Improving the Generalization of Meta-learning on Unseen Domains via Adversarial Shift

TL;DR

This paper introduces an adversarial shift layer to simulate domain shifts, enhancing meta-learning models' ability to generalize to unseen domains in few-shot classification and regression tasks.

Contribution

It proposes a novel model-agnostic shift layer and adversarial training mechanism to generate pseudo tasks for improved cross-domain meta-learning.

Findings

Achieves state-of-the-art results on cross-domain few-shot classification benchmarks.

Demonstrates improved generalization to unseen domains in regression tasks.

Applicable to various meta-learning frameworks.

Abstract

Meta-learning provides a promising way for learning to efficiently learn and achieves great success in many applications. However, most meta-learning literature focuses on dealing with tasks from a same domain, making it brittle to generalize to tasks from the other unseen domains. In this work, we address this problem by simulating tasks from the other unseen domains to improve the generalization and robustness of meta-learning method. Specifically, we propose a model-agnostic shift layer to learn how to simulate the domain shift and generate pseudo tasks, and develop a new adversarial learning-to-learn mechanism to train it. Based on the pseudo tasks, the meta-learning model can learn cross-domain meta-knowledge, which can generalize well on unseen domains. We conduct extensive experiments under the domain generalization setting. Experimental results demonstrate that the proposed shift layer is applicable to various meta-learning frameworks. Moreover, our method also leads to state-of-the-art performance on different cross-domain few-shot classification benchmarks and produces good results on cross-domain few-shot regression.