NexaNota: An AI-Powered Smart Linked Lecture Note-Taking System Leveraging Large Language Models
Mi Jiang, Junran Gao, Zeyu Pan, Yue Wu, Zile Wang · 2025
Taking lecture note is an essential pedagogical method for aiding memory and understanding. Compared to unstructured lecture notes, well-structured lecture notes are highly beneficial that students would read frequently for reviewing and organizing their idea. However, it can be challenging to effectively take notes while listen to the lecturer simultaneously in a lecture, not to mention recall the whole lecture content after class. Particularly when the lecture content involves complex topics and intricates connections between topics, students feel vulnerable to review the lecture. To solve these difficulties, we design and propose an automated note-taking system, NexaNota. The system leverages the advanced Large Language Model (LLMs) to generate smart linked and structured lecture notes by constructing topics network through knowledge graphs, and providing additional web resources to complement each topic. In a within-subjects study with 17 participants (12 students and 5 experts), we found that NexaNota generates highly organized notes with 97.7% accuracy in topic identification and 86.7% accuracy in the connections between topics. Our results suggest that NexaNota enhances student learning efficiency by providing smart-linked, high-quality lecture notes.