Modulated Policy Hierarchies
This addresses the problem of sparse rewards in reinforcement learning for robotics, offering an incremental improvement over existing hierarchical methods.
The paper tackles the challenge of sparse rewards in reinforcement learning by introducing modulated policy hierarchies (MPH), which learn end-to-end without manual intervention, and it outperforms baselines on robotics tasks like pushing and sparse block stacking.
Solving tasks with sparse rewards is a main challenge in reinforcement learning. While hierarchical controllers are an intuitive approach to this problem, current methods often require manual reward shaping, alternating training phases, or manually defined sub tasks. We introduce modulated policy hierarchies (MPH), that can learn end-to-end to solve tasks from sparse rewards. To achieve this, we study different modulation signals and exploration for hierarchical controllers. Specifically, we find that communicating via bit-vectors is more efficient than selecting one out of multiple skills, as it enables mixing between them. To facilitate exploration, MPH uses its different time scales for temporally extended intrinsic motivation at each level of the hierarchy. We evaluate MPH on the robotics tasks of pushing and sparse block stacking, where it outperforms recent baselines.