Feudal Graph Reinforcement Learning

Graph-based representations and weight-sharing modular policies constitute prominent approaches to tackling composable control problems in Reinforcement Learning (RL). However, as shown by recent graph deep learning literature, message-passing operators can create bottlenecks in information propagation and hinder global coordination. The issue becomes dramatic in tasks where high-level planning is needed. In this work, we propose a novel methodology, named Feudal Graph Reinforcement Learning (FGRL), that addresses such challenges by relying on hierarchical RL and a pyramidal message-passing architecture. In particular, …

bottlenecks control cs.lg deep learning global graph graph-based hinder information issue literature modular operators planning propagation reinforcement reinforcement learning tasks work