Contrastive Multiview Coding

TLDR

Humans perceive the world through multiple noisy sensory channels that share underlying factors such as physics, geometry, and semantics. We investigate whether a powerful representation models view‑invariant factors. We use multiview contrastive learning that maximizes mutual information between different views of the same scene while remaining compact, and it scales to any number of view modalities. The contrastive loss outperforms cross‑view prediction, and adding more views improves semantic capture, yielding state‑of‑the‑art results on image and video unsupervised learning benchmarks. Code is available at http://github.com/HobbitLong/CMC/.

Abstract

Humans view the world through many sensory channels, e.g., the long-wavelength light channel, viewed by the left eye, or the high-frequency vibrations channel, heard by the right ear. Each view is noisy and incomplete, but important factors, such as physics, geometry, and semantics, tend to be shared between all views (e.g., a "dog" can be seen, heard, and felt). We investigate the classic hypothesis that a powerful representation is one that models view-invariant factors. We study this hypothesis under the framework of multiview contrastive learning, where we learn a representation that aims to maximize mutual information between different views of the same scene but is otherwise compact. Our approach scales to any number of views, and is view-agnostic. We analyze key properties of the approach that make it work, finding that the contrastive loss outperforms a popular alternative based on cross-view prediction, and that the more views we learn from, the better the resulting representation captures underlying scene semantics. Our approach achieves state-of-the-art results on image and video unsupervised learning benchmarks. Code is released at: http://github.com/HobbitLong/CMC/.