DAC: The Double Actor-Critic Architecture for Learning Options
This work addresses the challenge of efficient option learning in hierarchical reinforcement learning for robotics, offering a novel architecture that improves transfer performance.
The authors tackled the problem of learning options in hierarchical reinforcement learning by reformulating the option framework as two parallel augmented MDPs, enabling the use of standard policy optimization algorithms, and demonstrated that DAC outperforms hierarchy-free and previous gradient-based methods in transfer learning on robot simulation tasks.
We reformulate the option framework as two parallel augmented MDPs. Under this novel formulation, all policy optimization algorithms can be used off the shelf to learn intra-option policies, option termination conditions, and a master policy over options. We apply an actor-critic algorithm on each augmented MDP, yielding the Double Actor-Critic (DAC) architecture. Furthermore, we show that, when state-value functions are used as critics, one critic can be expressed in terms of the other, and hence only one critic is necessary. We conduct an empirical study on challenging robot simulation tasks. In a transfer learning setting, DAC outperforms both its hierarchy-free counterpart and previous gradient-based option learning algorithms.