One approach to deal with the statistical inefficiency of neural networks is to rely on auxiliary losses that help to build useful representations. However, it is not always trivial to know if an auxiliary task will be helpful for the main task and when it could start hurting. We propose to use the cosine similarity between gradients of tasks as an adaptive weight to detect when an auxiliary loss is helpful to the main loss. We show that our approach is guaranteed to converge to critical points of the main task and demonstrate the practical usefulness of the proposed algorithm in a few domains: multi-task supervised learning on subsets of ImageNet, reinforcement learning on gridworld, and reinforcement learning on Atari games.

翻译：解决神经网络统计效率低下问题的一种办法是依靠辅助损失来建立有用的表达方式,然而,知道辅助任务是否对主要任务有帮助以及何时开始造成伤害,并非总是微不足道的。我们提议使用任务梯度的相似性作为适应权重,以便在辅助损失对主要损失有帮助时进行检测。我们表明,我们的方法有保证会与主要任务的关键点汇合,并表明拟议的算法在几个领域的实际效用:在图像网络子集上进行多任务监督学习,加强在网格世界上的学习,加强在Atari游戏上的学习。