Reproduce “OmniPresent: Generating Coherent Presentation Suites from Scientific Papers” 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
Qianli Ma, Jipeng Xiao, Siyu Wang, Zhiheng Tian, Wangyu Feng, Shibo Wang, Chang Guo, Shuochen Chang and 2 more
Researchers from Shanghai Jiao Tong University introduced OmniPresent, a framework designed to generate coherent presentation suites (project page, poster, slides, video) from scientific papers, demonstrating superior performance in content faithfulness and visual quality compared to specialized baselines and human-authored artifacts on the OmniPreBench dataset.
Key insights
- A unified framework with a centralized knowledge source and cross-artifact verification significantly enhances the quality of individual presentation artifacts and the overall coherence of the entire suite.
- Utilizing renderable HTML as an intermediate representation facilitates precise, code-level content verification and enables flexible rendering across diverse static (posters, slides) and dynamic (videos, project pages) formats.
- The 'Plan-Verify-Render' architecture provides a robust constraint framework that can improve performance even with less capable underlying LLM backbones, acting as a system-level enhancement that bounds the stochastic nature of language models.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.