Continual Optimization with Symmetry Teleportation for Multi-Task Learning

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhou, Zhipeng, Meng, Ziqiao, Wu, Pengcheng, Zhao, Peilin, Miao, Chunyan
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912262099304448
author Zhou, Zhipeng
Meng, Ziqiao
Wu, Pengcheng
Zhao, Peilin
Miao, Chunyan
author_facet Zhou, Zhipeng
Meng, Ziqiao
Wu, Pengcheng
Zhao, Peilin
Miao, Chunyan
contents Multi-task learning (MTL) is a widely explored paradigm that enables the simultaneous learning of multiple tasks using a single model. Despite numerous solutions, the key issues of optimization conflict and task imbalance remain under-addressed, limiting performance. Unlike existing optimization-based approaches that typically reweight task losses or gradients to mitigate conflicts or promote progress, we propose a novel approach based on Continual Optimization with Symmetry Teleportation (COST). During MTL optimization, when an optimization conflict arises, we seek an alternative loss-equivalent point on the loss landscape to reduce conflict. Specifically, we utilize a low-rank adapter (LoRA) to facilitate this practical teleportation by designing convergent, loss-invariant objectives. Additionally, we introduce a historical trajectory reuse strategy to continually leverage the benefits of advanced optimizers. Extensive experiments on multiple mainstream datasets demonstrate the effectiveness of our approach. COST is a plug-and-play solution that enhances a wide range of existing MTL methods. When integrated with state-of-the-art methods, COST achieves superior performance.
format Preprint
id arxiv_https___arxiv_org_abs_2503_04046
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Continual Optimization with Symmetry Teleportation for Multi-Task Learning
Zhou, Zhipeng
Meng, Ziqiao
Wu, Pengcheng
Zhao, Peilin
Miao, Chunyan
Machine Learning
Artificial Intelligence
Multi-task learning (MTL) is a widely explored paradigm that enables the simultaneous learning of multiple tasks using a single model. Despite numerous solutions, the key issues of optimization conflict and task imbalance remain under-addressed, limiting performance. Unlike existing optimization-based approaches that typically reweight task losses or gradients to mitigate conflicts or promote progress, we propose a novel approach based on Continual Optimization with Symmetry Teleportation (COST). During MTL optimization, when an optimization conflict arises, we seek an alternative loss-equivalent point on the loss landscape to reduce conflict. Specifically, we utilize a low-rank adapter (LoRA) to facilitate this practical teleportation by designing convergent, loss-invariant objectives. Additionally, we introduce a historical trajectory reuse strategy to continually leverage the benefits of advanced optimizers. Extensive experiments on multiple mainstream datasets demonstrate the effectiveness of our approach. COST is a plug-and-play solution that enhances a wide range of existing MTL methods. When integrated with state-of-the-art methods, COST achieves superior performance.
title Continual Optimization with Symmetry Teleportation for Multi-Task Learning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2503.04046