Vision-Language-Action Models

Also known as VLA

These models combine vision-language backbones with action generators that produce continuous motion or discrete action tokens, often in chunks and with proprioceptive input. They are commonly trained on robot demonstrations, with flow-matching and diffusion policies, reinforcement fine-tuning, and predictive world models used to improve control and planning. Research spans tabletop and bimanual manipulation, mobile and aerial navigation, autonomous driving, and humanoid control, evaluated on benchmarks such as LIBERO, CALVIN, RoboTwin, and driving simulators. A central challenge is making broad visual and language competence translate into reliable closed-loop behavior under new objects, viewpoints, embodiments, delays, and safety constraints; benchmark success can still conceal weak instruction grounding and sim-to-real failures.

Latest papers

1,154 papers249 papers in the last four weeks, up 144% on the four weeks before. 1.5% of all new papers.

Uncovering and Mitigating Positional Blind Spots in Vision-Language-Action Models
Aug 3, 2026Dongdong An, Pengjie Zhao, Yihao Huang +5Vision-Language-Action Models VLMs for Robotics
Shanghai Normal University Shanghai, China · East China Normal University Shanghai, China · Northwest A&F University Yangling, Shaanxi, China +1
Demystifying When and Why VLAs Fail in Contact-Rich Tasks and How to Fix Them
Aug 2, 2026Carlota Parés-Morlans, Nils Kuhn, Isabel Liu +2Robotic Control Contact-Rich Robotic Manipulation
Stanford University
DreamTrajectory: Trajectory-Guided Action Generation with World Model Alignment for Mobile Manipulation
Aug 2, 2026Zheng Yang, Wenjie Zhang, Xiangyu Chen +7Trajectory Generation World Models for Robotics
Hong Kong University of Science and Technology (Guangzhou) · Ola Dimensions
Hermite Curves as Trajectory Priors for Vision-Language-Action Models
Aug 2, 2026Qi Lv, Jianming Xing, Zhao Yang +5Robot Motion Generation Robotic Manipulation
School of Computer Science and Technology, Harbin Institute of Technology (Shenzhen), China, 518055 · Department of Electrical and Computer Engineering and Show Lab, National University of Singapore, Singapore, 117583 · Jiangsu Cytoderm Intelligent Technology Co., Ltd., China, 215334
WAM-Diff2: Hierarchical AR-to-Diffusion Distillation for Highly Efficient Autonomous Driving VLA
Aug 2, 2026Zhihao Zhu, Hanlin Shang, Mingwang Xu +6End-to-End Autonomous Driving Diffusion Model Distillation
Fudan University · Yinwang Inttelligent Technology Co.,Ltd
Latency-Tolerant Cloud-Edge Collaborative Vision-Language-Action Models via Emergent Representational Specialization
Aug 1, 2026Daojie Peng, Fulong Ma, Bingtao Wang +2Robotic Control Edge Computing
HKUST(GZ) · Shandong University · HKUST
The Gate, Not the Cache: Gate Provenance Bounds the Closed-Loop Reliability of Training-Free VLA Token Skipping
Aug 1, 2026Qi Luo, Shuaijun Liu, Hao Zhao +5Efficient VLA Model Inference Vision-Language-Action Models
The Hong Kong University of Science and Technology (Guangzhou) · Shenzhen University of Advanced Technology · Sangfor Technologies Inc. +1
WCM: A World Critic Model for Vision-Language-Action Reinforcement Learning
Jul 31, 2026Senyu Fei, Xiaopeng Yu, Siyin Wang +3World Model Learning Vision-Language-Action Models
1Tongji University · 3Fudan University · 2Shanghai Innovation Institute
FibVLA: An Efficient Temporal Vision-Language-Action Model with Fibonacci Sampling
Jul 31, 2026Li Lin, Wujun Xu, Weiwei Meng +3Efficient VLA Model Inference Efficient VLA Models
Southeast University, Nanjing, Jiangsu, China · Nanyang Technological University, Singapore.
Safe Vision Language Action Models via Barrier Enhanced Flow Matching
Jul 31, 2026Kasra Sinaei, Hung-Chieh Wu, Donald EbeigbeFlow Matching Control Barrier Functions
The authors are with the Department of Electrical Engineering, The Pennsylvania State University
ActFovea: Runtime Safeguarding for VLA Policies via Spatiotemporal Visual-Action Consistency
Jul 31, 2026Wenda Yu, Tianshi Wang, Fengling Li +3Robot Safety Robotic Manipulation
Tongji University · Mohamed bin Zayed University of Artificial Intelligence · Shanghai Artificial Intelligence Laboratory +1
RoboBRIDGE: A Modular Framework for Bridging Policies to Robust Real-World Robotic Agents
Jul 30, 2026Sihyung Yoon, Minjong Yoo, Sanghyun Ahn +2Long-Horizon Robotic Manipulation Robotics
Department of Computer Science and Engineering, Sungkyunkwan University
Cross-Embodiment Transfer via Behavior-Aligned Representations
Jul 30, 2026Ajay Sridhar, Jensen Gao, Jonathan Yang +3Representation Learning Vision-Language-Action Models
1Stanford University · 2Toyota Research Institute
TurboVLA: Real-Time Vision-Language-Action Model at 32 Hz on an RTX 4090 with <1 GB VRAM
Jul 29, 2026Hengyi Xie, Chenfei Yao, Xianjin Wu +4Efficient VLA Model Inference Efficient VLA Models
Huazhong University of Science and Technology
DLAM: Distributional Latent Actions with Temporal Constraints
Jul 29, 2026Zuojin Tang, Feifan Luo, Haoyun Liu +10Latent Action Learning Latent Dynamics Modeling
Zhejiang University · Nanjing University · Shanghai Jiaotong University +4
RL $^2$ -VLA: Adaptive RL Latent Compositional Steering with Test-Time Scaling for Vision-Language-Action Models
Jul 29, 2026Derek Ming Siang Tan, Shailesh Shailesh, Srikrishna Iyer +4Vision-Language-Action Models Test-Time Adaptation of VLA Models
1National University of Singapore · 3Singapore Technologies Engineering · University of Toronto
CheckVLA: Execution-Time Verification with Action-Conditioned World Model for Long-Horizon Mobile Manipulation
Jul 29, 2026Yushan Liu, Peibo Sun, Xintao Chao +8Long-Horizon Robotic Manipulation World Models for Robotics
Tsinghua University · Shanghai Jiao Tong University · Peking University +2
Enfold: Folding World Model Imagination into Predictive Representations for Ultra-Efficient Embodied Control
Jul 29, 2026Weili Zeng, Yitong Xing, Fulong Liu +10World Models Efficient VLA Model Inference
Shanghai Jiao Tong University · Qilu University of Technology (Shandong Academy of Sciences)
Explicit Kinematic Guidance from Analytic Concepts for Vision-Language-Action Models
Jul 29, 2026Mingyang Sun, Jiude Wei, Xiujian Liang +4Robotic Manipulation Vision-Language-Action Models
Zhejiang University, Hangzhou, China · Shanghai Innovation Institute, Shanghai, China · Westlake University, Hangzhou, China +2
SAM3D-Guided Object-Centric Representation Alignment for Vision-Language-Action Models
Jul 28, 2026Zonghe Liu, Shanyuan Jie, Xiaoquan Sun +4Long-Horizon Robotic Manipulation Object-Centric Representation Learning
University of Hong Kong · Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences · Huazhong University of Science and Technology +2
A Causality-aware Infer-diagnose-refine Framework for Test-time Modality Adaptation in VLA Models
Jul 28, 2026Haoyu Zhang, Yuwei Wu, Jin Chen +6Robotic Manipulation Vision-Language-Action Models
1Beijing Institute of Technology · 2AInnovation Co. Ltd. · 3Beijing Innovation Center of Humanoid Robotics
CoTinyVLA: Chain-of-Thought Distillation for a Sub-Billion-Parameter Vision-Language-Action Model
Jul 28, 2026Minhyeok Lee, Chiyoung Kim, Chanhoe Gu +5CoT Distillation Efficient VLA Model Inference
Chung-Ang University, Republic of Korea
τ: Learning Touch-Augmented Vision-Language-Action Models from Future Visual Supervision
Jul 27, 2026Ning Cheng, Jinan Xu, Wanlin Li +5Visuo-Tactile Representation Learning Vision-Language-Action Models
Beijing Jiaotong University · Beijing Institute for General Artificial Intelligence
DeVA: Decoupled Video-Action Model with physical guidance for robot policy learning
Jul 27, 2026Mengqi Zhang, Sahil Khose, Simar Kareer +3Visuomotor Policy Learning Vision-Language-Action Models
University of California, Irvine · Georgia Institute of Technology
A Motion-Aware Vector Quantization Framework with Centroid Reuse for Efficient VLA Inference
Jul 27, 2026Zhuoran Song, Haozhe Jiang, Chunyu Qi +4AI Accelerator Inference Efficient VLA Model Inference
School of Computer Science, Shanghai Jiao Tong University, 2 Institute of Automation, Chinese Academy of Sciences China
FutureRTC: Real-Time Robot Execution with Anticipatory-Conditioned Action Chunking
Jul 27, 2026Hai Jiang, Yixian Zou, Binbin Liang +3Vision-Language-Action Models Action Chunking
School of Aeronautics and Astronautics, Sichuan University · University of Electronic Science and Technology of China · University of Alberta
Real2Sim2Real for Vision-Language-Action Manipulation: An AMD ROCm-Based Pipeline
Jul 25, 2026Qing Yang, Xun Wang, Ziguan Wang +3Language-Conditioned Robot Manipulation Robotics Simulation
AMD AIG Team · Beijing Institute of Technology
HyWorldVLA: A Vision-Language-Action Model with Hybrid World Modeling for Autonomous Driving
Jul 23, 2026Quanfu Yu, Xian Wu, Hao Xu +1World Models Autonomous Driving
Automotive New Technology Research Institute, BYD Company Limited
Closing the Lab-to-Store Gap: A Data-Efficient Post-Training and Experience-Driven Learning VLA Framework for Retail Humanoids
Jul 22, 2026Roger Sala Sisó, Tiago Silvério, Jakob Sand +1Robotic Manipulation Vision-Language-Action Models
HIVE Robots, Denmark · Department of Engineering Technology, Technical University of Denmark, Denmark
FM-VLA: Force-based Memory for Vision-Language-Action Models in Contact-Rich Manipulation
Jul 20, 2026Ruicheng Li, Qixiu Li, Ruichun Ma +8Memory-Augmented VLMs Language-Conditioned Robot Manipulation
Tsinghua University · Microsoft Research · Fudan University +1

Aug 3, 2026J/K move · Enter open · S save

Uncovering and Mitigating Positional Blind Spots in Vision-Language-Action Models

Dongdong An, Pengjie Zhao, Yihao Huang, Wenbing Tang+4

Shanghai Normal University Shanghai, China · East China Normal University Shanghai, China · Northwest A&F University Yangling, Shaanxi, China+1

Vision-Language-Action Models VLMs for Robotics Robustness of VLA Models

Open paper AskPDF arXiv

Recent Vision-Language-Action (VLA) models achieve promising performance in robotic manipulation, typically measured by success rates aggregated over predefined object configurations, an evaluation that implicitly assumes spatially uniform competence across the workspace. However, this assumption does not hold: even with the instruction and every other scene factor held fixed, merely relocating a task-irrelevant distractor can sharply raise the failure probability within localized, spatially coherent regions, which we term Positional Blind Spots (PBS). In this paper, we propose a two-stage black-box framework to uncover and mitigate PBS. During the uncovering stage, we grid the workspace and apply a one-sided log-likelihood-ratio test to localize PBS cells with significantly elevated risk. During the mitigation stage, we fine-tune the policy via LoRA on demonstrations collected from these PBS regions, improving competence there while largely preserving performance across the rest of the workspace. We evaluate our framework on five state-of-the-art VLA policies across two benchmarks, and find that PBS are pervasive and spatially concentrated in all of them, with failure rates up to 0.58. Our search strategy achieves an average F1-score of 0.678, outperforming random search and adaptive sampling baselines by 0.268 and 0.178, respectively. Guided by the discovered regions, targeted fine-tuning reduces the overall failure rate by 40.00%--85.19%.