Domain-Generalization to Improve Learning in Meta-Learning Algorithms

TL;DR

This paper presents DGS-MAML, a meta-learning algorithm that combines gradient matching and sharpness-aware minimization to improve generalization and robustness across tasks with limited data, supported by theoretical analysis and strong experimental results.

Contribution

Introduces DGS-MAML, a novel meta-learning method integrating gradient matching with sharpness-aware minimization within a bi-level framework for better task generalization.

Findings

DGS-MAML outperforms existing methods on benchmark datasets.

Theoretical analysis confirms convergence and generalization guarantees.

Effective for few-shot learning and rapid adaptation scenarios.

Abstract

This paper introduces Domain Generalization Sharpness-Aware Minimization Model-Agnostic Meta-Learning (DGS-MAML), a novel meta-learning algorithm designed to generalize across tasks with limited training data. DGS-MAML combines gradient matching with sharpness-aware minimization in a bi-level optimization framework to enhance model adaptability and robustness. We support our method with theoretical analysis using PAC-Bayes and convergence guarantees. Experimental results on benchmark datasets show that DGS-MAML outperforms existing approaches in terms of accuracy and generalization. The proposed method is particularly useful for scenarios requiring few-shot learning and quick adaptation, and the source code is publicly available at GitHub.