论文标题
自主驾驶系统中多任务网络的动态任务加权方法
Dynamic Task Weighting Methods for Multi-task Networks in Autonomous Driving Systems
论文作者
论文摘要
深度多任务网络对于自动驾驶系统特别感兴趣。他们可能会在预测性能，硬件约束和从多种类型的注释和方式中有效利用信息之间进行良好的权衡。但是，培训这种模型是不平凡的，需要在所有任务上平衡学习，因为它们各自的损失在整个培训中都表现出不同的规模，范围和动态。最近在不同数据集和任务组合上提出了以自适应方式调整损失的多种任务加权方法，因此很难比较它们。在这项工作中，我们在三个汽车数据集（Kitti，CityScapes和Woodscape）上共同审查并系统地评估了九种任务加权策略。然后，我们提出了一种新的方法，结合了进化元学习和基于任务的选择性反向流动，用于计算任务权重，从而进行可靠的网络培训。我们的方法在两任任务应用程序上的大幅度优于最先进的方法。
Deep multi-task networks are of particular interest for autonomous driving systems. They can potentially strike an excellent trade-off between predictive performance, hardware constraints and efficient use of information from multiple types of annotations and modalities. However, training such models is non-trivial and requires balancing learning over all tasks as their respective losses display different scales, ranges and dynamics across training. Multiple task weighting methods that adjust the losses in an adaptive way have been proposed recently on different datasets and combinations of tasks, making it difficult to compare them. In this work, we review and systematically evaluate nine task weighting strategies on common grounds on three automotive datasets (KITTI, Cityscapes and WoodScape). We then propose a novel method combining evolutionary meta-learning and task-based selective backpropagation, for computing task weights leading to reliable network training. Our method outperforms state-of-the-art methods by a significant margin on a two-task application.
