MetaMix: Improved Meta-Learning with Interpolation-based Consistency Regularization

TL;DR

MetaMix enhances MAML-based meta-learning by generating virtual feature-target pairs within episodes, improving decision boundary generalization and achieving state-of-the-art results on few-shot classification benchmarks.

Contribution

Introduces MetaMix, a novel regularization method that generates virtual pairs to improve generalization in MAML-based meta-learning algorithms.

Findings

MetaMix improves performance of MAML variants on mini-ImageNet, CUB, and FC100.

MetaMix achieves state-of-the-art results when combined with Meta-Transfer Learning.

Experimental results demonstrate better decision boundary generalization.

Abstract

Model-Agnostic Meta-Learning (MAML) and its variants are popular few-shot classification methods. They train an initializer across a variety of sampled learning tasks (also known as episodes) such that the initialized model can adapt quickly to new tasks. However, current MAML-based algorithms have limitations in forming generalizable decision boundaries. In this paper, we propose an approach called MetaMix. It generates virtual feature-target pairs within each episode to regularize the backbone models. MetaMix can be integrated with any of the MAML-based algorithms and learn the decision boundaries generalizing better to new tasks. Experiments on the mini-ImageNet, CUB, and FC100 datasets show that MetaMix improves the performance of MAML-based algorithms and achieves state-of-the-art result when integrated with Meta-Transfer Learning.