基于对称传送的持续优化在多任务学习中的应用
Continual Optimization with Symmetry Teleportation for Multi-Task Learning
AI总结:
针对多任务学习中优化冲突和任务不平衡问题,提出基于对称传送的持续优化方法(COST),利用低秩适配器寻找损失等价点以减少冲突,并引入历史轨迹重用策略,作为即插即用方案提升现有方法性能。
摘要:
多任务学习(MTL)是一种广泛探索的范式,能够使用单一模型同时学习多个任务。尽管已有众多解决方案,但优化冲突和任务不平衡这两个关键问题仍未得到充分解决,限制了性能。现有的基于优化的方法通常通过对任务损失或梯度进行重加权来缓解冲突或促进进展,与此不同,我们提出了一种基于持续优化与对称传送(COST)的新颖方法。在MTL优化过程中,当出现优化冲突时,我们寻找损失景观上的一个替代的损失等价点以减少冲突。具体而言,我们利用低秩适配器(LoRA),通过设计收敛的、损失不变的目标函数来实现这种实用的传送。此外,我们引入了一种历史轨迹重用策略,以持续利用先进优化器的优势。在多个主流数据集上的大量实验证明了我们方法的有效性。COST是一种即插即用的解决方案,能够增强广泛的现有MTL方法。当与最先进的方法结合时,COST实现了更优的性能。
英文摘要:
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.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。