Broaden Your Views for Self-Supervised Video Learning

TL;DR

BraVe introduces a novel self-supervised video learning framework that leverages different temporal views and modalities, achieving state-of-the-art results on multiple video and audio benchmarks.

Contribution

It proposes a new approach using narrow and broad temporal views with different backbones, incorporating various modalities for improved video representation learning.

Findings

Achieves state-of-the-art results on UCF101 and HMDB51

Outperforms previous methods on Kinetics and AudioSet

Effectively integrates multiple modalities like optical flow and audio

Abstract

Most successful self-supervised learning methods are trained to align the representations of two independent views from the data. State-of-the-art methods in video are inspired by image techniques, where these two views are similarly extracted by cropping and augmenting the resulting crop. However, these methods miss a crucial element in the video domain: time. We introduce BraVe, a self-supervised learning framework for video. In BraVe, one of the views has access to a narrow temporal window of the video while the other view has a broad access to the video content. Our models learn to generalise from the narrow view to the general content of the video. Furthermore, BraVe processes the views with different backbones, enabling the use of alternative augmentations or modalities into the broad view such as optical flow, randomly convolved RGB frames, audio or their combinations. We demonstrate that BraVe achieves state-of-the-art results in self-supervised representation learning on standard video and audio classification benchmarks including UCF101, HMDB51, Kinetics, ESC-50 and AudioSet.