Alpha MAML: Adaptive Model-Agnostic Meta-Learning

TL;DR

Alpha MAML introduces an online hyperparameter adaptation scheme to improve training stability and reduce tuning in meta-learning, making it more practical for few-shot learning tasks.

Contribution

The paper proposes Alpha MAML, an extension of MAML that automatically adapts hyperparameters during training, eliminating the need for manual tuning.

Findings

Reduces hyperparameter tuning requirements in MAML

Improves training stability and robustness

Demonstrates effectiveness on Omniglot dataset

Abstract

Model-agnostic meta-learning (MAML) is a meta-learning technique to train a model on a multitude of learning tasks in a way that primes the model for few-shot learning of new tasks. The MAML algorithm performs well on few-shot learning problems in classification, regression, and fine-tuning of policy gradients in reinforcement learning, but comes with the need for costly hyperparameter tuning for training stability. We address this shortcoming by introducing an extension to MAML, called Alpha MAML, to incorporate an online hyperparameter adaptation scheme that eliminates the need to tune meta-learning and learning rates. Our results with the Omniglot database demonstrate a substantial reduction in the need to tune MAML training hyperparameters and improvement to training stability with less sensitivity to hyperparameter choice.