aiqbench analysis
This arXiv preprint proposes a lightweight predictor that decides when a navigation vision-language-action model can reuse its previous command instead of running costly semantic reasoning again. The reported gains are modest: 9.5% fewer VLA calls and 8.1% lower estimated inference time, with comparable navigation performance on one benchmark (RxR-CE). Readers should note that the results are from a single frozen base model and simulated evaluation, and that the "estimated" timing does not reflect measured on-robot latency.
ASPEN: Adaptive Semantic Reasoning via Command Persistence Prediction for Efficient Navigation VLAs
arXiv:2610.36285v1 Announce Type: new Abstract: Navigation vision-language-action (VLA) policies repeatedly invoke costly semantic reasoning even when the next command remains unchanged. Command reuse can reduce computation but risks missing instruction-dependent turns or stops. We introduce ASPEN, an event-triggered scheduler based on command persistence prediction: estimating whether the frozen base policy would reissue its previous command. A lightweight multimodal predictor combines temporal observations, language instructions, execution history, and cached VLA confidence. At eligible command boundaries, sufficient predicted persistence permits command reuse. Otherwise, ASPEN triggers fresh semantic reasoning from the VLA. Evaluated with frozen NaVILA on cross-dataset RxR-CE val-unseen, ASPEN reduces VLA calls by 9.5% and estimated inference time by 8.1%, including scheduling overhead. It maintains comparable navigation performance and preserves identical baseline action sequences on 92.1% of episodes.
Abstract:Navigation vision-language-action (VLA) policies repeatedly invoke costly semantic reasoning even when the next command remains unchanged. Command reuse can reduce computation but risks missing instruction-dependent turns or stops. We introduce ASPEN, an event-triggered scheduler based on command persistence prediction: estimating whether the frozen base policy would reissue its previous command. A lightweight multimodal predictor combines temporal observations, language instructions, execution history, and cached VLA confidence. At eligible command boundaries, sufficient predicted persistence permits command reuse. Otherwise, ASPEN triggers fresh semantic reasoning from the VLA. Evaluated with frozen NaVILA on cross-dataset RxR-CE val-unseen, ASPEN reduces VLA calls by 9.5% and estimated inference time by 8.1%, including scheduling overhead. It maintains comparable navigation performance and preserves identical baseline action sequences on 92.1% of episodes.
Robotics (cs.RO)
arXiv:2610.36285 [cs.RO]
(or arXiv:2610.36285v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2610.36285 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 215 words
Related coverage
LandscapeSHAP: Which Persistent Homology Class Gets the Credit?
arXiv · Sep 28, 2026
Age of Learning: Temporal Persistence of Prediction Errors as a Learning Signal
arXiv · Sep 30, 2026
Exact Action Values Are Not Enough: Rollout-Verified Reinforcement Fine-Tuning of a Reasoning Model for Multi-Zone VAV Control
arXiv · Jul 31, 2026
Predictive Safety Guardrails Do Not Always Make LLM Agents Safer
arXiv · Nov 4, 2026