# Action Chunking and Exploratory Data Collection Yield Exponential Improvements in Behavior Cloning for Continuous Control

arXiv: https://arxiv.org/abs/2507.09061 · HTML: https://mlsanity.com/paper/2507.09061/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data.
- First posted on arXiv 11 Jul 2025.

## Placed in

- Offline RL and imitation learning from data / [Robot manipulation imitation learning](https://mlsanity.com/offlinerl/robot-manipulation-imitation-learning/)

## Built on by (most cited first, 15 of 15)

- [Much Ado About Noising: Dispelling the Myths of Generative Robotic Control](https://mlsanity.com/paper/2512.01809/) (2025)
- [OAT: Ordered Action Tokenization](https://mlsanity.com/paper/2602.04215/) (2026)
- [Turning Video Models into Generalist Robot Policies](https://mlsanity.com/paper/2605.27817/) (2026)
- [From Prior to Pro: Efficient Skill Mastery via Distribution Contractive RL Finetuning](https://mlsanity.com/paper/2603.10263/) (2026)
- [Demystifying Action Space Design for Robotic Manipulation Policies](https://mlsanity.com/paper/2602.23408/) (2026)
- [πR2: Reactive Real-time Flow Policies](https://mlsanity.com/paper/2607.26055/) (2026)
- [MePoly: Max Entropy Polynomial Policy Optimization](https://mlsanity.com/paper/2602.17832/) (2026)
- [Ordered Action Tokens for Visuomotor Policy Learning](https://mlsanity.com/paper/2607.21670/) (2026)
- [Three Steps at a Time: Learning Representations from Action Sequences in Contrastive RL](https://mlsanity.com/paper/2608.30640/) (2026)
- [Overcoming Dynamics-Blindness: Training-Free Pace-and-Path Correction for VLA Models](https://mlsanity.com/paper/2605.11459/) (2026)
- [FocalPolicy: Frequency-Optimized Chunking and Locally Anchored Flow Matching for Coherent Visuomotor Policy](https://mlsanity.com/paper/2605.15944/) (2026)
- [Spatial Attention: Adapting Execution Horizons for Diffusion Policies via Observation Sensitivity](https://mlsanity.com/paper/2607.04739/) (2026)
- [Action Chunking Proximal Policy Optimization with Feedback Correction](https://mlsanity.com/paper/2609.36250/) (2026)
- [SplineWAM: Adaptive Action Horizons for World Action Models via B-Spline Representations](https://mlsanity.com/paper/2609.39873/) (2026)
- [Measuring the Stability Assumption Behind Action Chunking](https://mlsanity.com/paper/2610.01626/) (2026)

## Abstract (arXiv)

This paper presents a theoretical analysis of two of the most impactful interventions in modern learning from demonstration in robotics and continuous control: the practice of action-chunking (predicting sequences of actions in open-loop) and exploratory augmentation of expert demonstrations. Though recent results show that learning from demonstration, also known as imitation learning (IL), can suffer errors that compound exponentially with task horizon in continuous settings, we demonstrate that action chunking and exploratory data collection circumvent exponential compounding errors in different regimes. Our results identify control-theoretic stability as the key mechanism underlying the benefits of these interventions. On the empirical side, we validate our predictions and the role of control-theoretic stability through experimentation on popular robot learning benchmarks. On the theoretical side, we demonstrate that the control-theoretic lens provides fine-grained insights into how compounding error arises, leading to tighter statistical guarantees on imitation learning error when these interventions are applied than previous techniques based on information-theoretic considerations alone.