Self-Supervised Representation Learning: Introduction, advances, and challenges

Self-supervised representation learning (SSRL) methods aim to provide powerful, deep feature learning without the requirement of large annotated data sets, thus alleviating the annotation bottleneck—one of the main barriers to the practical deployment of deep learning today. These techniques have advanced rapidly in recent years, with their efficacy approaching and sometimes surpassing fully supervised pretraining alternatives across a variety of data modalities, including image, video, sound, text, and graphs. This article introduces this vibrant area, including key concepts, the four main families of approaches and associated state-of-the-art techniques, and how self-supervised methods are applied to diverse…

• EA
  Engineering and Physical Sciences Research Council

  Awards: EP/S000631/1, EP/R026173/1, EP/L016427/1, EP/S000631/1, EP/R026173