A Probabilistically Motivated Learning Rate Adaptation for Stochastic Optimization
Machine Learning
2021-02-23 v1
Abstract
Machine learning practitioners invest significant manual and computational resources in finding suitable learning rates for optimization algorithms. We provide a probabilistic motivation, in terms of Gaussian inference, for popular stochastic first-order methods. As an important special case, it recovers the Polyak step with a general metric. The inference allows us to relate the learning rate to a dimensionless quantity that can be automatically adapted during training by a control algorithm. The resulting meta-algorithm is shown to adapt learning rates in a robust manner across a large range of initial values when applied to deep learning benchmark problems.
Cite
@article{arxiv.2102.10880,
title = {A Probabilistically Motivated Learning Rate Adaptation for Stochastic Optimization},
author = {Filip de Roos and Carl Jidling and Adrian Wills and Thomas Schön and Philipp Hennig},
journal= {arXiv preprint arXiv:2102.10880},
year = {2021}
}