Timezone: »

Gradient Surgery for Multi-Task Learning
Tianhe Yu · Saurabh Kumar · Abhishek Gupta · Sergey Levine · Karol Hausman · Chelsea Finn

Wed Dec 09 09:00 PM -- 11:00 PM (PST) @ Poster Session 4 #1208

While deep learning and deep reinforcement learning (RL) systems have demonstrated impressive results in domains such as image classification, game playing, and robotic control, data efficiency remains a major challenge. Multi-task learning has emerged as a promising approach for sharing structure across multiple tasks to enable more efficient learning. However, the multi-task setting presents a number of optimization challenges, making it difficult to realize large efficiency gains compared to learning tasks independently. The reasons why multi-task learning is so challenging compared to single-task learning are not fully understood. In this work, we identify a set of three conditions of the multi-task optimization landscape that cause detrimental gradient interference, and develop a simple yet general approach for avoiding such interference between task gradients. We propose a form of gradient surgery that projects a task's gradient onto the normal plane of the gradient of any other task that has a gradient. On a series of challenging multi-task supervised and multi-task RL problems, this approach leads to substantial gains in efficiency and performance. Further, it is model-agnostic and can be combined with previously-proposed multi-task architectures for enhanced performance.

Author Information

Tianhe (Kevin) Yu (Stanford University)
Saurabh Kumar (Stanford University)
Abhishek Gupta (UC Berkeley)
Sergey Levine (UC Berkeley)
Karol Hausman (Google Brain)
Chelsea Finn (Stanford)

More from the Same Authors