Improving Generative Behavior Cloning via Self-Guidance and Adaptive Chunking

TL;DR

The paper tackles the fidelity and reactivity bottlenecks of Generative Behavior Cloning using diffusion policies under open-loop control. It introduces Self Guidance, which exploits the model’s own past outputs as negative guidance to sharpen action distributions and enable forward-looking adaptation, and Adaptive Chunking, which selectively replans based on action similarity to balance reactivity with temporal consistency. Across simulation and real-world robotic manipulation tasks, the approach yields substantial performance gains over Vanilla Diffusion Policy and BID, while reducing computational burden relative to heavy inference strategies. The methods generalize to Vision-Language-Action models, suggesting broad applicability to real-world robotic planning under uncertainty.

Abstract

Generative Behavior Cloning (GBC) is a simple yet effective framework for robot learning, particularly in multi-task settings. Recent GBC methods often employ diffusion policies with open-loop (OL) control, where actions are generated via a diffusion process and executed in multi-step chunks without replanning. While this approach has demonstrated strong success rates and generalization, its inherent stochasticity can result in erroneous action sampling, occasionally leading to unexpected task failures. Moreover, OL control suffers from delayed responses, which can degrade performance in noisy or dynamic environments. To address these limitations, we propose two novel techniques to enhance the consistency and reactivity of diffusion policies: (1) self-guidance, which improves action fidelity by leveraging past observations and implicitly promoting future-aware behavior; and (2) adaptive chunking, which selectively updates action sequences when the benefits of reactivity outweigh the need for temporal consistency. Extensive experiments show that our approach substantially improves GBC performance across a wide range of simulated and real-world robotic manipulation tasks. Our code is available at https://github.com/junhyukso/SGAC

Figures (18)

• Figure 1: Illustration of our Self Guidance(SG). By using the past state distribution as negative guidance, SG effectively sharpens the distribution or proactively reacts to environmental perturbations.
• Figure 2: (a) Visualization of the action distribution of Diffusion Policy (DP) chi2023diffusion. (b) The sharpened distribution after applying our Self-Guidance(SG). (c) Their respective performances. Standard DP often generates low-fidelity actions, which can harm sequential control performance.
• Figure 3: Effect of SG guidance scale($w$) on varying noise levels (P)
• Figure 4: Similarity between actions from a previously planned chunk and newly replanned actions at each time step. The similarity tends to be high during simple movements (e.g., moving , transporting). Conversely, it tends to be low when high precision is required (e.g., attempting to grasp).
• Figure 5: Simulation Experiments : Stochastic(top) & Static(bottom) : Performance comparison in the 6 simulated environment. Results are averaged over 100 episodes across three random seeds.