Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Loss is its own Reward: Self-Supervision for Reinforcement Learning

About

Reinforcement learning optimizes policies for expected cumulative reward. Need the supervision be so narrow? Reward is delayed and sparse for many tasks, making it a difficult and impoverished signal for end-to-end optimization. To augment reward, we consider a range of self-supervised tasks that incorporate states, actions, and successors to provide auxiliary losses. These losses offer ubiquitous and instantaneous supervision for representation learning even in the absence of reward. While current results show that learning from reward alone is feasible, pure reinforcement learning methods are constrained by computational and data efficiency issues that can be remedied by auxiliary losses. Self-supervised pre-training and joint optimization improve the data efficiency and policy returns of end-to-end reinforcement learning.

Evan Shelhamer, Parsa Mahmoudieh, Max Argus, Trevor Darrell• 2016

Related benchmarks

TaskDatasetResultRank
ControlDMControl
DMControl: Ball in Cup Catch Score871
29
Reinforcement LearningAtari 7-game suite ALE (test)
Relative Score1.692
13
Showing 2 of 2 rows

Other info

Follow for update