Exploring Simple Siamese Representation Learning

TL;DR

This paper demonstrates that simple Siamese networks can learn meaningful visual representations without negative pairs, large batches, or momentum encoders, primarily relying on a stop-gradient operation to prevent collapse, achieving competitive results.

Contribution

The paper introduces SimSiam, a simple Siamese network approach that challenges existing beliefs by showing effective unsupervised learning without complex components.

Findings

Simple Siamese networks can learn meaningful representations.

Stop-gradient operation is crucial to prevent collapse.

SimSiam achieves competitive results on ImageNet.

Abstract

Siamese networks have become a common structure in various recent models for unsupervised visual representation learning. These models maximize the similarity between two augmentations of one image, subject to certain conditions for avoiding collapsing solutions. In this paper, we report surprising empirical results that simple Siamese networks can learn meaningful representations even using none of the following: (i) negative sample pairs, (ii) large batches, (iii) momentum encoders. Our experiments show that collapsing solutions do exist for the loss and structure, but a stop-gradient operation plays an essential role in preventing collapsing. We provide a hypothesis on the implication of stop-gradient, and further show proof-of-concept experiments verifying it. Our "SimSiam" method achieves competitive results on ImageNet and downstream tasks. We hope this simple baseline will motivate people to rethink the roles of Siamese architectures for unsupervised representation learning. Code will be made available.