frozen-vla-policies-executing-repetitive-perceptually-aliased-manipulation-tasks CLM-0395For frozen VLA policies executing repetitive, perceptually aliased manipulation tasks (counting pick-and-place), a progress memory that advances after each attempted action converts transient grasp or placement failures into persistent task-state errors, so writes must be gated on observed physical evidence of the intended transition; gating this way raised success far above the frozen policy and above memory-augmented baselines.
Capability: Turning plans into physical action
Observed on
Repetitive counting manipulation where the current image cannot reveal which repetition is underway; subgoals anchored to gripper open/close events; verified in RoboMME Counting si.
Sources
- Validated by: Success rate on RoboMME Counting 50-episode official test split: 100.0% PickXTimes and 84.0% BinFill, four-task average 55.96% vs 48.83% for MemER and 28.78% for frozen pi-0.5; past-action conditioning gave 29.09%. Per-count breakdown shows frozen policy collapsing to 0% by N=5 while AGM holds. Real robot: 100.0% and 82.0%. Ablations: proprioception-only grasp check 98.0/78.0; flattened thresholds 98.0/74.0.. [truncated]
Disagreeing is the most useful thing you can do here. Nothing on this page waits for a reviewer: the counts above move when evidence lands, and every source stays attached with who brought it.
Related claims
- For memory-augmented VLA manipulation policies that must actively move the viewpoint to find task-relevant evidence, writing one memory frame at each fixed action-chunk boundary generalizes better than event-triggered writing (on predicted subtask change or on gripper/head/torso motion), and extra memory capacity helps only under the regular schedule — event-triggered policies degrade further out of distribution because mistimed triggers omit informative views and control jitter fills memory with redundant frames.Turning plans into physical action · reviewed by AI
- For manipulation policies whose large VLA backbone cannot run within the control period, letting the slow model propose action chunks from a stale observation while a small RL-trained edit policy adjusts and Q-selects among those chunks using the latest observation recovers more performance than applying RL on top of asynchronous imitation-learning schemes alone, and the gap widens as inference delay or environment speed grows.Turning plans into physical action · reviewed by AI
- On language-conditioned manipulation tasks where the same base task is instantiated at increasing levels of spatial ambiguity and procedural length, current VLA models degrade on fine-grained spatial relations, precise low-level execution, and memory-dependent multi-step planning, so success measured only on predefined simple settings overstates their competence.Turning plans into physical action · reviewed by AI
- When adapting a pretrained VLA manipulation policy with only a handful of target demonstrations, retrieving prior demonstrations at the level of LLM-decomposed subtasks (then reranking by behavior-feature similarity) raises success more than retrieval by visual, optical-flow, or state-action similarity, and the margin is concentrated on long-horizon composite tasks rather than single-skill pick-and-place.Turning plans into physical action · reviewed by AI
- For VLA and world-action manipulation policies that score near-ceiling on standard LIBERO, recovery success after a real execution failure falls sharply as recovery requires deeper state reasoning \u2014 retrying or adapting the immediate action works far more often than restoring a displaced task object or clearing an environmental obstruction \u2014 and rankings on the standard benchmark do not predict recovery rankings.Turning plans into physical action · reviewed by AI
Notes
Drafted from the paper by a model and filed unreviewed. Visible here so it can be read, not because anyone has vouched for it: it does not move any technique's standing and does not count toward the internal scorecard. Drafted confidence: medium. Falsifier as drafted: An attempt-based progress memory matching or beating verification-gated writing on the same frozen policy and subgoal interface, or verification gating failing to improve over no progress memory on comparable repetitive manipulation tasks.