通过锥形约束动态调节梯度方向，解决多任务学习中的冲突问题。
Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone
- 引入角度约束，将梯度更新限制在参考梯度附近的锥形区域内。
- 在多个监督与强化学习基准上达到当前最优性能。
- 方法高效可扩展，适合高维参数空间的多任务场景。
多任务学习中平衡相互竞争的目标仍是一个根本性挑战，主要源于各任务间梯度的冲突。现有方法通常通过计算动态梯度更新向量来随优化进程平衡不同任务。基于此思路，我们提出ConicGrad，一种原则性强、可扩展且鲁棒的多任务学习方法，将其建模为带约束的优化问题。该方法引入角度约束，动态调节梯度更新方向，使其被限制在以整体目标参考梯度为中心的锥形区域内。通过在不过度约束方向或大小的情况下平衡任务特定梯度，ConicGrad有效缓解了任务间的梯度冲突。此外，该框架确保了计算效率，并可扩展至高维参数空间。我们在标准监督学习和强化学习多任务基准上进行了广泛实验，结果表明ConicGrad在多种任务上均实现了当前最优性能。
原文摘要 · Abstract (English)
Balancing competing objectives remains a fundamental challenge in multi-task learning (MTL), primarily due to conflicting gradients across individual tasks. A common solution relies on computing a dynamic gradient update vector that balances competing tasks as optimization progresses. Building on this idea, we propose ConicGrad, a principled, scalable, and robust MTL approach formulated as a constrained optimization problem. Our method introduces an angular constraint to dynamically regulate gradient update directions, confining them within a cone centered on the reference gradient of the overall objective. By balancing task-specific gradients without over-constraining their direction or magnitude, ConicGrad effectively resolves inter-task gradient conflicts. Moreover, our framework ensures computational efficiency and scalability to high-dimensional parameter spaces. We conduct extensive experiments on standard supervised learning and reinforcement learning MTL benchmarks, and demonstrate that ConicGrad achieves state-of-the-art performance across diverse tasks.
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。