix). The capability of correlation as an unsupervised metric and the gains from feature decorrelation are verified empirically to highlight the importance and the potential of this insight.

* The first two authors contribute equally 

   Access Paper Source

Share this with someone who'll enjoy it: