Researchers have developed a chatbot designed to answer student questions using only course-specific lecture videos, providing timestamped citations for its answers. During a semester-long deployment, the chatbot successfully retrieved information from the active course materials, with 70.5% of messages including citations. The system was designed to decline answering when no relevant lecture evidence was found, and students found the citations to be the most useful feature, while requesting more practice-question generation capabilities. The chatbot also demonstrated improved lecture retrieval accuracy compared to standard methods on the EduVidQA benchmark. AI
IMPACT This system demonstrates a practical application of retrieval-augmented generation for educational tools, potentially improving student study habits and instructor support.
RANK_REASON This is a research paper detailing a novel chatbot system for educational purposes. [lever_c_demoted from research: ic=1 ai=1.0]
Read on arXiv cs.IR (Information Retrieval) →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →