Video World Models · 2026-10-15
Generative World Simulation with DAO and JING
XGEN Team, Guohao Cai, Haoxuan Che, Frank Fan, Chenran Huang, Junjia Huang, Jiayi Jiang, Yanheng Li, Leon Liu, Shaoteng Liu, Xi Liu, Feng Nie, Zhefan Rao, An Wang, Hanwen Wang, Pengxiang Yan, Jieming Yu, Qian Zha, Wentao Zheng
AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.
TL;DR
The authors present XGEN, a Generative World Simulation framework that couples DAO, a computable persistent shared world engine, with JING, an egocentric interactive audiovisual generator. DAO maintains characters/objects/spaces and updates them from actions and world time; it then derives observer-specific conditions for JING. Together, they form a continuous loop where action consequences persist and shape future first-person experiences.
Why it matters
The paper targets an open challenge in interactive world modeling: keeping an explicit shared world state that can be updated by actions, queried across observers and time, and used to generate each observer’s local audiovisual experience. The DAO+JING coupling connects world-state evolution with experience generation so that past actions can influence what is seen after revisits and absences.
Method
- DAO: explicitly represents characters, objects, spaces, and their relationships, advances shared state via interaction rules and world time, and preserves consequences of past actions.
- JING: an egocentric interactive model that generates continuous first-person audiovisual experiences autoregressively from reference conditions, observation history, and current actions (with camera trajectory control).
- Closed loop: actions update the shared DAO state; DAO converts shared state into observer-specific conditions for JING, grounding generation in persistent world history.
Limitation
Future work will focus on improving motion robustness and long-term memory, reducing response latency, and strengthening consistency between shared world state and generated audiovisual content.
Abstract (from arXiv)
We present a Generative World Simulation framework that couples persistent shared world state with continuous first-person audiovisual generation. The framework integrates DAO, a computable world engine, and JING, an Egocentric Interactive Experience Model, to support sustained interaction for humans and agents. DAO maintains the states of characters, objects, and spaces and their relationships, advancing the world according to interaction rules and world time while preserving the consequences of past actions. From this evolving state, DAO derives observer-specific conditions for JING, which combines these conditions with character, object, and scene references, observation history, and current actions to generate each observer's audiovisual experience. Users can explore environments, manipulate objects, and communicate with characters through viewpoint control and text instructions, with their actions shaping both the shared state and subsequent experiences. By connecting world-state evolution with experience generation in a continuous interaction loop, the framework offers an approach to building interactive worlds that retain their history, evolve through action, and persist beyond any individual observer.
Related papers
- Oneira: From Open-Ended Generation to Open-World Interaction in Video World Models
- RIGen: Beyond Action Following with Reciprocal Interaction in Generative Gameplay
- Learning Interactive Real-World Simulators
- World Editing: Intervening on Executable Worlds at Increasing Depth
- Genie: Generative Interactive Environments
Publication record
Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.
First public release: 2026-10-14 · Indexed by Pub-AI: 2026-10-15 UTC
Generative World Simulation with DAO and JING
Preprint posted: 2026-10-14 · Original record ↗
- arxiv
- 2610.19726
Explicit source versions (1)
- v1 · 2026-10-14
- Preprint available
Source date: 2026-10-14 · First detected: 2026-10-25 UTC · Historical baseline
Source evidence (1)
- Official evidence ↗ · atom:published (official metadata)
Abstract · arxiv
We present a Generative World Simulation framework that couples persistent shared world state with continuous first-person audiovisual generation. The framework integrates DAO, a computable world engine, and JING, an Egocentric Interactive Experience Model, to support sustained interaction for humans and agents. DAO maintains the states of characters, objects, and spaces and their relationships, advancing the world according to interaction rules and world time while preserving the consequences of past actions. From this evolving state, DAO derives observer-specific conditions for JING, which combines these conditions with character, object, and scene references, observation history, and current actions to generate each observer's audiovisual experience. Users can explore environments, manipulate objects, and communicate with characters through viewpoint control and text instructions, with their actions shaping both the shared state and subsequent experiences. By connecting world-state evolution with experience generation in a continuous interaction loop, the framework offers an approach to building interactive worlds that retain their history, evolve through action, and persist beyond any individual observer.