Physical Reasoning in Video

Follow this topic Ask about this topic

Momentum

10 papers in the last four weeks, up 100% on the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 54

All topics

World Models' Last Exam in Physics
Oct 6, 2026Mingju Gao, Qingle Liu, Yuzhao Peng +8Physical Consistency in Video Generation World Model Evaluation
Navers Lab, Einsia.AI · Peking University · Tsinghua University
How Does Geometry Enter Generated Motion?
Oct 4, 2026Weihan Li, Junhao Wu, Yuhan Song +2Physical Consistency in Video Generation Image-to-Video Generation
The University of Tokyo · RWTH Aachen University · Harbin Institute of Technology, Shenzhen
PhysicsLENS: Diagnosing Physical Property Blindness in Video Generation Models
Oct 1, 2026Isaiah Milkey, Som Sagar, Aditya Taparia +3Physical Consistency in Video Generation Physical Reasoning in Video
School of Computing and Augmented Intelligence Arizona State University Tempe, AZ, USA
Physis-Lang: Self-Evolving Language as a Physical Representation for Video World Model
Sep 30, 2026Liming Lu, Xianzheng Ma, Wenkun He +14Physical Consistency in Video Generation Video Captioning
MIT · University of Oxford · NVIDIA
Asking the World: Generalist Physical Reasoning through Agentic World Modeling and Probing
Sep 30, 2026Shenxiang Zeng, Chen Yang, Peiyao Chen +3World Model Learning Physics-Based Simulation
Tsinghua University
On Parameters of Nonlinear Scalar Dynamics from Video: Invariants, Calibration, and Identifiability
Sep 30, 2026Wenjie Wang, Yuanyuan Wang, Zixiang Jiang +2Parameter Estimation Parameter Identifiability
University of Melbourne · Mohamed bin Zayed University of Artificial Intelligence · Carnegie Mellon University
Scaling Video Generation for Reasoning: At What Cost?
Sep 29, 2026Weihang Guo, Xiaoyu Wu, Yifei Wang +2Video Reasoning Autoregressive Video Generation
Rice University · Carnegie Mellon University
Foresight at the Event Boundary: Evaluating Physical Prediction in Video World Models
Sep 29, 2026Estela Monserrat Arriaga Santana, Julian Rosas Scull, Ehécatl Sacamch'en Núñez Rico +1Physical Consistency in Video Generation World Model Evaluation
National Autonomous University of Mexico · National Autonomous University of Mexico, Mexico · University of Texas at El Paso +1
VideoPhysEdit: Physical Counterfactual Video Editing via Rigid-Body Physical Scene Reconstruction
Sep 28, 2026Conghan Yue, Yuanjie Chen, Yue Han +4Physical Consistency in Video Generation Video Editing
Institute of Trustworthy Embodied AI, Fudan University
Training Object Permanence in World Models
Sep 23, 2026Haotian Zhang, Fengyuan Yu, Dezhi Luo +28Object Permanence Physical Reasoning in Video
University of Southern California · Carnegie Mellon University · University of Michigan +13
PIVOT: Physics-Grounded Verification for AI-Generated Audio-Video Detection
Sep 14, 2026Bo Zheng, Kangran Zhao, Xiaoyu Zhang +7Explainable Deepfake Detection Deepfake Detection
School of Artificial Intelligence, Chinese University of Hong Kong, Shenzhen, Shenzhen 518172, China · Nanjing University of Posts and Telecommunications, Nanjing 210023, China · University at Buffalo, State University of New York, New York 14260, USA
Principia: Relational Physics Tests for Video Models
Sep 3, 2026Varun Varma Thozhiyoor, Shivam Tripathi, Venkatesh Babu Radhakrishnan +1VLM Evaluation Physical Consistency in Video Generation
Indian Institute of Science · Johns Hopkins University
VeriPhy: Agentic Physical Reasoning for World Model Evaluation and Refinement
Sep 2, 2026Wenzhuo Xu, Yuchen Zhu, Chongjian Ge +8Physical Consistency in Video Generation LLM Agent Verification
1Carnegie Mellon University · 2Georgia Institute of Technology · 4Adobe Research +1
What, Where, and How: Probing Spatiotemporal Representations in Video Foundation Models
Sep 1, 2026Sharon S. Musa, Fereshteh Forghani, Harrish Thasarathan +3Vision Foundation Models Video Representation Learning
Lassonde School of Engineering, York University, Toronto, Canada · Vector Institute, Toronto, Canada · FAIR, Meta Superintelligence Labs +2
Physically Plausible Video Generation via Visual-Semantic Chain-of-Events Conditioning
Sep 1, 2026Zixuan Wang, Yixin Hu, Wen Li +4Physical Consistency in Video Generation Video Diffusion Models
College of Electronics and Information Engineering, Sichuan University, Chengdu, 610064, China · School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu, 611731 · School of Computer Science, University of Adelaide, 5005, Adelaide, Australia +2
LaGSplat: Inferring Physics-Governed Interactive Simulation from Monocular Video Using Latent Lagrangian Gaussian Splatting
Aug 17, 2026Louen Pottier3D Gaussian Splatting Latent Dynamics Modeling
Université Paris-Saclay, CEA, List, F-91120 Palaiseau, France
EffectLearner: World-Aware Object-Effect Reasoning for Real-World Video Object Removal
Aug 6, 2026Feier Wu, Wanke Xia, Xu He +8Video Diffusion Models Video Editing
Tsinghua University · China Agricultural University · Jiangnan University +1
HERA: Historical Evidence Routing Adapter for Physical Prediction in Latent World Models
Aug 6, 2026Yuanruyi, Yue Cao, Haojia Gao +7Video Prediction Latent World Models
Chongqing University · Everwise-Tech Co., Ltd. · Research Institute of Tsinghua University in Shenzhen +2
PhysMind: From Video to Executable Worlds for Training-Free Physical Reasoning
Aug 5, 2026Chen Yang, Shenxiang Zeng, Haoyang Zhao +6Counterfactual Reasoning Physical Reasoning in Video
1Tsinghua University · 2The University of Hong Kong
PhyCheck: Fine-Grained Evidence-Grounded Dataset for Physical Law Understanding in Video-LLMs
Aug 3, 2026Zhongjie Ba, Shengwang Xu, Peng Cheng +4Video-Language Model Evaluation Video QA
State Key Lab. of Blockchain and Data Security, Zhejiang University, Hangzhou, China · Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security Hangzhou, Zhejiang, China · Mohamed Bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates
PhiZero: A World Model Built Around Physical Language
Jul 30, 2026Shuyao Shang, Yuqi Wang, Ruopeng Gao +4Action-Conditioned Video Generation Physical Consistency in Video Generation
NLPR, Institute of Automation, Chinese Academy of Sciences (CASIA) · Independent Researcher
DeforM: Reasoning-Guided Physics-Aware Video Generation via Spatial-Temporal Masking
Jul 21, 2026Yunyi Li, Yu Qiao, Yaohui Wang +1Physics-Informed Generative Modeling Physical Consistency in Video Generation
Shanghai Artificial Intelligence Laboratory, Shanghai, China
Apple- $π$ : Benchmarking Thinking with Video Towards Law-Grounded Physical Intelligence
Jul 17, 2026Runmao Yao, Kairui Hu, Yukang Cao +11Physical Consistency in Video Generation Physical Reasoning in Video
1S-Lab, Nanyang Technological University · 2The Chinese University of Hong Kong
PhysPlan: Grounded Physical State Reasoning and Graph-Guided Optimization for Physically Plausible Video Generation
Jul 16, 2026Minh-Loi Nguyen, Xuan-Vu Le, Trung-Nghia Le +3Physical Consistency in Video Generation Image-to-Video Generation
University of Science, Ho Chi Minh City, Vietnam · Vietnam National University, Ho Chi Minh City, Vietnam · University of Dayton, Dayton, Ohio, US +1
SportD: How do VLMs physically strategize?
Jul 16, 2026Jasin Cekinmez, Addison J. Wu, Haotian Xia +6VLM Evaluation Vision-Language Models
Princeton University · Rice University · UC Irvine +2
The Seriality Gap in Video Diffusion Models
Jul 14, 2026Jorge Diaz Chao, Konpat Preechakul, Yuxi Liu +1Video Diffusion Models Physical Reasoning in Video
UC Berkeley
PhysMRV: Physical Memory Retrieval and Verification for Physics Plausibility Reasoning
Jul 11, 2026Wenyuan Wang, Lianyu Hu, Hao Wang +1Memory-Augmented VLMs Physical Reasoning
1Nanyang Technological University · 2Rutgers University · University of Illinois Urbana-Champaign
SafeGuard: A Multi-Agent Perception-Reasoning Framework for Social-Risk AI-Generated Video Detection
Jul 3, 2026Wenlin Wu, Sheng Zhou, Peipei Song +3Deepfake Detection Multi-Agent Reasoning
University of Science and Technology of China · King Abdullah University of Science and Technology · Vast Intelligence Lab
PhysRAG: Enhancing Physics-Awareness in Video Generation via Retrieval-Augmented Generation
Jun 25, 2026Kexu Cheng, Zicheng Liu, Mingju Gao +2Multimodal RAG Physical Consistency in Video Generation
School of Computer Science, Peking University, Beijing, China · Institute of AI for Industries, Chinese Academy of Sciences, Nanjing, China
Physics-IQ Verified
Jun 17, 2026Tim Rädsch, Yuki M Asano, Hilde Kuehne +4Physical Consistency in Video Generation Image-to-Video Generation
Anates Labs · Technical University of Munich · University of Technology Nuremberg +3

Oct 6, 2026, cs.CVJ/K move · Enter open · S save

World Models' Last Exam in Physics

Mingju Gao, Qingle Liu, Yuzhao Peng, Xinjie Lin+7

Navers Lab, Einsia.AI · Peking University · Tsinghua University

Physical Consistency in Video Generation World Model Evaluation Physical Reasoning in Video Video World Models

Open paper AskPDF arXiv

Video world models can produce visually convincing yet physically inconsistent sequences, raising concerns about their reliability for prediction and planning in embodied AI systems. Existing evaluations often rely on model-based judgments or reference videos, while direct physical tests largely focus on mechanics. We introduce World Models' Last Exam in Physics, a measurement-based benchmark for evaluating physical consistency in video world models. The benchmark comprises 40 controlled tasks spanning mechanics, optics, fluids, thermal and phase-change phenomena, electromagnetism, and surface tension. Each task pairs an initial image and a generation prompt with predefined physical criteria, enabling interpretable tests of observable physical relationships without requiring reference videos. Its evaluator combines task-observability screening with task-specific quantitative physical measurements. Experiments on eight video generation models across 1,280 videos reveal persistent physical inconsistencies and substantial variation across tasks, with the best model achieving an overall score of 57.76 out of 100. Evaluation on synthetic videos with known physical relationships provides evidence for the validity of the measurement module under controlled conditions. The evaluator also achieves higher agreement with human judgments than a direct vision-language model baseline in both within-task rankings and pairwise comparisons. By combining coverage across physical domains with scores grounded in measurable evidence and explicit measurement limitations, the benchmark provides an interpretable basis for diagnosing physical inconsistencies and tracking progress toward physically consistent video world models.

Momentum

Related topics

Latest papers 54