Reproduce “Slide2Text: Leveraging LLMs for Personalized Textbook Generation from PowerPoint Presentations” 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
Yizhou Zhou
The Slide2Text system, developed at East China Normal University, automates the conversion of PowerPoint presentations into personalized, high-quality textbooks using Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG). It demonstrated multilingual output (English and Japanese) and customization of content and style, while establishing an evaluation framework to assess content quality and pedagogical impact.
Key insights
- LLMs can effectively generate structured and personalized educational content from semi-structured input like PowerPoint presentations, offering customization in style, difficulty, and language.
- Combining LLMs with RAG techniques significantly improves the accuracy and scholarly depth of generated educational materials by integrating authoritative external knowledge.
- A human-AI collaborative evaluation framework is crucial for ensuring the quality, cultural inclusivity, and pedagogical appropriateness of automatically generated educational content.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.