On Feature Decorrelation in Self-Supervised Learning

Abstract

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 gains from feature decorrelation are verified empirically to highlight the importance and the potential of this insight.

arXiv commentICCV 2021 Oral. The first two authors contribute equally

References in corpus (19)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Bootstrap your own latent: A new approach to self-supervised Learning
• Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
• Improved Baselines with Momentum Contrastive Learning
• Unsupervised Learning of Visual Features by Contrasting Cluster Assignments
• Learning deep representations by mutual information estimation and maximization
• SGDR: Stochastic Gradient Descent with Warm Restarts
• What Makes for Good Views for Contrastive Learning?
• Large Batch Training of Convolutional Networks
• Prototypical Contrastive Learning of Unsupervised Representations
• Achieving Human Parity in Conversational Speech Recognition
• Exploring Simple Siamese Representation Learning
• Whitening for Self-Supervised Representation Learning
• Understanding self-supervised Learning Dynamics without Contrastive Pairs
• BYOL works even without batch statistics
• Self-supervised Pre-training with Hard Examples Improves Visual Representations
• Whitening and Coloring batch transform for GANs
• Mine Your Own vieW: Self-Supervised Learning Through Across-Sample Prediction
• Training Deep Neural Networks Without Batch Normalization

Cited by in corpus (2)

• Understanding Dimensional Collapse in Contrastive Self-supervised Learning
• From Canonical Correlation Analysis to Self-supervised Graph Neural Networks