Newest papers

Post-hoc Epistemic Uncertainty for Vision-Language-Action Policies

Johannes Keller, Albrecht Bloße, Adrian Lindenmeyer, Thomas Neumuth, Daniel Schneider · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.29730 · cited by 0 works (OpenAlex) · open-access PDF

Japanese

ビジョン・ジャンル・アクション(VLA)は、新しい状況で有用なデモや援助を求めるための訓練の知識のギャップを認識する必要があります、特に失敗が費用がかかる場合。独立して訓練されたポリシーのセット内の不一致は、これらの決定を導くことができますが、トレーニングや複数のポリシーの実行を必要とします。私たちは、単一のポリシーの組み込みからエピステミック不確実性を推定するポストホックカーネル変数ヘッドを提案します。それぞれのヘッドは、トレーニング・デモからの組み込みに基づいて評価されたガウシアンプロセスの条件変数を計算します。フィッティングは、回帰目標やグレードアップの不適切なアップデートを必要とし、ポリシー

Machine translation (Meta M2M100); the English original follows.

Abstract

Vision-language-action (VLA) policies need to recognize gaps in their training knowledge to request useful demonstrations or assistance in new situations, especially when failures are costly. Disagreement within an ensemble of independently trained policies can guide these decisions, but requires training and running multiple policies. We propose post-hoc kernel-variance heads that estimate epistemic uncertainty from the embeddings of a single policy. Each head computes the conditional variance of a Gaussian process conditioned on embeddings from the training demonstrations. Fitting requires no regression targets or gradient updates and leaves the policy unchanged. In experiments with SmolVLA on LIBERO-10, the kernel variance heads perform competitively with two-policy ensemble scores in task-level correlation with task success, in demonstration acquisition during active learning, and in referral to a stronger supervisor policy as a proxy for human assistance. For SmolVLA, the head on the vision-language backbone reduces uncertainty-scoring time by up to approximately a factor of 50 compared with ensemble scores, while also substantially reducing setup costs. We also show that the established evaluation protocol on LIBERO-10 primarily rewards recognition of coverage differences between tasks, while execution variability weakens assessment of within-task rankings. Thus, controlled perturbations provide a complementary evaluation: modality-specific heads identify unfamiliar camera, instruction, or state inputs, while fused heads can detect unfamiliar combinations of individually familiar inputs. With a supervisor proxy trained on the evaluated perturbation families, the decoder head improves within-task referral over ensemble disagreement. These findings show that useful uncertainty estimates for adaptive robot learning can be obtained from a single policy.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Supervisor, Computer science, Variance (accounting), Artificial intelligence, Proxy (statistics), Machine learning

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange