Data-Efficient Image Recognition with Contrastive Predictive Coding
This addresses the problem of reducing labeled data requirements for image recognition tasks, offering a significant but incremental advance over existing unsupervised methods.
The paper tackled data-efficient image recognition by improving Contrastive Predictive Coding to learn representations that reduce the need for labeled data, achieving state-of-the-art linear classification on ImageNet and enabling 2-5x fewer labels for non-linear classification while improving transfer to object detection on PASCAL VOC.
Human observers can learn to recognize new categories of images from a handful of examples, yet doing so with artificial ones remains an open challenge. We hypothesize that data-efficient recognition is enabled by representations which make the variability in natural signals more predictable. We therefore revisit and improve Contrastive Predictive Coding, an unsupervised objective for learning such representations. This new implementation produces features which support state-of-the-art linear classification accuracy on the ImageNet dataset. When used as input for non-linear classification with deep neural networks, this representation allows us to use 2-5x less labels than classifiers trained directly on image pixels. Finally, this unsupervised representation substantially improves transfer learning to object detection on the PASCAL VOC dataset, surpassing fully supervised pre-trained ImageNet classifiers.