Preacher: Paper-to-Video Agentic System

2025-08-13

AI Summary

Current video generation models face critical limitations—including narrow context windows, restricted output duration, monolithic stylistic outputs, and imprecise domain-knowledge representation—when converting academic papers into structured video summaries. To address these challenges, we propose the first agent-based system specifically designed for the “paper-to-video” task, adopting a two-stage paradigm that synergistically integrates top-down content decomposition with bottom-up clip generation. Our approach innovatively introduces key-scene definition and Progressive Chain-of-Thought (P-CoT) reasoning to enable fine-grained cross-modal alignment and accurate domain-specific knowledge modeling. The system unifies large language model–driven reasoning, multi-granularity summarization, controllable video generation, and compositional synthesis, supporting end-to-end task planning and content orchestration. Evaluated across five academic disciplines, our generated video summaries demonstrate statistically significant improvements over baselines in domain expertise, narrative coherence, and stylistic diversity.

Technology Category

• Computer Vision: Video Understanding & Activity Analysis
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Multiagent Systems: Multiagent Learning

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

The paper-to-video task converts a research paper into a structured video abstract, distilling key concepts, methods, and conclusions into an accessible, well-organized format. While state-of-the-art video generation models demonstrate potential, they are constrained by limited context windows, rigid video duration constraints, limited stylistic diversity, and an inability to represent domain-specific knowledge. To address these limitations, we introduce Preacher, the first paper-to-video agentic system. Preacher employs a top- down approach to decompose, summarize, and reformulate the paper, followed by bottom-up video generation, syn- thesizing diverse video segments into a coherent abstract. To align cross-modal representations, we define key scenes and introduce a Progressive Chain of Thought (P-CoT) for granular, iterative planning. Preacher successfully gener- ates high-quality video abstracts across five research fields, demonstrating expertise beyond current video generation models. Code will be released at: https://github.com/Gen- Verse/Paper2Video

Problem

• Converts research papers into structured video abstracts
• Overcomes limitations of current video generation models
• Aligns cross-modal representations for coherent video synthesis

Innovation

• Top-down paper decomposition and summarization
• Bottom-up diverse video segment synthesis
• Progressive Chain of Thought for planning