Meta-strategy for Learning Tuning Parameters with Guarantees
Machine Learning
2021-11-15 v3 Machine Learning
Statistics Theory
Computation
Statistics Theory
Abstract
Online learning methods, like the online gradient algorithm (OGA) and exponentially weighted aggregation (EWA), often depend on tuning parameters that are difficult to set in practice. We consider an online meta-learning scenario, and we propose a meta-strategy to learn these parameters from past tasks. Our strategy is based on the minimization of a regret bound. It allows to learn the initialization and the step size in OGA with guarantees. It also allows to learn the prior or the learning rate in EWA. We provide a regret analysis of the strategy. It allows to identify settings where meta-learning indeed improves on learning each task in isolation.
Cite
@article{arxiv.2102.02504,
title = {Meta-strategy for Learning Tuning Parameters with Guarantees},
author = {Dimitri Meunier and Pierre Alquier},
journal= {arXiv preprint arXiv:2102.02504},
year = {2021}
}