VLA / 视觉-语言-动作模型

2602.17951 2026-02-23 cs.CV cs.AI 88%

ROCKET: Residual-Oriented Multi-Layer Alignment for Spatially-Aware Vision-Language-Action Models

ROCKET：基于残差的多层对齐用于空间感知的视觉-语言-动作模型

Guoheng Sun, Tingting Du, Kaixi Feng, Chenxiang Luo, Xingguo Ding, Zheyu Shen, Ziyao Wang, Yexiao He, Ang Li

机构 * University of Maryland, College Park ； University of Wisconsin, Madison ； City University of Hong Kong ； St.\ Paul's School

专题命中 VLA模型：vision-language-action(title,abstract)；action model(title)；VLA(abstract)；分类 cs.CV、cs.AI

AI总结 ROCKET通过共享投影器和稀疏激活方案实现多层对齐，提升3D空间理解能力，在LIBERO等任务中达到98.5%的高成功率。

2602.18224 2026-02-23 cs.RO cs.LG 84%

SimVLA: A Simple VLA Baseline for Robotic Manipulation

SimVLA：一种简单的VLA基线用于机器人操作

Yuankai Luo, Woping Chen, Tong Liang, Baiqiao Wang, Zhenguo Li

专题命中 VLA模型：VLA(title,abstract)；vision-language-action(abstract)；分类 cs.RO、cs.LG

AI总结 SimVLA通过简化设计实现了最先进的机器人操作性能，展示了最小参数模型在模拟和真实环境中的优越表现。

2602.18397 2026-02-23 cs.RO 83%

How Fast Can I Run My VLA? Demystifying VLA Inference Performance with VLA-Perf

我如何快速跑动我的VLA？通过VLA-Perf解密VLA推理性能

Wenqi Jiang, Jason Clemons, Karu Sankaralingam, Christos Kozyrakis

机构 * NVIDIA Research（英伟达研究）

专题命中 VLA模型：VLA(title,abstract)；vision-language-action(abstract)；分类 cs.RO

AI总结本文通过VLA-Perf分析VLA模型的推理性能，探讨了模型设计和部署对实时推理的影响，提出了15项关键结论以指导未来VLA系统设计。

2602.18118 2026-02-23 astro-ph.HE 78%

Updated Air-Shower $X_{\rm max}$ Moment Parametrizations for UHECR Composition with Latest Hadronic Interaction Models

更新的空气 shower X_max 矩参数化用于 UHECR 组成与最新核相互作用模型

Carmelo Evoli, Igor Vaiman, Sergio Petrera, Francesco Salamida

专题命中 VLA模型：action model(title,abstract)

AI总结本文基于最新核相互作用模型，更新了 UHECR 组成的 $X_{\rm max} $ 矩参数化，用于高精度的宇宙射线研究。

Comments 19 pages, 10 figures, 5 tables. All data products and fit coefficients used in this work are publicly available via Zenodo at https://doi.org/10.5281/zenodo.18710119

2602.18300 2026-02-23 quant-ph cond-mat.mes-hall 71%

Impossibility of Refrigeration and Engine Operation in Minimal Qubit Repeated-Interaction Models

在最小量子比特重复相互作用模型中制冷和引擎操作的不可能性

Gabrielle Barsky-Giles, Alessandro Prositto, Matthew Gerry, Dvira Segal

专题命中 VLA模型：action model(title)

AI总结研究揭示了在最小量子比特重复相互作用模型中，制冷和引擎操作的不可能性，并展示了纯热传导的非单调行为。

2602.17836 2026-02-23 nlin.CD 50%

Nonlinear dynamics of a vertical pendulum driven by magnetic field provided by two coils magnets: analytical, numerical and experimental studies

由两个线圈磁铁提供的磁场驱动的垂直摆的非线性动力学：分析、数值和实验研究

Bonaventure Nana, Krystian Polczyński, Paul Woafo, Jan Awrejcewicz

专题命中 VLA模型：action model(abstract)

AI总结研究通过磁铁与线圈相互作用分析垂直摆的非线性动力学，揭示双稳和三稳势能特性，通过理论与实验验证多稳态行为。

Comments 25 pages, 21 figures

Journal ref Nana, B., Polczyński, K., Woafo, P., Awrejcewicz, J. Nonlinear dynamics of a vertical pendulum driven by magnetic field provided by two coils magnets: analytical, numerical and experimental studies. Nonlinear Dyn 113, 22585-22602 (2025)

2506.10058 2026-02-23 astro-ph.GA astro-ph.CO 50%

The Cosmic Owl: Twin Active Collisional Ring Galaxies with Starburst Merging Front at $z=1.14$

宇宙猫：z=1.14处的双主动碰撞环星系与星暴合并前沿

Mingyu Li, Bjorn H. C. Emonts, Zheng Cai, Takumi S. Tanaka, Wilfried Mercier, Yunjing Wu, Fujiang Yu, Fengwu Sun, Fuyan Bian, Emanuele Daddi, Xiaohui Fan, Xiaojing Lin, Jianwei Lyu, Jeyhan S. Kartaltepe, Francesco Valentino

专题命中 VLA模型：VLA(abstract)

AI总结宇宙猫在z=1.14处发现，由双碰撞环星系和AGN射流引发的星暴揭示了碰撞和AGN驱动的恒星形成机制。

Comments Accepted by ApJ, 17 pages, 9 figures