Injecting Imbalance Sensitivity for Multi-Task Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhou, Zhipeng, Liu, Liu, Zhao, Peilin, Gong, Wei
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912269711966208
author Zhou, Zhipeng
Liu, Liu
Zhao, Peilin
Gong, Wei
author_facet Zhou, Zhipeng
Liu, Liu
Zhao, Peilin
Gong, Wei
contents Multi-task learning (MTL) has emerged as a promising approach for deploying deep learning models in real-life applications. Recent studies have proposed optimization-based learning paradigms to establish task-shared representations in MTL. However, our paper empirically argues that these studies, specifically gradient-based ones, primarily emphasize the conflict issue while neglecting the potentially more significant impact of imbalance/dominance in MTL. In line with this perspective, we enhance the existing baseline method by injecting imbalance-sensitivity through the imposition of constraints on the projected norms. To demonstrate the effectiveness of our proposed IMbalance-sensitive Gradient (IMGrad) descent method, we evaluate it on multiple mainstream MTL benchmarks, encompassing supervised learning tasks as well as reinforcement learning. The experimental results consistently demonstrate competitive performance.
format Preprint
id arxiv_https___arxiv_org_abs_2503_08006
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Injecting Imbalance Sensitivity for Multi-Task Learning
Zhou, Zhipeng
Liu, Liu
Zhao, Peilin
Gong, Wei
Machine Learning
Artificial Intelligence
Multi-task learning (MTL) has emerged as a promising approach for deploying deep learning models in real-life applications. Recent studies have proposed optimization-based learning paradigms to establish task-shared representations in MTL. However, our paper empirically argues that these studies, specifically gradient-based ones, primarily emphasize the conflict issue while neglecting the potentially more significant impact of imbalance/dominance in MTL. In line with this perspective, we enhance the existing baseline method by injecting imbalance-sensitivity through the imposition of constraints on the projected norms. To demonstrate the effectiveness of our proposed IMbalance-sensitive Gradient (IMGrad) descent method, we evaluate it on multiple mainstream MTL benchmarks, encompassing supervised learning tasks as well as reinforcement learning. The experimental results consistently demonstrate competitive performance.
title Injecting Imbalance Sensitivity for Multi-Task Learning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2503.08006