CoRL 2020, Spotlight Talk 171: Safe Policy Learning for Continuous Control

“**Safe Policy Learning for Continuous Control**
Yinlam Chow (Google AI)*; Ofir Nachum (Google); Aleksandra Faust (Google Brain); Edgar Dueñez-Guzman (DeepMind); Mohammad Ghavamzadeh (Google Research)

We study continuous action reinforcement learning problems in which it is crucial that the agent interacts with the environment only through near-safe policies, i.e.,~policies that keep the agent in desirable situations, both during training and at convergence. We formulate these problems as constrained Markov decision processes (CMDPs) and present safe policy optimization algorithms that are based on a Lyapunov approach to solve them. Our algorithms can use any standard policy gradient (PG) method, such as deep deterministic policy gradient (DDPG) or proximal policy optimization (PPO), to train a neural network policy, while enforcing near-constraint satisfaction for every policy update by projecting either the policy parameter or the selected action onto the set of feasible solutions induced by the state-dependent linearized Lyapunov constraints. Compared to the existing constrained PG algorithms, ours are more data efficient as they are able to utilize both on-policy and off-policy data. Moreover, in practice our action-projection algorithm often leads to less conservative policy updates and allows for natural integration into an end-to-end PG training pipeline. We evaluate our algorithms and compare them with the state-of-the-art baselines on several simulated (MuJoCo) tasks, as well as a real-world robot obstacle-avoidance problem, demonstrating their effectiveness in terms of balancing performance and constraint satisfaction.

YouTube Source for this Robot AI Video

AI video(s) you might be interested in …