Guided Policy Optimization under Partial Observability
This addresses the problem of effectively using additional information in RL for partially observable environments, which is an incremental improvement over existing methods.
The paper tackles reinforcement learning in partially observable environments by introducing Guided Policy Optimization (GPO), a framework that co-trains a guider and learner to leverage privileged information, achieving optimality comparable to direct RL and significantly outperforming existing methods in tasks like continuous control and memory-based challenges.
Reinforcement Learning (RL) in partially observable environments poses significant challenges due to the complexity of learning under uncertainty. While additional information, such as that available in simulations, can enhance training, effectively leveraging it remains an open problem. To address this, we introduce Guided Policy Optimization (GPO), a framework that co-trains a guider and a learner. The guider takes advantage of privileged information while ensuring alignment with the learner's policy that is primarily trained via imitation learning. We theoretically demonstrate that this learning scheme achieves optimality comparable to direct RL, thereby overcoming key limitations inherent in existing approaches. Empirical evaluations show strong performance of GPO across various tasks, including continuous control with partial observability and noise, and memory-based challenges, significantly outperforming existing methods.