Optimistic Meta-Gradients
Machine Learning
2023-01-10 v1 Artificial Intelligence
Optimization and Control
Abstract
We study the connection between gradient-based meta-learning and convex op-timisation. We observe that gradient descent with momentum is a special case of meta-gradients, and building on recent results in optimisation, we prove convergence rates for meta-learning in the single task setting. While a meta-learned update rule can yield faster convergence up to constant factor, it is not sufficient for acceleration. Instead, some form of optimism is required. We show that optimism in meta-learning can be captured through Bootstrapped Meta-Gradients (Flennerhag et al., 2022), providing deeper insight into its underlying mechanics.
Cite
@article{arxiv.2301.03236,
title = {Optimistic Meta-Gradients},
author = {Sebastian Flennerhag and Tom Zahavy and Brendan O'Donoghue and Hado van Hasselt and András György and Satinder Singh},
journal= {arXiv preprint arXiv:2301.03236},
year = {2023}
}