动作分块的视觉-语言-动作（VLA）模型用于预测机器人未来动作序列。在固定时域执行模式下，机器人每次执行预测分块中固定长度的前缀后才再次调用策略，剩余预测结果未被使用。对于给定的已执行动作数量，较短的执行时域需要更多次策略调用；延长执行时域可将策略推理成本分摊到更多动作上，但在不同场景中统一采用长时域会限制精确修正的机会。为解决该问题，本文提出PICES（Perturbation-Informed Chunk and Execution Horizon Selection，即扰动感知分块与执行时域选择），这是一种无需训练的方法，可联合选择动作分块及下一次策略调用前需执行的动作数量。在单次策略调用中，动作专家基于多个初始噪声样本及其局部扰动对应样本，利用所有样本相同的条件特征生成一批动作分块。PICES利用每个名义分块与其对应分块中对应动作的差异，确定候选特定执行时域并选择用于执行的名义分块。在仿真基准测试和真实世界操作任务上的实验表明，与固定时域基线相比，PICES在Jetson Orin和Thor平台上的策略推理挂钟时间实现了最高2.3倍的加速，同时提升了任务成功率。
英文摘要：
Action-chunking vision-language-action (VLA) models predict sequences of future robot actions. Under fixed-horizon execution, the robot executes a fixedlength prefix of each predicted chunk before invoking the policy again, leaving the remaining predictions unused. For a given number of executed actions, short execution horizons require more policy calls. Extending the execution horizon amortizes policy inference cost over more actions. However, uniformly applying long horizons across episodes can limit opportunities for precise corrections. To address this limitation, we propose PICES (Perturbation-Informed Chunk and Execution Horizon Selection), a training-free method that jointly selects an action chunk and the number of actions executed before the next policy call. Within a single policy call, the action expert generates a batch of action chunks from multiple initial noise samples and their locally perturbed counterparts, using the same conditioning features across all samples. PICES exploits differences between corresponding actions in each nominal chunk and its counterpart to determine candidate-specific execution horizons and select a nominal chunk for execution. Experiments on simulation benchmarks and real-world manipulation tasks demonstrate that PICES achieves speedup by up to 2.3x in wall-clock policy inference time on Jetson Orin and Thor relative to fixed horizon baselines while improving the success rate.
Never Too Late for Force: Accelerating VLA Post-Training with Reactive Force Injection
Shanghai Jiao Tong University（上海交通大学）；Shanghai Innovation Institute（上海创新研究院）；Southern University of Science and Technology（南方科技大学）；Noematrix Ltd.（诺玛矩阵有限公司）
Fast-dVLA: Accelerating Discrete Diffusion VLA to Real-Time Performance
The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））；ShanghaiTech University（上海科技大学）；Shanghai Institute of Technical Physics, CAS（中国科学院上海技术物理研究所）；AIR, Tsinghua University（清华大学智能产业研究院）；Westlake University（西湖大学）；Zhejiang University（浙江大学）
OrthoSkillVLA: Continual Skill Learning via Gradient-Informed Skill Subspace Adaptation
School of Computer Science and Engineering, Southeast University（东南大学计算机科学与工程学院）；School of Electronic Science and Engineering, Southeast University（东南大学电子科学与工程学院）
An Anatomy of Vision-Language-Action Models: From Modules to Milestones and Challenges
IROOTECH TECHNOLOGY（IROOTECH技术公司）；Wolf 1069 b Lab, Sany Group（Sany集团沃尔夫1069b实验室）；Department of Engineering, King’s College London（伦敦国王学院工程系）；Hong Kong Polytechnic University（香港理工大学）；Computer Science Department of the Technische Universität Darmstadt（德累斯顿技术大学计算机科学系）；Department of ICT and Center for AI Research, University of Agder (UiA)（阿格德大学信息与通信技术系及人工智能研究中心）；Department of Computing, Imperial College London（伦敦帝国理工学院计算系）
FORCE: Efficient VLA Reinforcement Fine-Tuning via Value-Calibrated Warm-up and Self-Distillation
Institute of Automation, Chinese Academy of Sciences, Beijing, China（中国科学院自动化研究所）；Beijing Academy of Artificial Intelligence, Beijing, China（北京人工智能研究院）；State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University, Beijing, China（北京大学计算机科学学院多媒体信息处理国家重点实验室）
ActDistill: General Action-Guided Self-Derived Distillation for Efficient Vision-Language-Action Models
School of Computer Science and Technology, Tongji University（同济大学计算机科学与技术学院）；University of Technology Sydney（悉尼科技大学）；Advanced Institute of Big Data（大数据先进研究院）
