TAAC: Temporally Abstract Actor-Critic for Continuous Control

TL;DR

TAAC introduces a temporally abstract actor-critic algorithm that enhances continuous control by incorporating action repetition and informed decision-making, leading to improved exploration and performance.

Contribution

The paper proposes TAAC, a novel off-policy RL algorithm with a second-stage binary policy for action repetition, improving exploration and policy quality in continuous control tasks.

Findings

TAAC outperforms strong baselines on 14 continuous control tasks.

TAAC discovers significant action repetition even in complex tasks.

Action repetition can be beneficial in continuous control policies.

Abstract

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.