Reproduce “ContextRenderer: Dynamic-Context Video Generation for Interactive Game Worlds” 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

Runze Wang, Fei Zha, Bing He, Ke Xu, Zhongyuan Wu, Yuxuan Xie, Wenzheng Wu, Zhongyuan Liu and 2 more

ContextRenderer renders game scenes from engine-provided proxy videos while retaining world state in the engine. For each chunk, it combines updated prompts and subject references with compressed history; training uses dynamic condition learning and shared reconstruction across proxy representations. On GWM-Bench’s visual-variation task, its 33B model scores 0.527 DOVER versus 0.477 for SolarRenderer-AR, but gains vary across metrics and benchmarks.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.