MGDA Converges under Generalized Smoothness, Provably

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Qi, Xiao, Peiyao, Zou, Shaofeng, Ji, Kaiyi
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912265035317248
author Zhang, Qi
Xiao, Peiyao
Zou, Shaofeng
Ji, Kaiyi
author_facet Zhang, Qi
Xiao, Peiyao
Zou, Shaofeng
Ji, Kaiyi
contents Multi-objective optimization (MOO) is receiving more attention in various fields such as multi-task learning. Recent works provide some effective algorithms with theoretical analysis but they are limited by the standard $L$-smooth or bounded-gradient assumptions, which typically do not hold for neural networks, such as Long short-term memory (LSTM) models and Transformers. In this paper, we study a more general and realistic class of generalized $\ell$-smooth loss functions, where $\ell$ is a general non-decreasing function of gradient norm. We revisit and analyze the fundamental multiple gradient descent algorithm (MGDA) and its stochastic version with double sampling for solving the generalized $\ell$-smooth MOO problems, which approximate the conflict-avoidant (CA) direction that maximizes the minimum improvement among objectives. We provide a comprehensive convergence analysis of these algorithms and show that they converge to an $ε$-accurate Pareto stationary point with a guaranteed $ε$-level average CA distance (i.e., the gap between the updating direction and the CA direction) over all iterations, where totally $\mathcal{O}(ε^{-2})$ and $\mathcal{O}(ε^{-4})$ samples are needed for deterministic and stochastic settings, respectively. We prove that they can also guarantee a tighter $ε$-level CA distance in each iteration using more samples. Moreover, we analyze an efficient variant of MGDA named MGDA-FA using only $\mathcal{O}(1)$ time and space, while achieving the same performance guarantee as MGDA.
format Preprint
id arxiv_https___arxiv_org_abs_2405_19440
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MGDA Converges under Generalized Smoothness, Provably
Zhang, Qi
Xiao, Peiyao
Zou, Shaofeng
Ji, Kaiyi
Machine Learning
Optimization and Control
Multi-objective optimization (MOO) is receiving more attention in various fields such as multi-task learning. Recent works provide some effective algorithms with theoretical analysis but they are limited by the standard $L$-smooth or bounded-gradient assumptions, which typically do not hold for neural networks, such as Long short-term memory (LSTM) models and Transformers. In this paper, we study a more general and realistic class of generalized $\ell$-smooth loss functions, where $\ell$ is a general non-decreasing function of gradient norm. We revisit and analyze the fundamental multiple gradient descent algorithm (MGDA) and its stochastic version with double sampling for solving the generalized $\ell$-smooth MOO problems, which approximate the conflict-avoidant (CA) direction that maximizes the minimum improvement among objectives. We provide a comprehensive convergence analysis of these algorithms and show that they converge to an $ε$-accurate Pareto stationary point with a guaranteed $ε$-level average CA distance (i.e., the gap between the updating direction and the CA direction) over all iterations, where totally $\mathcal{O}(ε^{-2})$ and $\mathcal{O}(ε^{-4})$ samples are needed for deterministic and stochastic settings, respectively. We prove that they can also guarantee a tighter $ε$-level CA distance in each iteration using more samples. Moreover, we analyze an efficient variant of MGDA named MGDA-FA using only $\mathcal{O}(1)$ time and space, while achieving the same performance guarantee as MGDA.
title MGDA Converges under Generalized Smoothness, Provably
topic Machine Learning
Optimization and Control
url https://arxiv.org/abs/2405.19440