Reproduce “Generative World Simulation with DAO and JING” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
XGEN Team, Guohao Cai, Haoxuan Che, Frank Fan, Chenran Huang, Junjia Huang, Jiayi Jiang, Yanheng Li and 11 more
XGEN separates shared-state updates from observer-specific audiovisual generation: DAO maintains the world, while JING generates a first-person view conditioned on it. In a bakery demonstration, an attendant extinguishes a burning napkin while the observer is away, changing the return view. On WBench Full, four-step AR scores 79.9, second of 15; this benchmark does not measure persistence or real-time latency.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.