# Marionette: Predicting World States, Rendering Geometry, Painting Appearance

arXiv: https://arxiv.org/abs/2608.14530 · HTML: https://mlsanity.com/paper/2608.14530/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Placed in Video-generation world models and simulators, one of 11 directions in World models.
- Builds on 53 of its 102 references that are in the map.
- First posted on arXiv 14 Aug 2026.

## Placed in

- World models / [Video-generation world models and simulators](https://mlsanity.com/wm/video-generation-world-models-and-simulators/)

## Builds on (most cited first, 30 of 53)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Toolformer: Language Models Can Teach Themselves to Use Tools](https://mlsanity.com/paper/2302.04761/) (2023)
- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [VBench: Comprehensive Benchmark Suite for Video Generative Models](https://mlsanity.com/paper/2311.17982/) (2023)
- [HunyuanVideo: A Systematic Framework For Large Video Generative Models](https://mlsanity.com/paper/2412.03603/) (2024)
- [World Models](https://mlsanity.com/paper/1803.10122/) (2018)
- [Towards Accurate Generative Models of Video: A New Metric & Challenges](https://mlsanity.com/paper/1812.01717/) (2018)
- [Human Motion Diffusion Model](https://mlsanity.com/paper/2209.14916/) (2022)
- [Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks](https://mlsanity.com/paper/1506.03099/) (2015)
- [Animate Anyone: Consistent and Controllable Image-to-Video Synthesis for Character Animation](https://mlsanity.com/paper/2311.17117/) (2023)
- [PAL: Program-aided Language Models](https://mlsanity.com/paper/2211.10435/) (2022)
- [Faith and Fate: Limits of Transformers on Compositionality](https://mlsanity.com/paper/2305.18654/) (2023)
- [Diffusion Models Are Real-Time Game Engines](https://mlsanity.com/paper/2408.14837/) (2024)
- [Champ: Controllable and Consistent Human Image Animation with 3D Parametric Guidance](https://mlsanity.com/paper/2403.14781/) (2024)
- [Navigation World Models](https://mlsanity.com/paper/2412.03572/) (2024)
- [OmniControl: Control Any Joint at Any Time for Human Motion Generation](https://mlsanity.com/paper/2310.08580/) (2023)
- [Rolling Forcing: Autoregressive Long Video Diffusion in Real Time](https://mlsanity.com/paper/2509.25161/) (2025)
- [OccWorld: Learning a 3D Occupancy World Model for Autonomous Driving](https://mlsanity.com/paper/2311.16038/) (2023)
- [How Far is Video Generation from World Model: A Physical Law Perspective](https://mlsanity.com/paper/2411.02385/) (2024)
- [MimicMotion: High-Quality Human Motion Video Generation with Confidence-aware Pose Guidance](https://mlsanity.com/paper/2406.19680/) (2024)
- [Matrix-game 2.0: An open-source real-time and streaming interactive world model](https://mlsanity.com/paper/2508.13009/) (2025)
- [CamCo: Camera-Controllable 3D-Consistent Image-to-Video Generation](https://mlsanity.com/paper/2406.02509/) (2024)
- [Cosmos 3: Omnimodal World Models for Physical AI](https://mlsanity.com/paper/2606.02800/) (2026)
- [Seedance 2.0: Advancing Video Generation for World Complexity](https://mlsanity.com/paper/2604.14148/) (2026)
- [Training Agents Inside of Scalable World Models](https://mlsanity.com/paper/2509.24527/) (2025)
- [CamI2V: Camera-Controlled Image-to-Video Diffusion Model](https://mlsanity.com/paper/2410.15957/) (2024)
- [Is Sora a World Simulator? A Comprehensive Survey on General World Models and Beyond](https://mlsanity.com/paper/2405.03520/) (2024)
- [Hunyuan-GameCraft: High-dynamic Interactive Game Video Generation with Hybrid History Condition](https://mlsanity.com/paper/2506.17201/) (2025)
- [MineWorld: a Real-Time and Open-Source Interactive World Model on Minecraft](https://mlsanity.com/paper/2504.08388/) (2025)
- [WoVoGen: World Volume-aware Diffusion for Controllable Multi-camera Driving Scene Generation](https://mlsanity.com/paper/2312.02934/) (2023)

## Built on by (most cited first, 6 of 6)

- [AI for Games in the Foundation Model Era](https://mlsanity.com/paper/2609.16679/) (2026)
- [GameDirector: Decoupling Gameplay Logic from Rendering for Player-Configurable Game World Models](https://mlsanity.com/paper/2609.25652/) (2026)
- [Code Plans, Diffusion Renders: Open-Ended Generative World Modeling](https://mlsanity.com/paper/2609.26458/) (2026)
- [Proxy2World: Learning to Generate Worlds From Lightweight Proxies without Seeing Them](https://mlsanity.com/paper/2609.35023/) (2026)
- [AgentGarten: Code Worlds for Evolving Agents](https://mlsanity.com/paper/2610.12374/) (2026)
- [Generative World Simulation with DAO and JING](https://mlsanity.com/paper/2610.19726/) (2026)

## Abstract (arXiv)

Interactive game world models typically autoregress visual observations directly in pixel or latent space, forcing structured properties such as pose, geometry, and occlusion to be implicitly maintained by the same generative sequence. Over long horizons, errors in these latent world properties accumulate, making consistency and controllability fragile. We explicitly model the evolving world state, delegate exact geometric computation to a fixed, zero-parameter renderer, and leave the neural model to synthesize appearance. We instantiate this idea as Marionette, a world model for interactive games with articulated characters. First, a two-stage autoregressive dynamics model predicts an explicit and interpretable 276-dimensional 3D world state comprising multi-entity articulated skeletons, metric root trajectories, and rotations. Second, a zero-parameter graphics bridge converts the predicted state into pose-control videos, computing world-space geometry and occlusion in closed form. Third, a control-conditioned video-diffusion observation model synthesizes photorealistic RGB observations from the resulting structured controls. Our experiments establish two properties of Marionette. First, the predicted world state is directly controllable. Forcing a mismatched action stream changes root-aligned joint error by 31% across 48 held-out segments. Second, long-horizon behaviour is determined in the state, and can be repaired there. Left free, the two generated characters drift to 21.2 m apart (recorded sessions stay near 5 m) and a third of frames show ground penetration. Two rules imposed on the explicit state, a terrain collider and a separation cap, cut penetration by 66% and keep the pair engaged, with no change to the observation model. Routing appearance through the predicted state costs no fidelity we can detect, at an FVD of 831 against 799 for recorded pose.