Efficient Automatic Meta Optimization Search for Few-Shot Learning

TL;DR

This paper introduces an automated neural architecture search framework for meta-learning in few-shot tasks, significantly reducing search time and achieving high transferability across datasets.

Contribution

It presents a universal NAS-based framework for optimizing meta-learning architectures automatically, outperforming hand-designed models in few-shot learning.

Findings

Achieves competitive results on Mini-ImageNet and Omniglot.

Search process takes only 1-2 GPU days.

Discovered architectures transfer well across datasets.

Abstract

Previous works on meta-learning either relied on elaborately hand-designed network structures or adopted specialized learning rules to a particular domain. We propose a universal framework to optimize the meta-learning process automatically by adopting neural architecture search technique (NAS). NAS automatically generates and evaluates meta-learner's architecture for few-shot learning problems, while the meta-learner uses meta-learning algorithm to optimize its parameters based on the distribution of learning tasks. Parameter sharing and experience replay are adopted to accelerate the architectures searching process, so it takes only 1-2 GPU days to find good architectures. Extensive experiments on Mini-ImageNet and Omniglot show that our algorithm excels in few-shot learning tasks. The best architecture found on Mini-ImageNet achieves competitive results when transferred to Omniglot, which shows the high transferability of architectures among different computer vision problems.