Provable Guarantees for Gradient-Based Meta-Learning

TL;DR

This paper introduces a meta-learning algorithm based on online convex optimization that offers strong theoretical guarantees, good sample efficiency, and scalability to deep learning, with empirical validation.

Contribution

It presents the first scalable meta-learning method with theoretical guarantees that improve with task similarity and matches lower bounds under natural assumptions.

Findings

The algorithm achieves sample efficiency in convex settings.

It generalizes better as task similarity increases.

Experimental results confirm theoretical predictions in convex and deep learning contexts.

Abstract

We study the problem of meta-learning through the lens of online convex optimization, developing a meta-algorithm bridging the gap between popular gradient-based meta-learning and classical regularization-based multi-task transfer methods. Our method is the first to simultaneously satisfy good sample efficiency guarantees in the convex setting, with generalization bounds that improve with task-similarity, while also being computationally scalable to modern deep learning architectures and the many-task setting. Despite its simplicity, the algorithm matches, up to a constant factor, a lower bound on the performance of any such parameter-transfer method under natural task similarity assumptions. We use experiments in both convex and deep learning settings to verify and demonstrate the applicability of our theory.