Beyond Simple Meta-Learning: Multi-Purpose Models for Multi-Domain, Active and Continual Few-Shot Learning

TL;DR

This paper introduces variance-sensitive models for few-shot learning that outperform existing methods on multiple benchmarks, including meta-learning, transductive, continual, and active learning scenarios.

Contribution

It proposes Simple CNAPS and Transductive CNAPS, novel models that address feature variance assumptions and improve classification accuracy across diverse few-shot tasks.

Findings

Simple CNAPS achieves strong results on Meta-Dataset, mini-ImageNet, and tiered-ImageNet.

Transductive CNAPS outperforms previous methods on Meta-Dataset with unlabelled data.

Models demonstrate robustness and versatility in continual and active learning settings.

Abstract

Modern deep learning requires large-scale extensively labelled datasets for training. Few-shot learning aims to alleviate this issue by learning effectively from few labelled examples. In previously proposed few-shot visual classifiers, it is assumed that the feature manifold, where classifier decisions are made, has uncorrelated feature dimensions and uniform feature variance. In this work, we focus on addressing the limitations arising from this assumption by proposing a variance-sensitive class of models that operates in a low-label regime. The first method, Simple CNAPS, employs a hierarchically regularized Mahalanobis-distance based classifier combined with a state of the art neural adaptive feature extractor to achieve strong performance on Meta-Dataset, mini-ImageNet and tiered-ImageNet benchmarks. We further extend this approach to a transductive learning setting, proposing Transductive CNAPS. This transductive method combines a soft k-means parameter refinement procedure with a two-step task encoder to achieve improved test-time classification accuracy using unlabelled data. Transductive CNAPS achieves state of the art performance on Meta-Dataset. Finally, we explore the use of our methods (Simple and Transductive) for "out of the box" continual and active learning. Extensive experiments on large scale benchmarks illustrate robustness and versatility of this, relatively speaking, simple class of models. All trained model checkpoints and corresponding source codes have been made publicly available.