PACT-VLA: Speculative Action-Chunk Verification with Pre-Commit Observations
Overview
This paper presents PACT-VLA, a training-free framework that improves the success rate of vision-language-action (VLA) policies by using pre-commit observations to select the best action chunk during inference delay. It outperforms existing methods on both simulated and real-robot tasks, achieving significant improvements in task success rates.
From arXiv Robotics — research abstracts
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Reported quantity · delay steps: 4 other · Basis not reported
“ity predicted by the frozen policy under the pre-commit observation, requiring no separate verifier. A pipelined implementation overlaps observation processing with candidate generation and batches scoring, so selection adds no control-step delay on the evaluated hardware. At a four-step delay, PACT-VLA improves average success by 15.0 percentage points over RTC on Kinetix and by 5.2 points over single-candidate asyn”
Exact source · revision 1Source owner not reported
Evidence dependency changes (2)
- 2026-10-30T21:58:52.679Z · evidence updated · source revision 1. Evidence extraction was updated; current source attributions are shown above.
- 2026-10-30T21:58:33.868Z · evidence updated · source revision 1. Evidence extraction was updated; current source attributions are shown above.
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsSignalPACT-VLA: Speculative Action-Chunk Verification with Pre-Commit Observations
This paper presents PACT-VLA, a training-free framework that improves the success rate of vision-language-action (VLA) policies by using pre-commit observations to select the best action chunk during inference delay. It outperforms existing methods on both simulated and real-robot tasks, achieving significant improvements in task success rates.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).