Learning Engagement Assistant (LEA): Cross-Course Scalability and Classroom Evaluation of an Agentic AI Tutoring System
3.20T1 sourcearXiv cs.HC
Source record
Published by arXiv cs.HC (T1 source). The original is at https://arxiv.org/abs/2607.13370.
Pipeline notes
The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.
SummaryLEA is an adaptive AI tutoring agent combining retrieval-augmented generation with knowledge component models, offering Chat, Tutor, and Quiz modes. A real classroom deployment (n=8) and cross-course evaluation across three courses found simulation predictions diverged from real use, with answer relevancy stable (0.88–0.94) but faithfulness declining with curriculum distance (0.69 to 0.50).
Why it mattersReports a real classroom deployment of an agentic tutoring system with concrete cross-course RAGAS metrics, showing where simulation predictions break down. Useful for builders of multi-agent learning tools, though the sample is small.
Cited by
No citations on record.