ELASTIC: Efficiently Learning to Adaptively Scale Test-Time Compute for Generative Control Policies

• URL: http://arxiv.org/abs/2606.31132v1
• Date: Tue, 30 Jun 2026 05:00:32 GMT
• Title: ELASTIC: Efficiently Learning to Adaptively Scale Test-Time Compute for Generative Control Policies
• Abstract summary: Generative control policies (GCPs) enable test-time scaling in robot control.<n>Test-time compute can be allocated along two axes: sequential scaling and parallel scaling.<n>We present ELASTIC, an algorithm that learns state-dependent test-time compute schedules for GCPs.
• Score: 36.28264388380867
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Generative control policies (GCPs), such as diffusion policies and flow-based vision-language-action models, enable test-time scaling in robot control. Test-time compute can be allocated along two axes: sequential scaling, which increases denoising steps to refine actions, and parallel scaling, which samples multiple candidate actions to search across modes of the policy distribution. However, the optimal allocation of sequential and parallel compute is hard to know a priori as it is state-, task-, and policy-dependent. For example, early stages of a grasp may benefit from broader parallel exploration, while near-contact phases may require more sequential refinement for precision. We present ELASTIC, an algorithm that learns state-dependent test-time compute schedules for GCPs. We formulate compute allocation as a meta-Markov Decision Process in which a meta-policy interacts with a frozen pretrained robot policy and selects sequential steps and parallel samples at each denoising iteration to maximize task success while minimizing compute. Using reinforcement learning, this meta-policy also learns adaptive compute schedules without access to the GCP's training data. Across simulated manipulation benchmarks with diffusion policies, ELASTIC Pareto-dominates fixed and single-axis scaling baselines at matched compute budgets. On real-world robot manipulation with the $π_{0.5}$ vision-language-action model, ELASTIC matches best-of-$10$ success while reducing wall-clock latency by 34%.

Related papers

• $τ_0$-VLA: a Hierarchical Robot Foundation Model with World-Model-Guided Test-Time Computation [51.05516457377339]
  Long-horizon robot manipulation requires a robot to both execute individual skills reliably and sequence them coherently over extended tasks.<n>We introduce $_0$-VLA, a hierarchical robot foundation model that formulates high-level subtask generation as a compute-scalable inference problem.<n>A low-level policy then executes the generated subtask across multiple robot embodiments.
  arXiv  Detail & Related papers  (2026-08-17T17:59:11Z)
• Test-Time Gradient Guidance of Flow Policies in Reinforcement Learning [50.738952715864116]
  Expressive continuous control policies form the backbone of advances in scaling imitation learning for simulated and real robot control.<n>We propose QGF (Q-Guided Flow), an RL algorithm that performs policy optimization entirely at test time.<n> Empirically, QGF outperforms prior test-time RL methods on single-task and goal-conditioned offline RL benchmarks.
  arXiv  Detail & Related papers  (2026-06-09T16:45:57Z)
• Compute Where it Counts: Self Optimizing Language Models [10.058821474955177]
  We study dynamic budget allocation for autoregressive decoding.<n>We train the policy with group-relative policy optimization on teacher-forced episodes.<n>Our reward trades off language-model quality against soft penalties that encourage episode-average budget usage.
  arXiv  Detail & Related papers  (2026-05-11T17:27:15Z)
• Breaking the Computational Barrier: Provably Efficient Actor-Critic for Low-Rank MDPs [53.412166189410904]
  We use supervised learning as a computational proxy to establish a hierarchy of commonly adopted RL oracles under low-rank Markov Decision Processes (MDPs)<n>Motivated by this observation, we propose a novel optimistic actor-critic algorithm that relies solely on the policy evaluation oracle.<n>We prove that our algorithm outperforms the existing sample complexity guarantees for low-rank MDPs while avoiding computationally expensive planning or optimization oracles commonly assumed in prior works.
  arXiv  Detail & Related papers  (2026-05-02T04:46:54Z)
• FASTER: Value-Guided Sampling for Fast RL [103.55398181003262]
  FASTER is a method for getting the benefits of sampling-based test-time scaling of diffusion-based policies without the computational cost.<n>FASTER achieves the same performance while substantially reducing training and inference compute requirements.
  arXiv  Detail & Related papers  (2026-04-21T17:52:17Z)
• Dynamic Test-Time Compute Scaling in Control Policy: Difficulty-Aware Stochastic Interpolant Policy [40.173458986694584]
  We introduce Difficulty-Aware Interpolant Policy (DA-SIP), a framework that enables robotic controllers to adaptively adjust their integration horizon in real time based on task difficulty.<n>DA-SIP achieves 2.6-4.4x reduction in total time while maintaining task success rates comparable to fixed maximum-computation baselines.
  arXiv  Detail & Related papers  (2025-11-25T22:46:42Z)
• IMLE Policy: Fast and Sample Efficient Visuomotor Policy Learning via Implicit Maximum Likelihood Estimation [3.7584322469996896]
  IMLE Policy is a novel behaviour cloning approach based on Implicit Maximum Likelihood Estimation (IMLE)<n>It excels in low-data regimes, effectively learning from minimal demonstrations and requiring 38% less data on average to match the performance of baseline methods in learning complex multi-modal behaviours.<n>We validate our approach across diverse manipulation tasks in simulated and real-world environments, showcasing its ability to capture complex behaviours under data constraints.
  arXiv  Detail & Related papers  (2025-02-17T23:22:49Z)
• Automatically Planning Optimal Parallel Strategy for Large Language Models [9.804975588324035]
  We propose an automatic parallel algorithm that automatically plans the parallel strategy with maximum throughput.<n>By decoupling the training time into computation, communication, and overlap, we established a training duration simulation model.<n>The multi-node experiment results show that the algorithm can estimate the parallel training duration in real time with an average accuracy of 96%.
  arXiv  Detail & Related papers  (2024-12-31T03:51:14Z)
• Learning Optimal Deterministic Policies with Stochastic Policy Gradients [62.81324245896716]
  Policy gradient (PG) methods are successful approaches to deal with continuous reinforcement learning (RL) problems. In common practice, convergence (hyper)policies are learned only to deploy their deterministic version. We show how to tune the exploration level used for learning to optimize the trade-off between the sample complexity and the performance of the deployed deterministic policy.
  arXiv  Detail & Related papers  (2024-05-03T16:45:15Z)
• Better than the Best: Gradient-based Improper Reinforcement Learning for Network Scheduling [60.48359567964899]
  We consider the problem of scheduling in constrained queueing networks with a view to minimizing packet delay. We use a policy gradient based reinforcement learning algorithm that produces a scheduler that performs better than the available atomic policies.
  arXiv  Detail & Related papers  (2021-05-01T10:18:34Z)

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.