Momentum Contrast for Unsupervised Visual Representation Learning

TLDR

The paper introduces Momentum Contrast (MoCo) for unsupervised visual representation learning. MoCo constructs a dynamic dictionary using a queue and a moving‑averaged encoder to enable contrastive learning as dictionary lookup. MoCo achieves competitive ImageNet linear classification performance, transfers effectively to downstream tasks, and can surpass supervised pre‑training on several detection and segmentation benchmarks, indicating that the gap between unsupervised and supervised representation learning has largely closed.

Abstract

We present Momentum Contrast (MoCo) for unsupervised visual representation learning. From a perspective on contrastive learning as dictionary look-up, we build a dynamic dictionary with a queue and a moving-averaged encoder. This enables building a large and consistent dictionary on-the-fly that facilitates contrastive unsupervised learning. MoCo provides competitive results under the common linear protocol on ImageNet classification. More importantly, the representations learned by MoCo transfer well to downstream tasks. MoCo can outperform its supervised pre-training counterpart in 7 detection/segmentation tasks on PASCAL VOC, COCO, and other datasets, sometimes surpassing it by large margins. This suggests that the gap between unsupervised and supervised representation learning has been largely closed in many vision tasks.