Select before Act: Spatially Decoupled Action Repetition for Continuous Control

2025-02-10

AI Summary

Existing action repetition methods in continuous-control reinforcement learning treat all action dimensions holistically, ignoring dimensional heterogeneity—leading to rigid, insufficiently agile policies. Method: We propose a spatially decoupled action repetition framework that models, for each action dimension independently, a closed-loop “execute-or-repeat” decision, enabling the first dimension-level adaptive prediction of action duration and repetition control. Our approach integrates an RL-driven closed-loop selection mechanism with dimension-wise decoupled modeling. Contribution/Results: The framework significantly improves the trade-off between action persistence and diversity. Evaluated across diverse continuous-control benchmarks, it enhances sample efficiency and policy performance while reducing action jitter—demonstrating both the effectiveness and generalizability of spatial decoupling.

Technology Category

• Reasoning under Uncertainty: Sequential Decision Making
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Machine Learning: Reinforcement Learning

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Studies of user behavior, including longitudinal effects of personalized systems

Abstract

Reinforcement Learning (RL) has achieved remarkable success in various continuous control tasks, such as robot manipulation and locomotion. Different to mainstream RL which makes decisions at individual steps, recent studies have incorporated action repetition into RL, achieving enhanced action persistence with improved sample efficiency and superior performance. However, existing methods treat all action dimensions as a whole during repetition, ignoring variations among them. This constraint leads to inflexibility in decisions, which reduces policy agility with inferior effectiveness. In this work, we propose a novel repetition framework called SDAR, which implements Spatially Decoupled Action Repetition through performing closed-loop act-or-repeat selection for each action dimension individually. SDAR achieves more flexible repetition strategies, leading to an improved balance between action persistence and diversity. Compared to existing repetition frameworks, SDAR is more sample efficient with higher policy performance and reduced action fluctuation. Experiments are conducted on various continuous control scenarios, demonstrating the effectiveness of spatially decoupled repetition design proposed in this work.

Problem

• Decouples action repetition spatially
• Improves action persistence and diversity
• Enhances sample efficiency and performance

Innovation

• Spatially Decoupled Action Repetition
• Closed-loop act-or-repeat selection
• Improved balance persistence diversity