Matching Networks for One Shot Learning

TLDR

Learning from few examples remains a key challenge, and standard supervised deep learning does not provide a satisfactory solution for rapid concept learning from limited data. The authors aim to develop a one‑shot learning framework that leverages metric learning and external memory augmentation, and to demonstrate its applicability to both vision and language tasks. The framework maps a small labelled support set and an unlabelled example to a label without fine‑tuning, and defines one‑shot problems on Omniglot, ImageNet, and Penn Treebank. The algorithm raises one‑shot accuracy on ImageNet from 87.6 % to 93.2 % and on Omniglot from 88.0 % to 93.8 %, outperforming competing methods.

Abstract

Learning from a few examples remains a key challenge in machine learning. Despite recent advances in important domains such as vision and language, the standard supervised deep learning paradigm does not offer a satisfactory solution for learning new concepts rapidly from little data. In this work, we employ ideas from metric learning based on deep neural features and from recent advances that augment neural networks with external memories. Our framework learns a network that maps a small labelled support set and an unlabelled example to its label, obviating the need for fine-tuning to adapt to new class types. We then define one-shot learning problems on vision (using Omniglot, ImageNet) and language tasks. Our algorithm improves one-shot accuracy on ImageNet from 87.6% to 93.2% and from 88.0% to 93.8% on Omniglot compared to competing approaches. We also demonstrate the usefulness of the same model on language modeling by introducing a one-shot task on the Penn Treebank.