Enregistré dans:
Détails bibliographiques
Auteurs principaux: Chayti, El Mahdi, Karimireddy, Sai Praneeth
Format: Preprint
Publié: 2022
Sujets:
Accès en ligne:https://arxiv.org/abs/2206.00395
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866917152402964480
author Chayti, El Mahdi
Karimireddy, Sai Praneeth
author_facet Chayti, El Mahdi
Karimireddy, Sai Praneeth
contents We investigate the fundamental optimization question of minimizing a target function $f$, whose gradients are expensive to compute or have limited availability, given access to some auxiliary side function $h$ whose gradients are cheap or more available. This formulation captures many settings of practical relevance, such as i) re-using batches in SGD, ii) transfer learning, iii) federated learning, iv) training with compressed models/dropout, Et cetera. We propose two generic new algorithms that apply in all these settings; we also prove that we can benefit from this framework under the Hessian similarity assumption between the target and side information. A benefit is obtained when this similarity measure is small; we also show a potential benefit from stochasticity when the auxiliary noise is correlated with that of the target function.
format Preprint
id arxiv_https___arxiv_org_abs_2206_00395
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Optimization with Access to Auxiliary Information
Chayti, El Mahdi
Karimireddy, Sai Praneeth
Machine Learning
Optimization and Control
We investigate the fundamental optimization question of minimizing a target function $f$, whose gradients are expensive to compute or have limited availability, given access to some auxiliary side function $h$ whose gradients are cheap or more available. This formulation captures many settings of practical relevance, such as i) re-using batches in SGD, ii) transfer learning, iii) federated learning, iv) training with compressed models/dropout, Et cetera. We propose two generic new algorithms that apply in all these settings; we also prove that we can benefit from this framework under the Hessian similarity assumption between the target and side information. A benefit is obtained when this similarity measure is small; we also show a potential benefit from stochasticity when the auxiliary noise is correlated with that of the target function.
title Optimization with Access to Auxiliary Information
topic Machine Learning
Optimization and Control
url https://arxiv.org/abs/2206.00395