Reuse Before You Retrieve: Diagnosing Headroom and Complementarity for Test-Time Augmentation of Embodied Multimodal Policies

• URL: http://arxiv.org/abs/2608.17484v1
• Date: Tue, 18 Aug 2026 08:11:31 GMT
• Title: Reuse Before You Retrieve: Diagnosing Headroom and Complementarity for Test-Time Augmentation of Embodied Multimodal Policies
• Abstract summary: Frozen-language-action (VLA) policies are increasingly improved at test time.<n>Additional sampling is useful only when better behavior already exists within the policy's rollouts.<n>Retrieval is most useful when the relevant action prior is not reliably represented by the policy.
• Score: 56.86456077356948
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Frozen vision-language-action (VLA) policies are increasingly improved at test time by sampling additional policy behaviors or introducing external demonstrations. Yet there is little guidance for deciding which intervention a deployed policy actually needs. Additional sampling is useful only when better behavior already exists within the policy's stochastic rollouts and can be identified, whereas retrieval is most useful when the relevant action prior is not reliably represented by the policy. We study this decision through two measurable factors, recoverable headroom and retrieval complementarity, which characterize how much useful behavior is already available to recover and whether an external action prior fills a measurable gap. We evaluate an episode-level retry selector under retryable or parallel execution, together with retrieval across multiple frozen VLA policies and environments. The selector consistently recovers substantial latent capability across all tested VLA backbones on LIBERO, with gains of up to 21.0 success-rate points that closely track recoverable headroom. It also transfers to a different robot and simulator and remains effective under degraded observations, while experiments with autoregressive OpenVLA illustrate the distinction between available headroom and the ability to rank candidate rollouts. Retrieval behaves differently, improving the policy with the largest measured action-prior gap and providing further gains when combined with selection. Together, these results provide an empirical basis for characterizing test-time augmentation opportunities by separating capability that can be recovered from the frozen policy from behavioral priors that may need to be introduced externally.

Related papers

• SEED: Self-Evolving On-Policy Distillation for Agentic Reinforcement Learning [38.54413500261524]
  Outcome-based reinforcement learning (RL) provides a practical optimization paradigm, but its sparse trajectory-level rewards offer limited guidance on intermediate decisions.<n>We propose SEED (SElf-Evolving On-Policy Distillation), a self-evolving framework that converts completed on-policy trajectories into training-time hindsight skills.
  arXiv  Detail & Related papers  (2026-07-16T09:57:18Z)
• Experience Augmented Policy Optimization for LLM Reasoning [50.16045797892008]
  We argue that experience in RLVR should not be reused as fixed reasoning trajectories, but instead expressed in a policy-adaptive manner.<n>EAPO consistently improves reasoning performance over state-of-the-art RLVR methods.
  arXiv  Detail & Related papers  (2026-06-29T15:05:28Z)
• Value Function Decomposition in Markov Recommendation Process [19.082512423102855]
  We propose an online reinforcement learning framework to improve recommender performance.<n>We show that these two factors can be separately approximated by decomposing the original temporal difference loss.<n>The disentangled learning framework can achieve a more accurate estimation with faster learning and improved robustness against action exploration.
  arXiv  Detail & Related papers  (2025-01-29T04:22:29Z)
• Bidirectional Decoding: Improving Action Chunking via Guided Test-Time Sampling [51.38330727868982]
  We show how action chunking impacts the divergence between a learner and a demonstrator.<n>We propose Bidirectional Decoding (BID), a test-time inference algorithm that bridges action chunking with closed-loop adaptation.<n>Our method boosts the performance of two state-of-the-art generative policies across seven simulation benchmarks and two real-world tasks.
  arXiv  Detail & Related papers  (2024-08-30T15:39:34Z)
• CUER: Corrected Uniform Experience Replay for Off-Policy Continuous Deep Reinforcement Learning Algorithms [5.331052581441265]
  We develop a novel algorithm, Corrected Uniform Experience (CUER), which samples the stored experience while considering the fairness among all other experiences. CUER provides promising improvements for off-policy continuous control algorithms in terms of sample efficiency, final performance, and stability of the policy during the training.
  arXiv  Detail & Related papers  (2024-06-13T12:03:40Z)
• Beyond Expected Return: Accounting for Policy Reproducibility when Evaluating Reinforcement Learning Algorithms [9.649114720478872]
  Many applications in Reinforcement Learning (RL) have noise ority present in the environment. These uncertainties lead the exact same policy to perform differently, from one roll-out to another. Common evaluation procedures in RL summarise the consequent return distributions using solely the expected return, which does not account for the spread of the distribution. Our work defines this spread as the policy: the ability of a policy to obtain similar performance when rolled out many times, a crucial property in some real-world applications.
  arXiv  Detail & Related papers  (2023-12-12T11:22:31Z)
• Dichotomy of Control: Separating What You Can Control from What You Cannot [129.62135987416164]
  We propose a future-conditioned supervised learning framework that separates mechanisms within a policy's control (actions) from those beyond a policy's control (environmentity) We show that DoC yields policies that are consistent with their conditioning inputs, ensuring that conditioning a learned policy on a desired high-return future outcome will correctly induce high-return behavior.
  arXiv  Detail & Related papers  (2022-10-24T17:49:56Z)
• Offline Reinforcement Learning with Implicit Q-Learning [85.62618088890787]
  Current offline reinforcement learning methods need to query the value of unseen actions during training to improve the policy. We propose an offline RL method that never needs to evaluate actions outside of the dataset. This method enables the learned policy to improve substantially over the best behavior in the data through generalization.
  arXiv  Detail & Related papers  (2021-10-12T17:05:05Z)
• Accelerated Policy Evaluation: Learning Adversarial Environments with Adaptive Importance Sampling [19.81658135871748]
  A biased or inaccurate policy evaluation in a safety-critical system could potentially cause unexpected catastrophic failures. We propose the Accelerated Policy Evaluation (APE) method, which simultaneously uncovers rare events and estimates the rare event probability. APE is scalable to large discrete or continuous spaces by incorporating function approximators.
  arXiv  Detail & Related papers  (2021-06-19T20:03:26Z)
• Provably Good Batch Reinforcement Learning Without Great Exploration [51.51462608429621]
  Batch reinforcement learning (RL) is important to apply RL algorithms to many high stakes tasks. Recent algorithms have shown promise but can still be overly optimistic in their expected outcomes. We show that a small modification to Bellman optimality and evaluation back-up to take a more conservative update can have much stronger guarantees.
  arXiv  Detail & Related papers  (2020-07-16T09:25:54Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.