mlsanityData to 14 Oct 2026

Sekai: A Video Dataset towards World Exploration

arXiv 2506.15675 ↗First posted 18 Jun 2025cs.CV cs.AIcited by 67

Not on the map. Why?

Cited by 37 in World models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Video generation techniques have made remarkable progress, promising to be the foundation of interactive world exploration. However, existing video generation datasets are not well-suited for world exploration training as they suffer from some limitations: limited locations, short duration, static scenes, and a lack of annotations about exploration and the world. In this paper, we introduce Sekai (meaning "world" in Japanese), a high-quality first-person view worldwide video dataset with rich annotations for world exploration. It consists of over 5,000 hours of walking or drone view (FPV and UVA) videos from over 100 countries and regions across 750 cities. We develop an efficient and effective toolbox to collect, pre-process and annotate videos with location, scene, weather, crowd density, captions, and camera trajectories. Comprehensive analyses and experiments demonstrate the dataset's scale, diversity, annotation quality, and effectiveness for training video generation models. We believe Sekai will benefit the area of video generation and world exploration, and motivate valuable applications. The project page is https://lixsp11.github.io/sekai-project/.

Key facts. Cited by 67 papers in the map (data through 14 Oct 2026). First posted on arXiv 18 Jun 2025.

Builds on

0 of 0 references.

No references here yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 14; 2026: 53.

67 papers.

Sort:

By area:

  1. 2025Matrix-game 2.0: An open-source real-time and streaming interactive world modelcited by 186
  2. 2025WorldPlay: Towards Long-Term Geometric Consistency for Real-Time Interactive World Modelingcited by 170
  3. 2025Emu3.5: Native Multimodal Models are World Learnerscited by 132
  4. 2025RELIC: Interactive Video World Model with Long-Horizon Memorycited by 96
  5. 2025Yume: An Interactive World Generation Modelcited by 81
  6. 2025SpatialVID: A Large-Scale Video Dataset with Spatial Annotationscited by 77
  7. 2025OmniWorld: A Multi-Domain and Multi-Modal Dataset for 4D World Modelingcited by 62
  8. 2026Context Forcing: Consistent Autoregressive Video Generation with Long Contextcited by 59
  9. 2026SANA-WM: Efficient Minute-Scale World Modeling with Hybrid Linear Diffusion Transformercited by 55
  10. 2026Infinite-World: Scaling Interactive World Models to 1000-Frame Horizons via Pose-Free Hierarchical Memorycited by 40
  11. 2026DreamX-World 1.0: A General-Purpose Interactive World Modelcited by 40
  12. 2025Astra: General Interactive World Model with Autoregressive Denoisingcited by 33
38 more
  1. 2025Pack and Force Your Memory: Long-form and Consistent Video Generationcited by 29
  2. 2026VerseCrafter: Dynamic Realistic Video World Model with 4D Geometric Controlcited by 28
  3. 2026MosaicMem: Hybrid Spatial Memory for Controllable Video World Modelscited by 27
  4. 2026Out of Sight but Not Out of Mind: Hybrid Memory for Dynamic Video World Modelscited by 26
  5. 2026VGGT-Omegacited by 26
  6. 2026Omni-WorldBench: Towards a Comprehensive Interaction-Centric Evaluation for World Modelscited by 18
  7. 2026WildWorld: A Large-Scale Dataset for Dynamic World Modeling with Actions and Explicit State toward Generative ARPGcited by 13
  8. 2026Mode Seeking meets Mean Seeking for Fast Long Video Generationcited by 12
  9. 2026WorldCam: Interactive Autoregressive 3D Gaming Worlds with Camera Pose as a Unifying Geometric Representationcited by 12
  10. 2026SolarWM: Open Data and Scalable Training for Long-Horizon Video World Modelscited by 12
  11. 2026AlayaWorld: Interactive Long-Horizon World Modeling -- Full Technical Reportcited by 11
  12. 2026ReWorld: An Interactive World Model with Long-Horizon Memorycited by 10
  13. 2026MemLearner: Learning to Query Context memory for Video World Modelscited by 9
  14. 2025MVInverse: Feed-forward Multi-view Inverse Rendering in Secondscited by 6
  15. 2026BiWM: Advancing Open-Source Interactive Video World Models with Bidirectional Autoregressioncited by 6
  16. 2026OmniShotCut: Holistic Relational Shot Boundary Detection with Shot-Query Transformercited by 5
  17. 2026Worldscape-MoE: A Unified Mixture-of-Experts World Model for Scalable Heterogeneous Action Controlcited by 5
  18. 2026TrajVG: 3D Trajectory-Coupled Visual Geometry Learningcited by 4
  19. 2026From Pixels to States: Rethinking Interactive World Models as Game Enginescited by 4
  20. 2025UrbanVLA: A Vision-Language-Action Model for Urban Micromobilitycited by 3
  21. 2025MagicWorld: Towards Long-Horizon Stability for Interactive Video World Explorationcited by 3
  22. 2026SynthVerse: A Large-Scale Diverse Synthetic Dataset for Point Trackingcited by 3
  23. 2026Geo-Align: Video Generation Alignment via Metric Geometry Rewardcited by 3
  24. 2026Holo-World: Unified Camera, Object and Weather Control for Video World Modelcited by 3
  25. 2026HumanScale: Egocentric Human Video Can Outperform Real-Robot Data for Embodied Pretrainingcited by 3
  26. 2026Sekai2: From World Exploration to Interactive World Modelingcited by 3
  27. 2025Target-Bench: Can Video World Models Achieve Mapless Path Planning with Semantic Targets?cited by 2
  28. 2026DisCo: World Models with Discrete Camera Motion Controlcited by 2
  29. 2026PhysEditWorld: A Large-Scale Dataset Toward Physics-Editable World Modelscited by 2
  30. 2026PhiZero: A World Model Built Around Physical Languagecited by 2
  31. 2026Video-DeepResearch: Towards the Next-Generation Multimodal Deepresearch Agentcited by 2
  32. 2026WorldRover: A Scalable Synthetic Video Data Engine for World Exploration with Rich Annotationscited by 2
  33. 2026Mask Forcing: Improving Autoregressive Video Diffusion Distillation via Dual-Noise Masking Rolloutcited by 2
  34. 2025CineLOG: A Training Free Approach for Cinematic Long Video Generationcited by 1
  35. 2026SceneScribe-1M: A Large-Scale Video Dataset with Comprehensive Geometric and Semantic Annotationscited by 1
  36. 2026MotionScape: A Motion-Stratified UAV Video Benchmark for World Modeling and Future Video Generationcited by 1
  37. 2026HelloWorld: Enabling Socially Interactive Characters in Video World Modelscited by 1
  38. 2026Precise Editing and Flexible Referencing for Interactable Worldscited by 1

50 most cited of 67 All 67 (JSON)