In Multi-Task Learning, tasks may compete and limit the performance achieved on each other rather than guiding the optimization trajectory to a common solution, superior to its single-task counterparts. There is often not a single solution that is optimal for all tasks, leading practitioners to balance tradeoffs between tasks' performance, and to resort to optimality in the Pareto sense. Current Multi-Task Learning methodologies either completely neglect this aspect of functional diversity, and produce one solution in the Pareto Front predefined by their optimization schemes, or produce diverse but discrete solutions, each requiring a separate training run. In this paper, we conjecture that there exist Pareto Subspaces, i.e., weight subspaces where multiple optimal functional solutions lie. We propose Pareto Manifold Learning, an ensembling method in weight space that is able to discover such a parameterization and produces a continuous Pareto Front in a single training run, allowing practitioners to modulate the performance on each task during inference on the fly. We validate the proposed method on a diverse set of multi-task learning benchmarks, ranging from image classification to tabular datasets and scene understanding, and show that Pareto Manifold Learning outperforms state-of-the-art algorithms.
翻译:在多任务学习中,任务可能相互竞争,限制彼此实现的绩效,而不是将优化轨迹引导为共同的解决方案,优于单一任务对应方。常常没有一种对所有任务最合适的单一解决方案,让从业者平衡任务业绩之间的权衡,并采用Pareto意义的最佳方法。当前的多任务学习方法要么完全忽视功能多样性的这一方面,在Pareto Front中产生一种由其优化计划预先界定的解决方案,要么在Pareto Front中产生一种不同的但互不关联的解决方案,每个解决方案都需要单独的培训运行。在本文中,我们推测存在Pareto Subspace,即具有多重最佳功能解决方案的重量子空间。我们建议采用Pareto Manifold Learning,这是在重量空间中的一种组合方法,能够发现这种参数化,并在单一的培训运行中产生持续的Pareto Front,让从业者在对飞行的推论期间调整每一项任务的业绩。我们验证了一套关于多种任务学习基准的拟议方法,从图像分类到表格式的矩阵和图像式算法,显示从图像格式到图表式的矩阵到图表式的矩阵。