mlsanityData to 7 Oct 2026

Epona: Autoregressive Diffusion World Model for Autonomous Driving

arXiv 2506.24113 ↗First posted 30 Jun 2025cs.CVcited by 115

On the map: World models → World models for autonomous driving

also in 2 more
Abstract (arXiv)

Diffusion models have demonstrated exceptional visual quality in video generation, making them promising for autonomous driving world modeling. However, existing video diffusion-based world models struggle with flexible-length, long-horizon predictions and integrating trajectory planning. This is because conventional video diffusion models rely on global joint distribution modeling of fixed-length frame sequences rather than sequentially constructing localized distributions at each timestep. In this work, we propose Epona, an autoregressive diffusion world model that enables localized spatiotemporal distribution modeling through two key innovations: 1) Decoupled spatiotemporal factorization that separates temporal dynamics modeling from fine-grained future world generation, and 2) Modular trajectory and video prediction that seamlessly integrate motion planning with visual modeling in an end-to-end framework. Our architecture enables high-resolution, long-duration generation while introducing a novel chain-of-forward training strategy to address error accumulation in autoregressive loops. Experimental results demonstrate state-of-the-art performance with 7.4% FVD improvement and minutes longer prediction duration compared to prior works. The learned world model further serves as a real-time motion planner, outperforming strong end-to-end planners on NAVSIM benchmarks. Code will be publicly available at hrefhttps://github.com/Kevin-thu/Epona/https://github.com/Kevin-thu/Epona/.

Key facts. Cited by 115 papers in the map (data through 7 Oct 2026). Placed in World models for autonomous driving, one of 11 directions in World models. Builds on 52 of its 68 references that are in the map. First posted on arXiv 30 Jun 2025.

Builds on

52 of 68 references.

Closest ancestors · 13

Same direction, newest first

  1. 2024DrivingWorld: Constructing World Model for Autonomous Driving via Video GPT27 Dec
  2. 2024DrivingGPT: Unifying Driving World Modeling and Planning with Multi-modal Autoregressive Transformers24 Dec
  3. 2024Doe-1: Closed-Loop Autonomous Driving with Large World Model12 Dec
  4. 2024DOME: Taming Diffusion Model into High-Fidelity Controllable Occupancy World Model14 Oct
  5. 2024Enhancing End-to-End Autonomous Driving with Latent World Model12 Jun
  6. 2024Vista: A Generalizable Driving World Model with High Fidelity and Versatile Controllability27 May
  7. 2024GenAD: Generalized Predictive Model for Autonomous Driving14 Mar
  8. 2024GenAD: Generative End-to-End Autonomous Driving18 Feb
  9. 2023Driving into the Future: Multiview Visual Forecasting and Planning with World Model for Autonomous Driving29 Nov
  10. 2023OccWorld: Learning a 3D Occupancy World Model for Autonomous Driving27 Nov
  11. 2023GAIA-1: A Generative World Model for Autonomous Driving29 Sep
  12. 2023DriveDreamer: Towards Real-world-driven World Models for Autonomous Driving18 Sep
1 more
  1. 2021NuPlan: A closed-loop ML-based planning benchmark for autonomous vehicles22 Jun

Other references · 39

  1. 2017Attention Is All You Needcited by 59,932
  2. 2020Language Models are Few-Shot Learnerscited by 32,199
  3. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  4. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  5. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  6. 2022Scalable Diffusion Models with Transformerscited by 5,931
  7. 2017Neural Discrete Representation Learningcited by 4,650
  8. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,485
  9. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,550
  10. 2020Taming Transformers for High-Resolution Image Synthesiscited by 3,229
  11. 2023Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasetscited by 2,526
  12. 2024HunyuanVideo: A Systematic Framework For Large Video Generative Modelscited by 1,597
27 more
  1. 2023Align your Latents: High-Resolution Video Synthesis with Latent Diffusion Modelscited by 1,465
  2. 2018World Modelscited by 1,444
  3. 2018Towards Accurate Generative Models of Video: A New Metric & Challengescited by 1,204
  4. 2022Building Normalizing Flows with Stochastic Interpolantscited by 964
  5. 2025Cosmos World Foundation Model Platform for Physical AIcited by 833
  6. 2024Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusioncited by 634
  7. 2023VAD: Vectorized Scene Representation for Efficient Autonomous Drivingcited by 613
  8. 2022Phenaki: Variable Length Video Generation From Open Domain Textual Descriptioncited by 475
  9. 2021Multi-Modal Fusion Transformer for End-to-End Autonomous Drivingcited by 372
  10. 2022ST-P3: End-to-end Vision-based Autonomous Driving via Spatial-Temporal Feature Learningcited by 365
  11. 2024NAVSIM: Data-Driven Non-Reactive Autonomous Vehicle Simulation and Benchmarkingcited by 343
  12. 2024VADv2: End-to-End Vectorized Autonomous Driving via Probabilistic Planningcited by 287
  13. 2023MagicDrive: Street View Generation with Diverse 3D Geometry Controlcited by 250
  14. 2024Deep Compression Autoencoder for Efficient High-Resolution Diffusion Modelscited by 209
  15. 2023Scene as Occupancycited by 208
  16. 2023FreeNoise: Tuning-Free Longer Video Diffusion via Noise Reschedulingcited by 173
  17. 2023NUWA-XL: Diffusion over Diffusion for eXtremely Long Video Generationcited by 170
  18. 2024JanusFlow: Harmonizing Autoregression and Rectified Flow for Unified Multimodal Understanding and Generationcited by 162
  19. 2024Understanding World or Predicting Future? A Comprehensive Survey of World Modelscited by 161
  20. 2023Gen-L-Video: Multi-Text to Long Video Generation via Temporal Co-Denoisingcited by 135
  21. 2024FIFO-Diffusion: Generating Infinite Videos from Text without Trainingcited by 134
  22. 2021DriveGAN: Towards a Controllable High-Quality Neural Simulationcited by 125
  23. 2025GoalFlow: Goal-Driven Flow Matching for Multimodal Trajectories Generation in End-to-End Autonomous Drivingcited by 117
  24. 2024Loong: Generating Minute-level Long Videos with Autoregressive Language Modelscited by 99
  25. 2023WoVoGen: World Volume-aware Diffusion for Controllable Multi-camera Driving Scene Generationcited by 93
  26. 2024DRAMA: An Efficient End-to-end Motion Planner for Autonomous Driving with Mambacited by 87
  27. 2025Generating Multimodal Driving Scenes via Next-Scene Predictioncited by 12

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 24; 2026: 91.

115 papers.

Sort:

By area:

  1. 2025A Comprehensive Survey on World Models for Embodied AIcited by 65
  2. 20253D and 4D World Modeling: A Surveycited by 64
  3. 2025A Survey: Learning Embodied Intelligence from Physical Simulators and World Modelscited by 50
  4. 2025DriveLaW:Unifying Planning and Video Generation in a Latent Driving Worldcited by 43
  5. 2026Infinite-World: Scaling Interactive World Models to 1000-Frame Horizons via Pose-Free Hierarchical Memorycited by 37
  6. 2026DSpark: Confidence-Scheduled Speculative Decoding with Semi-Autoregressive Generationcited by 36
  7. 2026SGDrive: Scene-to-Goal Hierarchical World Cognition for Autonomous Drivingcited by 34
  8. 2025OmniNWM: Omniscient Driving Navigation World Modelscited by 32
  9. 2025Vision-Language-Action Models for Autonomous Driving: Past, Present, and Futurecited by 32
  10. 2026DriveWorld-VLA: Unified Latent-Space World Modeling with Vision-Language-Action for Autonomous Drivingcited by 31
  11. 2026Latent-WAM: Latent World Action Modeling for End-to-End Autonomous Drivingcited by 27
  12. 2026Drive-JEPA: Video JEPA Meets Multimodal Trajectory Distillation for End-to-End Drivingcited by 26
38 more
  1. 2025ImagiDrive: A Unified Imagination-and-Planning Framework for Autonomous Drivingcited by 24
  2. 2026UniDrive-WM: Unified Understanding, Planning and Generation World Model for Autonomous Drivingcited by 20
  3. 2025WorldRFT: Latent World Model Planning with Reinforcement Fine-Tuning for Autonomous Drivingcited by 19
  4. 2026DriveVA: Video Action Models are Zero-Shot Driverscited by 19
  5. 2025DIVER: Reinforced Diffusion Breaks Imitation Bottlenecks in End-to-End Autonomous Drivingcited by 17
  6. 2026DriveDreamer-Policy: A Geometry-Grounded World-Action Model for Unified Generation and Planningcited by 17
  7. 2025Simulating the Visual World with Artificial Intelligence: A Roadmapcited by 15
  8. 2025WAM-Flow: Parallel Coarse-to-Fine Motion Planning via Discrete Flow Matching for Autonomous Drivingcited by 15
  9. 2026DriveFuture: Future-Aware Latent World Models for Autonomous Drivingcited by 15
  10. 2025LiDARCrafter: Dynamic 4D World Modeling from LiDAR Sequencescited by 14
  11. 2025AD-R1: Closed-Loop Reinforcement Learning for End-to-End Autonomous Driving with Impartial World Modelscited by 13
  12. 2026Bridging Scene Generation and Planning: Driving with World Model via Unifying Vision and Motion Representationcited by 13
  13. 2026EponaV2: Driving World Model with Comprehensive Future Reasoningcited by 12
  14. 2026DriveWAM: Video Generative Priors Enable Scalable World-Action Modeling for Autonomous Drivingcited by 12
  15. 2025Towards Unified World Models for Visual Navigation via Memory-Augmented Planning and Foresightcited by 11
  16. 2026DynVLA: Learning World Dynamics for Action Reasoning in Autonomous Drivingcited by 11
  17. 2025UniUGP: Unifying Understanding, Generation, and Planing For End-to-end Autonomous Drivingcited by 10
  18. 2025GenieDrive: Towards Physics-Aware Driving World Model with 4D Occupancy Guided Video Generationcited by 10
  19. 2026ExploreVLA: Dense World Modeling and Exploration for End-to-End Autonomous Drivingcited by 10
  20. 2026CoWorld-VLA: Thinking in a Multi-Expert World Model for Autonomous Drivingcited by 10
  21. 2026Metis: A Generalizable and Efficient World-Action Model for Autonomous Driving and Urban Navigationcited by 10
  22. 2025MGVQ: Could VQ-VAE Beat VAE? A Generalizable Tokenizer with Multi-group Quantizationcited by 9
  23. 2026FLARE: Learning Future-Aware Latent Representations from Vision-Language Models for Autonomous Drivingcited by 9
  24. 2026Uni-World VLA: Interleaved World Modeling and Planning for Autonomous Drivingcited by 9
  25. 2026Qwen-Drive-1.0: An Initial Step towards a Vision-Language Foundation Model for Autonomous Drivingcited by 8
  26. 2025FutureX: Enhance End-to-End Autonomous Driving via Latent Chain-of-Thought World Modelcited by 7
  27. 2026SafeDrive: Fine-Grained Safety Reasoning for End-to-End Driving in a Sparse Worldcited by 7
  28. 2026Beyond Imitation: Learning Safe End-to-End Autonomous Driving from Hard Negativescited by 6
  29. 2026Towards Interactive Video World Modeling: Frontiers, Challenges, Benchmarks, and Future Trendscited by 6
  30. 2026GraphWorld: Long-Horizon Planning with World Models for End-to-End Autonomous Drivingcited by 6
  31. 2026Video Generation Models as World Models: Efficient Paradigms, Architectures and Algorithmscited by 5
  32. 2026UNIVERSE: Unified Video Action Models for Autonomous Driving with Flexible Mask-Modulated Modality Generationcited by 5
  33. 2026SimWAM: A Simple World Action Model for End-to-End Autonomous Drivingcited by 5
  34. 2025WPT: World-to-Policy Transfer via Online World Model Distillationcited by 4
  35. 2026Reasoning models do not yet follow their reasoning in autonomous driving: The KITScenes LongTail Datasetcited by 4
  36. 2026UniDWM: Towards a Unified Driving World Model via Multifaceted Representation Learningcited by 3
  37. 2026Latent World Models for Automated Driving: A Unified Taxonomy, Evaluation Framework, and Open Challengescited by 3
  38. 2026FAR-Drive: Frame-AutoRegressive Video Generation in Closed-Loop Autonomous Drivingcited by 3

50 most cited of 115 All 115 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025LongDWM: Cross-Granularity Distillation for Building a Long-Term Driving World Model
  2. 2025EP-Diffuser: An Efficient Diffusion Model for Traffic Scene Generation and Prediction via Polynomial Representations
  3. 2025ProphetDWM: A Driving World Model for Rolling Out Future Actions and Videos
  4. 2026ForgeDrive: Bidirectional Cross-Conditioning for Unified Visual-Action Generation in Autonomous Driving
  5. 2025A Survey of World Models for Autonomous Driving
  6. 2024DreamForge: Motion-Aware Autoregressive Video Generation for Multi-View Driving Scenes
  7. 2025Vid2World: Crafting Video Diffusion Models to Interactive World Models
  8. 2026Drive-HWM: Hierarchical World Models for Dynamic-Latent Guided Autonomous Driving