Distributionally-Constrained Policy Optimization via Unbalanced Optimal Transport
arXiv:2102.07889
Abstract
We consider constrained policy optimization in Reinforcement Learning, where the constraints are in form of marginals on state visitations and global action executions. Given these distributions, we formulate policy optimization as unbalanced optimal transport over the space of occupancy measures. We propose a general purpose RL objective based on Bregman divergence and optimize it using Dykstra's algorithm. The approach admits an actor-critic algorithm for when the state or action space is large, and only samples from the marginals are available. We discuss applications of our approach and provide demonstrations to show the effectiveness of our algorithm.
References in corpus (14)
- Trust Region Policy Optimization
- Equivalence Between Policy Gradients and Soft Q-Learning
- Stochastic Optimization for Large-scale Optimal Transport
- A unified view of entropy-regularized Markov decision processes
- AlgaeDICE: Policy Gradient from Arbitrary Experience
- Scaling Algorithms for Unbalanced Transport Problems
- Sinkhorn Distances: Lightspeed Computation of Optimal Transportation Distances
- Primal Wasserstein Imitation Learning
- Variational Policy Gradient Method for Reinforcement Learning with General Utilities
- Wasserstein Adversarial Imitation Learning
- Reinforcement Learning via Fenchel-Rockafellar Duality
- Policy Optimization as Wasserstein Gradient Flows
- Cautious Reinforcement Learning via Distributional Risk in the Dual Domain
- Imitation Learning with Sinkhorn Distances