从神经正切核视角缓解多任务学习中的任务不平衡问题。
NTKMTL: Mitigating Task Imbalance in Multi-Task Learning from Neural Tangent Kernel Perspective
- 基于神经正切核理论分析训练动态，设计新方法平衡多任务收敛速度。
- 在多个监督与强化学习基准上达到顶尖性能，提升训练效率。
- 适合关注多任务学习优化、模型训练稳定性的研究者参考。
多任务学习（MTL）使单一模型能同时学习多个任务，通过任务间知识迁移提升泛化能力，已广泛应用于各领域。然而，任务不平衡仍是主要挑战。尽管平衡不同任务的收敛速度是有效策略，但在复杂的MTL系统中准确刻画训练动态和收敛速度极具难度。为此，本文借助神经正切核（NTK）理论分析MTL训练动态，提出新方法NTKMTL。具体地，引入适用于MTL的扩展NTK矩阵，并通过谱分析平衡多任务收敛速度，缓解任务不平衡。基于共享表示近似，进一步提出NTKMTL-SR，在保持优异性能的同时提升训练效率。大量实验表明，该方法在多种基准上（涵盖多任务监督学习与强化学习）均达到领先水平。源代码见https://github.com/jianke0604/NTKMTL。
原文摘要 · Abstract (English)
Multi-Task Learning (MTL) enables a single model to learn multiple tasks simultaneously, leveraging knowledge transfer among tasks for enhanced generalization, and has been widely applied across various domains. However, task imbalance remains a major challenge in MTL. Although balancing the convergence speeds of different tasks is an effective approach to address this issue, it is highly challenging to accurately characterize the training dynamics and convergence speeds of multiple tasks within the complex MTL system. To this end, we attempt to analyze the training dynamics in MTL by leveraging Neural Tangent Kernel (NTK) theory and propose a new MTL method, NTKMTL. Specifically, we introduce an extended NTK matrix for MTL and adopt spectral analysis to balance the convergence speeds of multiple tasks, thereby mitigating task imbalance. Based on the approximation via shared representation, we further propose NTKMTL-SR, achieving training efficiency while maintaining competitive performance. Extensive experiments demonstrate that our methods achieve state-of-the-art performance across a wide range of benchmarks, including both multi-task supervised learning and multi-task reinforcement learning. Source code is available at https://github.com/jianke0604/NTKMTL.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。