Adaptive Action Chunking via Multi-Chunk Q Value Estimation

TL;DR

This paper introduces ACH, an adaptive RL algorithm that dynamically adjusts action chunk length using a Transformer-based approach, improving performance and generalization across diverse tasks.

Contribution

It presents a novel method for adaptive action chunking in RL that estimates all chunk lengths simultaneously, overcoming fixed-length limitations.

Findings

ACH outperforms fixed-length baselines on 34 tasks.

The method improves generalization and learning efficiency.

Adaptive chunking yields more cohesive action sequences.

Abstract

Action chunking emerged as a pivotal technique in imitation learning, enabling policies to predict cohesive action sequences rather than single actions. Recently, this approach has expanded to reinforcement learning (RL), enhancing behavioral consistency and reducing bootstrapping errors in value function estimation. However, existing methods rely on a fixed chunk length, creating a performance bottleneck as the optimal length varies across states and tasks. In this paper, we propose Adaptive Action CHunking (ACH), a novel offline-to-online RL algorithm that dynamically modulates chunk length during both training and inference. To find the optimal chunk length for a dynamically varying current state, we simultaneously estimate action-values for all candidate chunk lengths in a single forward pass, using a Transformer-based architecture. Our mechanism allows the agent to select the most effective chunk length adaptively based on the current state. Evaluated on 34 challenging tasks, ACH consistently outperforms fixed-length baselines, demonstrating superior generalization and learning efficiency in complex environments.