机器人 / 具身智能 - arXivDaily 专题

2601.13247 2026-01-21 cs.CL cs.AI cs.CV cs.LG cs.MM 82%

Aligning Agentic World Models via Knowledgeable Experience Learning

通过知识性经验学习对齐代理世界模型

Baochang Ren, Yunzhi Yao, Rui Sun, Shuofei Qiao, Ningyu Zhang, Huajun Chen

机构 * Zhejiang University（浙江大学）； University of California, Los Angeles（加州大学洛杉矶分校）

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

AI总结 WorldMind通过知识性经验学习对齐代理世界模型，解决物理模拟中的幻觉问题，实现跨环境的高效迁移。

Comments Ongoing work

2511.00423 2026-01-16 cs.LG cs.AI cs.RO 82%

Bootstrap Off-policy with World Model

通过世界模型进行Bootstrap离策略学习

Guojian Zhan, Likun Wang, Xiangteng Zhang, Jiaxin Gao, Masayoshi Tomizuka, Shengbo Eben Li

机构 * College of AI & School of Vehicle and Mobility, Tsinghua University（人工智能学院与车辆与移动系统学院，清华大学）； Berkeley AI Research (BAIR), UC Berkeley（伯克利人工智能研究（BAIR），加州大学伯克利分校）

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.AI、cs.LG

AI总结 BOOM通过整合规划与离策略学习，利用世界模型和无参数动作分布提升策略性能，在高维控制任务中实现训练稳定性和最终性能的突破。

Comments NeurIPS 2025

2601.06604 2026-01-13 cs.AI cs.LG cs.RO 82%

Object-Centric World Models Meet Monte Carlo Tree Search

以对象为中心的世界模型与蒙特卡洛树搜索

Rodion Vakhitov, Leonid Ugadiarov, Aleksandr Panov

机构 * MIPT Moscow（莫斯科米斯托夫物理技术学院）； AIRI（人工智能研究所）； FRC CSC RAS（俄罗斯科学院计算机科学与应用数学研究所）

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.AI、cs.LG

AI总结本文提出ObjectZero算法，通过对象中心表示和图神经网络，结合蒙特卡洛树搜索实现更有效的环境建模与强化学习。

2505.03176 2026-01-12 cs.CV cs.AI cs.LG 82%

seq-JEPA: Autoregressive Predictive Learning of Invariant-Equivariant World Models

seq-JEPA: 自回归预测学习不变-等变世界模型

Hafez Ghaemi, Eilif Muller, Shahab Bakhtiari

机构 * Université de Montréal（蒙特利尔大学）； Mila - Quebec AI Institute（魁北克人工智能研究所）； CHU Sainte-Justine（圣约斯特医院）

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

AI总结 seq-JEPA通过引入架构归纳偏置，解决SSL中不变性与等变性任务间的性能权衡，同时学习两种独立表示以提升下游任务表现。

2601.00755 2026-01-05 eess.SY cs.SY 82%

A formal theory on problem space as a semantic world model in systems engineering

系统工程中作为语义世界模型的问题空间形式理论

Mayuranath SureshKumar, Hanumanthrao Kannan

专题命中具身推理：world model(title,abstract)；navigation(abstract)

AI总结本文提出了一种形式理论，将问题空间作为语义世界模型，以严格定义问题领域与解决方案之间的区分，并提供情境依赖的交互可追溯性和无歧义的边界语义。

Comments Submitted to Wiley Systems Engineering Journal

2210.13461 2025-12-30 cs.LG cs.AI cs.CV cs.NE q-bio.NC 82%

Active Predictive Coding: A Unified Neural Framework for Learning Hierarchical World Models for Perception and Planning

主动预测编码：一种统一的神经框架，用于学习感知与规划的分层世界模型

Rajesh P. N. Rao, Dimitrios C. Gklezakos, Vishwas Sathish

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

AI总结本文提出主动预测编码框架，通过分层世界模型解决人工智能中感知与规划的两大核心问题。

Comments 15 pages, 10 figures, 2 supplementary figures

2512.11225 2025-12-15 cs.CV cs.AI cs.LG 82%

VFMF: World Modeling by Forecasting Vision Foundation Model Features

VFMF：通过预测视觉基础模型特征进行世界建模

Gabrijel Boduljak, Yushi Lan, Christian Rupprecht, Andrea Vedaldi

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

AI总结本文提出了一种基于VFM特征的生成预测器，通过自回归流匹配在潜在空间中实现更准确的预测，提升世界建模的效率和效果。

2507.13162 2025-12-12 cs.CV cs.AI cs.LG 82%

Orbis: Overcoming Challenges of Long-Horizon Prediction in Driving World Models

Orbis：克服驾驶世界模型中长周期预测的挑战

Arian Mousakhan, Sudhanshu Mittal, Silvio Galesso, Karim Farid, Thomas Brox

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

AI总结 Orbis通过简单设计实现自动驾驶世界模型在长周期预测和复杂场景中的高性能，采用连续自回归模型优于离散token模型。

Comments Project page: https://lmb-freiburg.github.io/orbis.github.io/

2511.19584 2025-12-03 cs.LG cs.CV cs.RO 82%

Learning Massively Multitask World Models for Continuous Control

学习大规模多任务世界模型用于连续控制

Nicklas Hansen, Hao Su, Xiaolong Wang

机构 * University of California San Diego（加州大学圣地亚哥分校）

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.CV、cs.LG

AI总结 Newt通过大规模预训练和在线交互，实现智能体在数百个任务上的多任务学习，提升连续控制的效率和适应性。

Comments Webpage: https://www.nicklashansen.com/NewtWM

2510.19654 2025-11-26 cs.CV cs.AI cs.CL cs.RO 82%

From Forecasting to Planning: Policy World Model for Collaborative State-Action Prediction

从预测到规划：用于协作状态-动作预测的策略世界模型

Zhida Zhao, Talas Fu, Yifan Wang, Lijun Wang, Huchuan Lu

机构 * Dalian University of Technology（大连理工大学）

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.AI、cs.CV

AI总结本文提出策略世界模型（PWM），通过整合世界建模与轨迹规划，并利用无动作未来状态预测提升规划性能，实现更可靠的自主驾驶决策。

Comments Accepted by NuerIPS 2025 (Poster)

2511.09057 2025-11-18 cs.CV cs.AI cs.CL cs.LG 82%

PAN: A World Model for General, Interactable, and Long-Horizon World Simulation

PAN Team, Jiannan Xiang, Yi Gu, Zihan Liu, Zeyu Feng, Qiyue Gao, Yiyan Hu, Benhao Huang, Guangyi Liu, Yichi Yang, Kun Zhou, Davit Abrahamyan, Arif Ahmad, Ganesh Bannur, Junrong Chen, Kimi Chen, Mingkai Deng, Ruobing Han, Xinqi Huang, Haoqiang Kang, Zheqi Liu, Enze Ma, Hector Ren, Yashowardhan Shinde, Rohan Shingre, Ramsundar Tanikella, Kaiming Tao, Dequan Yang, Xinle Yu, Cong Zeng, Binglin Zhou, Zhengzhong Liu, Zhiting Hu, Eric P. Xing

机构 * Mohamed bin Zayed University of Artificial Intelligence（穆罕默德·本·扎耶德人工智能大学）

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

2506.01622 2025-10-21 cs.AI cs.LG cs.RO stat.ML 82%

General agents contain world models

Jonathan Richens, David Abel, Alexis Bellot, Tom Everitt

机构 * Google DeepMind（谷歌DeepMind）

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.AI、cs.LG

Comments Accepted ICML 2025. Typos corrected

2510.14977 2025-10-17 cs.CV cs.AI cs.LG 82%

Terra: Explorable Native 3D World Model with Point Latents

Yuanhui Huang, Weiliang Chen, Wenzhao Zheng, Xin Tao, Pengfei Wan, Jie Zhou, Jiwen Lu

机构 * Tsinghua University（清华大学）； Kuaishou Technology（快手科技）

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

Comments Project Page: https://huang-yh.github.io/terra/

2510.03727 2025-10-07 cs.AI cs.CL cs.CV cs.LG 82%

Bridging the Gap Between Multimodal Foundation Models and World Models

Xuehai He

机构 * Computer Science and Engineering University of California, Santa Cruz（计算机科学与工程大学加州大学圣克ruz分校）

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

Comments PhD thesis

2510.00855 2025-10-02 cs.CV cs.AI cs.CL cs.LG 82%

Can World Models Benefit VLMs for World Dynamics?

Kevin Zhang, Kuangzhi Ge, Xiaowei Chi, Renrui Zhang, Shaojun Shi, Zhen Dong, Sirui Han, Shanghang Zhang

机构 * Peking University（北京大学）； Hong Kong University of Science and Technology（香港科学与技术大学）； Chinese University of Hong Kong（香港中文大学）； University of California, Santa Barbara（加州大学圣芭芭拉分校）

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

Comments Project page: https://dyva-worldlm.github.io

2509.09737 2025-09-15 cs.CV cs.AI cs.LG 82%

World Modeling with Probabilistic Structure Integration

Klemen Kotar, Wanhee Lee, Rahul Venkatesh, Honglin Chen, Daniel Bear, Jared Watrous, Simon Kim, Khai Loong Aw, Lilian Naing Chen, Stefan Stojanov, Kevin Feigelis, Imran Thobani, Alex Durango, Khaled Jedoui, Atlas Kazemian, Dan Yamins

机构 * Stanford NeuroAI Lab（斯坦福神经人工智能实验室）

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

2507.13871 2025-07-21 cs.RO cs.CV cs.LG cs.SY eess.SY 82%

Safety Certification in the Latent space using Control Barrier Functions and World Models

Mehul Anand, Shishir Kolathaya

机构 * Indian Institute of Technology, Roorkee（印度理工学院，罗尔基）

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.CV、cs.LG

Comments 6 pages, 6 figures. arXiv admin note: text overlap with arXiv:2409.12616

2507.12496 2025-07-18 cs.RO cs.AI cs.LG 82%

FOUNDER: Grounding Foundation Models in World Models for Open-Ended Embodied Decision Making

Yucen Wang, Rui Yu, Shenghua Wan, Le Gan, De-Chuan Zhan

机构 * School of Artificial Intelligence, Nanjing University, China（人工智能学院，南京大学）； National Key Laboratory for Novel Software Technology, Nanjing University, China（新型软件技术国家重点实验室，南京大学）

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.AI、cs.LG

Comments Accepted by Forty-Second International Conference on Machine Learning (ICML 2025)

2507.05198 2025-07-08 cs.RO cs.AI cs.CV 82%

EmbodieDreamer: Advancing Real2Sim2Real Transfer for Policy Training via Embodied World Modeling

Boyuan Wang, Xinpan Meng, Xiaofeng Wang, Zheng Zhu, Angen Ye, Yang Wang, Zhiqin Yang, Chaojun Ni, Guan Huang, Xingang Wang

机构 * GigaAI ； Institute of Automation, Chinese Academy of Sciences（自动化研究所）； Peking University（北京大学）

专题命中具身推理：world model(title)；embodied AI(abstract)；分类 cs.RO、cs.AI、cs.CV

Comments Project Page: https://embodiedreamer.github.io/

2506.01600 2025-06-03 cs.RO cs.AI cs.CV 82%

WoMAP: World Models For Embodied Open-Vocabulary Object Localization

Tenny Yin, Zhiting Mei, Tao Sun, Lihan Zha, Emily Zhou, Jeremy Bao, Miyu Yamane, Ola Shorinwa, Anirudha Majumdar

机构 * Princeton University（普林斯顿大学）； McGill University（麦吉尔大学）

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.AI、cs.CV

2503.04416 2025-05-27 cs.LG cs.AI cs.CV 82%

Learning Transformer-based World Models with Contrastive Predictive Coding

Maxime Burchi, Radu Timofte

机构 * Computer Vision Lab, CAIDAS & IFI, University of Würzburg（计算机视觉实验室、CAIDAS与IFI、乌尔姆大学）

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

2505.15589 2025-05-22 cs.LG cs.AI cs.RO cs.SY eess.SY 82%

World Models as Reference Trajectories for Rapid Motor Adaptation

Carlos Stein Brito, Daniel McNamee

机构 * NightCity Labs（NightCity实验室）； Champalimaud Centre for the Unknown（Champalimaud未知中心）

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.AI、cs.LG

2405.18418 2025-05-16 cs.LG cs.CV cs.RO 82%

Hierarchical World Models as Visual Whole-Body Humanoid Controllers

Nicklas Hansen, Jyothir S, Vlad Sobal, Yann LeCun, Xiaolong Wang, Hao Su

机构 * UC San Diego（UC圣地亚哥大学）； New York University（纽约大学）； Meta AI

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.CV、cs.LG

Comments Code and videos at https://nicklashansen.com/rlpuppeteer

2410.02253 2025-04-22 cs.AI cs.LG cs.RO 82%

From Imitation to Exploration: End-to-end Autonomous Driving based on World Model

Yueyuan Li, Mingyang Jiang, Songan Zhang, Wei Yuan, Chunxiang Wang, Ming Yang

机构 * Department of Automation, Shanghai Jiao Tong University（自动化系，上海交通大学）； Key Laboratory of System Control and Information Processing, Ministry of Education of China（系统控制与信息处理重点实验室，中华人民共和国教育部）； Global Institute of Future Technology, Shanghai Jiao Tong University（未来技术全球研究院，上海交通大学）

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.AI、cs.LG

Comments 12 pages, 4 figures, 3 tables; T-ITS under review

2501.14174 2025-04-11 cs.CV cs.AI cs.LG 82%

Dreamweaver: Learning Compositional World Models from Pixels

Junyeob Baek, Yi-Fu Wu, Gautam Singh, Sungjin Ahn

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

2504.02252 2025-04-04 cs.LG cs.AI cs.RO 82%

Adapting World Models with Latent-State Dynamics Residuals

JB Lanier, Kyungmin Kim, Armin Karamzade, Yifei Liu, Ankita Sinha, Kat He, Davide Corsi, Roy Fox

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.AI、cs.LG

Comments 15 pages, 11 figures. Project website at https://redraw.jblanier.net/

2503.20523 2025-03-27 cs.CV cs.AI cs.RO 82%

GAIA-2: A Controllable Multi-View Generative World Model for Autonomous Driving

Lloyd Russell, Anthony Hu, Lorenzo Bertoni, George Fedoseev, Jamie Shotton, Elahe Arani, Gianluca Corrado

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.AI、cs.CV

Comments Technical Report

2503.15168 2025-03-20 cs.AI cs.CV cs.ET cs.LG 82%

World Models in Artificial Intelligence: Sensing, Learning, and Reasoning Like a Child

Javier Del Ser, Jesus L. Lobo, Heimo Müller, Andreas Holzinger

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

Comments 11 pages, 1 figure

2503.08122 2025-03-12 cs.LG cs.AI cs.CV 82%

Toward Stable World Models: Measuring and Addressing World Instability in Generative Environments

Soonwoo Kwon, Jin-Young Kim, Hyojun Go, Kyungjune Baek

专题命中具身推理：world model(title,abstract)；分类 cs.AI、cs.CV、cs.LG

Comments Preprint

2501.16733 2025-01-29 cs.RO cs.CV cs.LG 82%

Dream to Drive with Predictive Individual World Model

Yinfeng Gao, Qichao Zhang, Da-wei Ding, Dongbin Zhao

专题命中具身推理：world model(title,abstract)；分类 cs.RO、cs.CV、cs.LG

Comments Codes: https://github.com/gaoyinfeng/PIWM

1. 具身推理 3110 篇

Aligning Agentic World Models via Knowledgeable Experience Learning

Bootstrap Off-policy with World Model

Object-Centric World Models Meet Monte Carlo Tree Search

seq-JEPA: Autoregressive Predictive Learning of Invariant-Equivariant World Models

A formal theory on problem space as a semantic world model in systems engineering

Active Predictive Coding: A Unified Neural Framework for Learning Hierarchical World Models for Perception and Planning

VFMF: World Modeling by Forecasting Vision Foundation Model Features

Orbis: Overcoming Challenges of Long-Horizon Prediction in Driving World Models

Learning Massively Multitask World Models for Continuous Control

From Forecasting to Planning: Policy World Model for Collaborative State-Action Prediction

PAN: A World Model for General, Interactable, and Long-Horizon World Simulation

General agents contain world models

Terra: Explorable Native 3D World Model with Point Latents

Bridging the Gap Between Multimodal Foundation Models and World Models

Can World Models Benefit VLMs for World Dynamics?

World Modeling with Probabilistic Structure Integration

Safety Certification in the Latent space using Control Barrier Functions and World Models

FOUNDER: Grounding Foundation Models in World Models for Open-Ended Embodied Decision Making

EmbodieDreamer: Advancing Real2Sim2Real Transfer for Policy Training via Embodied World Modeling

WoMAP: World Models For Embodied Open-Vocabulary Object Localization

Learning Transformer-based World Models with Contrastive Predictive Coding

World Models as Reference Trajectories for Rapid Motor Adaptation

Hierarchical World Models as Visual Whole-Body Humanoid Controllers

From Imitation to Exploration: End-to-end Autonomous Driving based on World Model

Dreamweaver: Learning Compositional World Models from Pixels

Adapting World Models with Latent-State Dynamics Residuals

GAIA-2: A Controllable Multi-View Generative World Model for Autonomous Driving

World Models in Artificial Intelligence: Sensing, Learning, and Reasoning Like a Child

Toward Stable World Models: Measuring and Addressing World Instability in Generative Environments

Dream to Drive with Predictive Individual World Model