pith. sign in

arxiv: 2602.21445 · v2 · pith:JFBKXMSSnew · submitted 2026-02-24 · 💻 cs.RO

VLA Knows Its Limits: Adaptive Execution Horizons for Robot Policies

classification 💻 cs.RO
keywords actionexecutionhorizonactionsflow-basedacrossautohorizonchunk
0
0 comments X
read the original abstract

Action chunking has recently emerged as a standard practice in flow-based Vision-Language-Action (VLA) models. However, the effect and choice of the execution horizon - the number of actions to be executed from each predicted chunk - remains underexplored. In this work, we first show that varying the execution horizon leads to substantial performance deviations, with performance initially improving and then declining as the horizon increases. To uncover the reasons, we analyze the cross- and self-attention weights in flow-based VLAs and reveal two key phenomena: (i) intra-chunk actions attend invariantly to vision-language tokens, limiting adaptability to environmental changes; and (ii) the initial and terminal action tokens serve as stable anchors, forming latent centers around which intermediate actions are organized. Motivated by these insights, we interpret action self-attention weights as a proxy for the model's predictive limit and propose AutoHorizon, the first test-time method that dynamically estimates the execution horizon for each predicted action chunk to adapt to changing perceptual conditions. Across simulated and real-world robotic manipulation tasks, AutoHorizon is performant, incurs negligible computational overhead, and generalizes across diverse tasks and flow-based models.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 12 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Denoising Tells When to Replan: Denoising-Variance Adaptive Chunking for Flow-Based Robot Policies

    cs.RO 2026-06 unverdicted novelty 7.0

    DVAC uses denoising variance as an intrinsic signal to adaptively chunk actions in flow-based robot policies, improving success rates and cutting replans on LIBERO, RoboTwin, CALVIN, and real-world tasks.

  2. Dynamic Execution Commitment of Vision-Language-Action Models

    cs.CV 2026-05 unverdicted novelty 7.0

    A3 reframes dynamic action chunk commitment in VLA models as self-speculative prefix verification, accepting the longest continuous sequence of actions that satisfies consensus-ordered conditional invariance and prefi...

  3. Dynamic Execution Commitment of Vision-Language-Action Models

    cs.CV 2026-05 unverdicted novelty 7.0

    A3 determines the execution horizon in VLA models as the longest prefix of actions that passes consensus-based verification and sequential consistency checks.

  4. AutoSpeed: Annotation-Free Stage-Adaptive Motion Speed Learning for Robot Manipulation

    cs.RO 2026-07 unverdicted novelty 6.0

    AutoSpeed optimizes visuomotor policies over candidate trajectories at varying speeds using a composite cost of prediction error versus horizon length, with DCT-based modulation, yielding shorter execution times and h...

  5. PolicyTrim: Boosting Intrinsic Policy Efficiency of Vision-Language-Action Models

    cs.CV 2026-06 unverdicted novelty 6.0

    PolicyTrim is an RL post-training framework that boosts VLA policy efficiency by 3x chunk utilization and 51.4% fewer steps, yielding up to 5.83x speedup.

  6. $\mu$VLA: On Recurrent Memory for Partially Observable Manipulation in VLA Models

    cs.LG 2026-06 unverdicted novelty 6.0

    Adding recurrent memory tokens to VLA models raises success rates on partially observable manipulation tasks from 0.42 to 0.84 on training and 0.07 to 0.23 on held-out tasks while preserving performance under full obs...

  7. Learning to Assist: Collaborative VLAs for Implicit Human-Robot Collaboration

    cs.RO 2026-06 conditional novelty 6.0

    VLA models with inference-time steering mitigate action leakage in implicit human-robot collaboration, supporting longer horizons and yielding faster, more reliable assembly than shorter-horizon baselines in a 16-pers...

  8. PACE: Phase-Aware Chunk Execution for Robot Policies with Action Chunking

    cs.RO 2026-05 unverdicted novelty 6.0

    PACE dynamically selects execution horizons for action chunks in robot policies by detecting low-speed transition points in predicted speed profiles, raising success rates from 57.8% to 64.2% on 50 simulation tasks an...

  9. When to Trust Imagination: Adaptive Action Execution for World Action Models

    cs.RO 2026-05 unverdicted novelty 6.0

    Future Forward Dynamics Causal Attention (FFDC) enables World Action Models to adaptively choose action chunk lengths based on prediction-observation consistency, cutting model inferences by 69% and improving real-wor...

  10. When to Trust Imagination: Adaptive Action Execution for World Action Models

    cs.RO 2026-05 unverdicted novelty 6.0

    A verifier called Future Forward Dynamics Causal Attention enables adaptive action execution in World Action Models, reducing model inferences by 69% and improving success rates in robotic tasks.

  11. Position: Vision-Language-Action Models Cannot Be Verified to Perform Physical Reasoning

    cs.RO 2026-06 conditional novelty 5.0

    VLA benchmark success rates cannot distinguish semantic generalization from physical reasoning due to an identifiability gap in current evaluation protocols.

  12. Dynamic Execution Commitment of Vision-Language-Action Models

    cs.CV 2026-05 unverdicted novelty 5.0

    A3 adaptively selects verifiable action prefixes in VLA models using group-sampled consensus and conditional re-decoding to balance robustness and speed without manual horizon tuning.