Vision-language-action (VLA) models integrate visual observations and language instructions to predict robot actions, demonstrating promising generalization in manipulation tasks. However, most existing approaches primarily rely on direct mappings from 2D visual inputs to action sequences, without explicitly modeling the underlying 3D spatial structure or temporal world dynamics. Such representations may limit spatial reasoning and long-horizon decision-making in dynamic environments. To address this limitation, we propose StemVLA, a novel framework that explicitly incorporates both future-oriented 3D spatial knowledge and historical 4D spatiotemporal representations into action prediction. First, instead of relying solely on observed images, StemVLA forecasts structured 3D future spatial-geometric world knowledge, enabling the model to anticipate upcoming scene geometry and object configurations. Second, to capture temporal consistency and motion dynamics, we feed historical image frames into a pretrained video-geometry transformer backbone to extract implicit 3D world representations, and further aggregate them across time using a temporal attention module, termed VideoFormer [20], forming a unified 4D historical spatiotemporal representation. By jointly modeling 2D observations, predicted 3D future structure, and aggregated 4D temporal dynamics, StemVLA enables more comprehensive world understanding for robot manipulation. Extensive experiments in simulation demonstrate that Stem-VLA achieves an average accuracy of 92.0% across the LIBERO subsets, and 86.0% on the long-horizon LIBERO-Long subset.
MapNav: A Novel Memory Representation via Annotated Semantic Maps for Vision-and-Language Navigation
The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州));Beijing Academy of Artificial Intelligence(北京人工智能研究院);Tsinghua University(清华大学);Beijing Innovation Center of Humanoid Robotics Co., Ltd.(北京人形机器人创新中心有限公司);School of Computer Science, Wuhan University(武汉大学计算机学院);State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University(北京大学多媒体信息处理国家重点实验室,计算机学院)
TemporalFlow-VLA: Learning Physically Grounded Execution History for Long-Horizon Robot Manipulation
The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州));Zhejiang University(浙江大学);Simon Fraser University(西蒙菲莎大学);AgiBot(云圣智能科技(上海)有限公司)
History-Conditioned Spatio-Temporal Visual Token Pruning for Efficient Vision-Language Navigation
Department of Computer and Information Sciences at the University of Delaware(德克萨斯大学达勒姆分校计算机与信息科学系);University of Maryland’s Department of Computer Science(马里兰大学计算机科学系);Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学)
Memorized action chunking with Transformers: Imitation learning for vision-based tissue surface scanning
Imperial College London(帝国理工学院);Hamlyn Centre for Robotic Surgery(哈姆林机器人外科中心);Institute of Global Health Innovation(全球健康创新研究所);Department of Surgery and Cancer(外科与癌症系)
LightEMMA: A Longitudinal Evaluation of Vision-Language Models for Autonomous Driving
Department of Civil and Environmental Engineering, University of Michigan(土木与环境工程系,密歇根大学);University of Michigan Transportation Research Institute(密歇根大学交通研究所)
E-TTS: A New Embodied Test-Time Scaling Framework for Robotic Manipulation
New Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所模式识别国家重点实验室);School of Artificial Intelligence, University of Chinese Academy of Sciences(中国科学院大学人工智能学院);Foundation Model Research Center, Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所基础模型研究中心);FiveAges(五时代)
CycleManip: Enabling Cyclic Task Manipulation via Effective Historical Perception and Understanding
School of Computer Science and Engineering, Sun Yat-sen University(中山大学计算机科学与工程学院);The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳));Shenzhen Loop Area Institute(深圳河套学院)