Meta-Learning with Implicit Gradients

TL;DR

This paper introduces implicit MAML, a meta-learning algorithm that uses implicit differentiation to efficiently compute meta-gradients, enabling scalable, memory-efficient, and optimizer-agnostic few-shot learning.

Contribution

The paper proposes implicit MAML, a novel meta-learning method that leverages implicit differentiation to decouple meta-gradient computation from the inner loop optimizer.

Findings

Implicit MAML achieves empirical improvements on few-shot image recognition tasks.

The method reduces memory usage and computational cost compared to traditional MAML.

It is robust to the number of inner loop gradient steps.

Abstract

A core capability of intelligent systems is the ability to quickly learn new tasks by drawing on prior experience. Gradient (or optimization) based meta-learning has recently emerged as an effective approach for few-shot learning. In this formulation, meta-parameters are learned in the outer loop, while task-specific models are learned in the inner-loop, by using only a small amount of data from the current task. A key challenge in scaling these approaches is the need to differentiate through the inner loop learning process, which can impose considerable computational and memory burdens. By drawing upon implicit differentiation, we develop the implicit MAML algorithm, which depends only on the solution to the inner level optimization and not the path taken by the inner loop optimizer. This effectively decouples the meta-gradient computation from the choice of inner loop optimizer. As a result, our approach is agnostic to the choice of inner loop optimizer and can gracefully handle many gradient steps without vanishing gradients or memory constraints. Theoretically, we prove that implicit MAML can compute accurate meta-gradients with a memory footprint that is, up to small constant factors, no more than that which is required to compute a single inner loop gradient and at no overall increase in the total computational cost. Experimentally, we show that these benefits of implicit MAML translate into empirical gains on few-shot image recognition benchmarks.