Broaden Your Views for Self-Supervised Video Learning
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.
Cite
@article{arxiv.2103.16559,
title = {Broaden Your Views for Self-Supervised Video Learning},
author = {Adrià Recasens and Pauline Luc and Jean-Baptiste Alayrac and Luyu Wang and Ross Hemsley and Florian Strub and Corentin Tallec and Mateusz Malinowski and Viorica Patraucean and Florent Altché and Michal Valko and Jean-Bastien Grill and Aäron van den Oord and Andrew Zisserman},
journal= {arXiv preprint arXiv:2103.16559},
year = {2021}
}
Comments
This paper is an extended version of our ICCV-21 paper. It includes more results as well as a minor architectural variation which improves results