LearnSmart: A Tool to Extract Key Concepts and Arguments From YouTube Lectures
LearnSmart is an AI-powered platform built to make it easier for learners to process and understand the vast amount of digital content available today. It automatically finds relevant YouTube videos based on a user's query and uses advanced natural language processing (NLP) techniques - like clustering and Retrieval-Augmented Generation - to extract and organize key ideas. These insights are then transformed into concise articles with the help of Mistral AI, narrated video summaries using text-to-video technology, and neatly structured PowerPoint presentations converted directly from PDF documents. To handle content efficiently, LearnSmart uses Mini Self-Organizing Map for grouping similar topics and Facebook Artificial Intelligence Similarity Search for fast, vector-based searching. The system delivers strong performance, generating articles in about 26.81 seconds and loading images in just over 33 seconds. While clustering shows a low silhouette score, it still effectively handles overlapping ideas. By offering content in multiple formats-text, video, and slides, LearnSmart helps reduce information overload and supports more effective, flexible learning. Ongoing improvements will focus on enhancing NLP accuracy and expanding the platform’s ability to support a wider range of educational needs.
Authors
- Paritosh Joshi (ORCID: https://orcid.org/0000-0001-5209-0608)
- Ayush R Gerra
- Lintomon S Chirakkara
- Sania Khan
- Harsh Dhebe
Publication Details
- Journal
- Cureus Journal of Computer Science.
- Published
- 2025-09-21
- DOI
- https://doi.org/10.7759/s44389-025-03599-0
- Primary Topic
- Online Learning and Analytics
- Type
- article
- Field-Weighted Citation Impact
- 0.00