Non-exponentially weighted aggregation: regret bounds for unbounded loss functions
Machine Learning
2022-01-17 v5 Machine Learning
Abstract
We tackle the problem of online optimization with a general, possibly unbounded, loss function. It is well known that when the loss is bounded, the exponentially weighted aggregation strategy (EWA) leads to a regret in after steps. In this paper, we study a generalized aggregation strategy, where the weights no longer depend exponentially on the losses. Our strategy is based on Follow The Regularized Leader (FTRL): we minimize the expected losses plus a regularizer, that is here a -divergence. When the regularizer is the Kullback-Leibler divergence, we obtain EWA as a special case. Using alternative divergences enables unbounded losses, at the cost of a worst regret bound in some cases.
Cite
@article{arxiv.2009.03017,
title = {Non-exponentially weighted aggregation: regret bounds for unbounded loss functions},
author = {Pierre Alquier},
journal= {arXiv preprint arXiv:2009.03017},
year = {2022}
}