Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Generative Teaching via Code

Yuheng Wang, Runde Yang, Lin Wu, Jie Zhang, Jingru Fan, Ruoyu Fu, Tianle Zhou, Huatao Li, Siheng Chen, Weinan E, Chen Qian

TL;DR

Generative Teaching addresses the scalability bottleneck of manual content creation by shifting educators from content producers to high-level directors. The paper introduces TeachMaster, a code-centric multi-agent pipeline that converts pedagogical intent into interpretable, curriculum-ready videos through three stages: Content Planning, Presentation Generation, and Quality Validation. Extensive experiments across languages and disciplines show improvements in structural coherence, cross-modal alignment, and production efficiency, with real-world deployments at top universities and bilingual case studies. This approach promises scalable, adaptable education that maintains content quality while dramatically reducing production costs.

Abstract

The scalability of high-quality online education is hindered by the high costs and slow cycles of labor-intensive manual content creation. Despite advancements in video generation, current approaches often fail to ensure pedagogical structure and precise control due to their pixel-level, black-box nature. In this paper, we propose Generative Teaching, a novel paradigm that transitions educators from manual creators to high-level directors, allowing them to focus on pedagogical intent while autonomous agents handle the execution. To realize this vision, we introduce TeachMaster, a multi-agent framework that leverages code as an intermediate semantic medium. Unlike traditional video generation methods, TeachMaster orchestrates a collaborative team of agents--spanning planning, design, and rendering--to automate the production of interpretable, editable, and curriculum-ready educational videos. Experiments validate that TeachMaster significantly boosts production efficiency without compromising structural coherence or visual fidelity, providing a robust solution for scalable education.

Generative Teaching via Code

TL;DR

Generative Teaching addresses the scalability bottleneck of manual content creation by shifting educators from content producers to high-level directors. The paper introduces TeachMaster, a code-centric multi-agent pipeline that converts pedagogical intent into interpretable, curriculum-ready videos through three stages: Content Planning, Presentation Generation, and Quality Validation. Extensive experiments across languages and disciplines show improvements in structural coherence, cross-modal alignment, and production efficiency, with real-world deployments at top universities and bilingual case studies. This approach promises scalable, adaptable education that maintains content quality while dramatically reducing production costs.

Abstract

The scalability of high-quality online education is hindered by the high costs and slow cycles of labor-intensive manual content creation. Despite advancements in video generation, current approaches often fail to ensure pedagogical structure and precise control due to their pixel-level, black-box nature. In this paper, we propose Generative Teaching, a novel paradigm that transitions educators from manual creators to high-level directors, allowing them to focus on pedagogical intent while autonomous agents handle the execution. To realize this vision, we introduce TeachMaster, a multi-agent framework that leverages code as an intermediate semantic medium. Unlike traditional video generation methods, TeachMaster orchestrates a collaborative team of agents--spanning planning, design, and rendering--to automate the production of interpretable, editable, and curriculum-ready educational videos. Experiments validate that TeachMaster significantly boosts production efficiency without compromising structural coherence or visual fidelity, providing a robust solution for scalable education.
Paper Structure (11 sections, 10 equations, 4 figures, 3 tables)

This paper contains 11 sections, 10 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. TeachMaster
  3. Content Planning
  4. Presentation Generation
  5. Quality Validation
  6. Evaluation
  7. Qualitative Analysis
  8. Real-World Applications
  9. Case Study
  10. Related Work
  11. Conclusion

Figures (4)

  • Figure 1: Under the paradigm of Generative Teaching, TeachMaster is a code-centric multi-agent framework that transforms abstract pedagogical intent into ready-to-teach videos for seamless classroom integration.
  • Figure 2: TeachMaster automates educational content generation through three stages—content planning, presentation generation, and quality validation —transforming teacher intent into coherent, multimodal educational materials.
  • Figure 3: Analysis of Real-World Classroom Feedback.
  • Figure 4: Examples of TeachMaster-generated bilingual course materials across multiple disciplines and languages. The system transforms textual outlines into multimodal teaching materials (including animated visuals, narration, voiceovers, and other customizable configurations).