Towards AI-Assisted Video Editing - Generating Shorts from Long Videos
Massachusetts Institute of Technology via YouTube
-
58
-
- Write review
Learn EDR Internals: Research & Development From The Masters
Live Online Classes in Design, Coding & AI — Small Classes, Free Retakes
Overview
Build a Learning Habit
Download Class Central's free printable study calendar
Download for Free
Explore cutting-edge machine learning approaches to automated video editing in this conference talk that addresses the time-intensive challenge of transforming raw footage into polished content. Learn about two innovative projects that demonstrate how AI can intelligently cut, select, and rearrange long-form videos into shorter formats. Discover the TeaserGen system, which leverages pretrained large language models and language-vision models to compress 30+ minute documentaries into compelling 3-minute teasers using a narration-centered approach. Examine the REGen framework, a retrieval-embedded generation system that enables large language models to quote multimodal resources while maintaining narrative coherence. Understand how these technologies can be integrated into existing creative workflows and explore the potential for next-generation video editing interfaces that combine multimodal LLMs with retrieval-augmented generation. Gain insights into the broader vision of democratizing professional content creation by lowering barriers to entry and augmenting human creativity with AI-powered tools designed specifically for video creators and content producers.
Syllabus
Hao-Wen (Herman) Dong - Towards AI-assisted Video Editing: Generating Shorts from Long Videos
Taught by
MIT Embodied Intelligence