VLA [Berkeley|Abbeel]
mjlab: A Lightweight Framework for GPU-Accelerated Robot Learning
轻量级GPU加速机器人学习框架，本周可复用
cs.RO
閱讀原文
VLA 研究日報 Pulsar
LIVE
— AI 線今日無資料 —— AI 线今日无资料 —
VLA 線VLA 线 · cs.RO · cs.AI · cs.LG
- Provably Safe Generative Sampling with Constricting Barrier Functions 提出安全生成采样方法，但无明确VLA应用路径 CS.AI
- mjlab: A Lightweight Framework for GPU-Accelerated Robot Learning [Berkeley|Abbeel] 轻量级GPU加速机器人学习框架，本周可复用 CS.RO
- Learning Dexterous Manipulation Skills from Imperfect Simulations 从不完美仿真中学习灵巧操作技能，有实际贡献 CS.RO
- Active Tactile Exploration for Rigid Body Pose and Shape Estimation 主动触觉探索刚体姿态和形状估计，解决触觉VLA问题 CS.RO
- Normalizing Flows are Capable Models for Bi-manual Visuomotor Policy 双臂视觉运动策略的归一化流模型，但实验较弱 CS.RO
- Synthetic vs. Real Training Data for Visual Navigation 比较合成与真实数据在视觉导航中的性能，无具体VLA应用 CS.RO
- EO-1: An Open Unified Embodied Foundation Model for General Robot Control 提出统一具身基础模型，但缺乏具体实验细节 [💧灌水] CS.RO
- ArtVIP: Articulated Digital Assets of Visual Realism, Modular Interaction, and Physical Fidelity for Robot Learning 提出高保真度数字资产用于机器人学习，但无具体VLA应用 CS.RO
- Adaptive Diffusion Constrained Sampling for Bimanual Robot Manipulation 双臂机器人操作的自适应扩散约束采样，但实验较弱 CS.RO
- PD-VLA: Accelerating Vision-Language-Action Model Integrated with Action Chunking via Parallel Decoding 通过并行解码加速VLA模型，有实质贡献 CS.RO
- A Distributional Treatment of Real2Sim2Real for Object-Centric Agent Adaptation in Vision-Driven Deformable Linear Object Manipulation 提出Real2Sim2Real框架，但实验设计较弱 CS.RO
- World Guidance: World Modeling in Condition Space for Action Generation 世界建模以促进动作生成，但无具体实验细节 CS.RO
VLA 研究日報VLA 研究日报
共 25 篇
🔧 技術技术
VLA
Learning Dexterous Manipulation Skills from Imperfect Simulations
从不完美仿真中学习灵巧操作技能，有实际贡献
cs.RO
閱讀原文
VLA
Active Tactile Exploration for Rigid Body Pose and Shape Estimation
主动触觉探索刚体姿态和形状估计，解决触觉VLA问题
cs.RO
閱讀原文
VLA
PD-VLA: Accelerating Vision-Language-Action Model Integrated with Action Chunking via Parallel Decoding
通过并行解码加速VLA模型，有实质贡献
cs.RO
閱讀原文
📖 背景閱讀背景阅读
VLA
Provably Safe Generative Sampling with Constricting Barrier Functions
提出安全生成采样方法，但无明确VLA应用路径
cs.AI
閱讀原文
VLA
Normalizing Flows are Capable Models for Bi-manual Visuomotor Policy
双臂视觉运动策略的归一化流模型，但实验较弱
cs.RO
閱讀原文
VLA
Synthetic vs. Real Training Data for Visual Navigation
比较合成与真实数据在视觉导航中的性能，无具体VLA应用
cs.RO
閱讀原文
VLA
EO-1: An Open Unified Embodied Foundation Model for General Robot Control
提出统一具身基础模型，但缺乏具体实验细节 [💧灌水]
cs.RO
閱讀原文
VLA
ArtVIP: Articulated Digital Assets of Visual Realism, Modular Interaction, and Physical Fidelity for Robot Learning
提出高保真度数字资产用于机器人学习，但无具体VLA应用
cs.RO
閱讀原文
VLA
Adaptive Diffusion Constrained Sampling for Bimanual Robot Manipulation
双臂机器人操作的自适应扩散约束采样，但实验较弱
cs.RO
閱讀原文
VLA
A Distributional Treatment of Real2Sim2Real for Object-Centric Agent Adaptation in Vision-Driven Deformable Linear Object Manipulation
提出Real2Sim2Real框架，但实验设计较弱
cs.RO
閱讀原文
VLA
World Guidance: World Modeling in Condition Space for Action Generation
世界建模以促进动作生成，但无具体实验细节
cs.RO
閱讀原文
VLA [DeepMind|Song]
Are Foundation Models the Route to Full-Stack Transfer in Robotics?
综述论文，讨论基础模型在机器人中的应用
cs.RO
閱讀原文
VLA
DexRepNet++: Learning Dexterous Robotic Manipulation with Geometric and Spatial Hand-Object Representations
提出几何和空间手-物体表示学习灵巧操作，但实验较弱
cs.RO
閱讀原文
VLA
Joint-Aligned Latent Action: Towards Scalable VLA Pretraining in the Wild
提出关节对齐潜在动作，但实验设计较弱
cs.RO
閱讀原文
VLA
Primary-Fine Decoupling for Action Generation in Robotic Imitation
提出主细解耦动作生成，但实验较弱
cs.RO
閱讀原文
VLA
Biomechanical Comparisons Reveal Divergence of Human and Humanoid Gaits
生物力学比较揭示人类和类人机器人步态差异，但无具体VLA应用
cs.RO
閱讀原文
VLA
Self-Correcting VLA: Online Action Refinement via Sparse World Imagination
提出自纠正VLA模型，但实验设计较弱
cs.RO
閱讀原文
VLA
Tacmap: Bridging the Tactile Sim-to-Real Gap via Geometry-Consistent Penetration Depth Map
提出几何一致穿透深度图，但实验较弱
cs.RO
閱讀原文
VLA
ADM-DP: Adaptive Dynamic Modality Diffusion Policy through Vision-Tactile-Graph Fusion for Multi-Agent Manipulation
提出多模态融合策略，但实验设计较弱
cs.RO
閱讀原文
VLA
Learning Agile and Robust Omnidirectional Aerial Motion on Overactuated Tiltable-Quadrotors
学习敏捷和鲁棒的全向空中运动，但无具体VLA应用
cs.RO
閱讀原文
VLA
LiLo-VLA: Compositional Long-Horizon Manipulation via Linked Object-Centric Policies
提出长时域操作策略，但实验设计较弱
cs.RO
閱讀原文
VLA
VLA Knows Its Limits
讨论VLA模型的局限性，但无具体实验细节
cs.RO
閱讀原文
VLA
CableRobotGraphSim: A Graph Neural Network for Modeling Partially Observable Cable-Driven Robot Dynamics
提出图神经网络建模部分可观测电缆驱动机器人动力学，但无具体VLA应用
cs.RO
閱讀原文
VLA
FlowCorrect: Efficient Interactive Correction of Generative Flow Policies for Robotic Manipulation
提出高效交互式修正生成流策略，但实验设计较弱
hf-papers
閱讀原文