TAAC: Temporally Abstract Actor-Critic for Continuous Control
This addresses the problem of efficient exploration and policy optimization in continuous control RL, offering a novel approach to action repetition that yields competitive results.
The paper tackles the challenge of incorporating temporal abstraction into continuous control RL by proposing TAAC, an off-policy algorithm that adds a binary policy to choose between repeating previous actions or taking new ones, achieving top performance across 14 tasks and revealing that trained policies naturally 'mine' repeated actions even in continuous domains.
We present temporally abstract actor-critic (TAAC), a simple but effective off-policy RL algorithm that incorporates closed-loop temporal abstraction into the actor-critic framework. TAAC adds a second-stage binary policy to choose between the previous action and a new action output by an actor. Crucially, its "act-or-repeat" decision hinges on the actually sampled action instead of the expected behavior of the actor. This post-acting switching scheme let the overall policy make more informed decisions. TAAC has two important features: a) persistent exploration, and b) a new compare-through Q operator for multi-step TD backup, specially tailored to the action repetition scenario. We demonstrate TAAC's advantages over several strong baselines across 14 continuous control tasks. Our surprising finding reveals that while achieving top performance, TAAC is able to "mine" a significant number of repeated actions with the trained policy even on continuous tasks whose problem structures on the surface seem to repel action repetition. This suggests that aside from encouraging persistent exploration, action repetition can find its place in a good policy behavior. Code is available at https://github.com/hnyu/taac.