Learning feed-forward one-shot learners

TL;DR

This paper introduces a novel deep learning approach called learnet, which predicts parameters of a pupil network from a single example, enabling efficient one-shot learning for classification and tracking tasks.

Contribution

The paper presents a learnet architecture that predicts deep model parameters from one example, enabling end-to-end one-shot learning without large training datasets.

Findings

Successful character recognition from single exemplars in Omniglot

Effective visual object tracking from a single initial exemplar

Demonstrated end-to-end training of one-shot learners

Abstract

One-shot learning is usually tackled by using generative models or discriminative embeddings. Discriminative methods based on deep learning, which are very effective in other learning scenarios, are ill-suited for one-shot learning as they need large amounts of training data. In this paper, we propose a method to learn the parameters of a deep model in one shot. We construct the learner as a second deep network, called a learnet, which predicts the parameters of a pupil network from a single exemplar. In this manner we obtain an efficient feed-forward one-shot learner, trained end-to-end by minimizing a one-shot classification objective in a learning to learn formulation. In order to make the construction feasible, we propose a number of factorizations of the parameters of the pupil network. We demonstrate encouraging results by learning characters from single exemplars in Omniglot, and by tracking visual objects from a single initial exemplar in the Visual Object Tracking benchmark.