Research on a Vertical-Domain Personalized Learning System Based on RAG and Multi-Agent Collaboration
Journal of Higher Education Teaching · 2026-08 · 3 (4) · 148-157
DOI: 10.62517/jhet.202615423 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Higher Education Teaching, 2026-08. Verfasst von Chao Geng und 5 weiteren Autor:innen. Die Arbeit nennt 16 Literaturangaben.
Abstract (Originalsprache)
Specialized courses frequently encounter fragmented knowledge resources, insufficient adaptation to individual learning needs, and delayed feedback. This work presents a personalized learning system for vertical domains that combines Retrieval-Augmented Generation (RAG) with a multi-agent framework. Learner profiles incorporate prior knowledge, learning goals, weak knowledge points, and preferred difficulty levels, while a domain-specific knowledge base supplies external retrieval content. During resource generation, retrieved knowledge and learner profile data jointly inform the production of reading materials, quizzes, practical tasks, and animation scripts. A review agent evaluates generated resources prior to learner delivery. Learning records—including quiz results, coding performance, task completion, and learning time—update learner profiles and guide subsequent resource and path adjustments. Experimental results show that the proposed system achieves better resource adaptability than direct LLM generation and supports personalized learning in vertical domains.
Zitieren
APA
Geng, C., Fan, X., Zheng, Y., Zhang, R., Zhao, Y., & Gong, M. (2026). Research on a Vertical-Domain Personalized Learning System Based on RAG and Multi-Agent Collaboration. Journal of Higher Education Teaching, 3(4), 148-157. https://doi.org/10.62517/jhet.202615423
Vancouver
Geng C, Fan X, Zheng Y, Zhang R, Zhao Y, Gong M. Research on a Vertical-Domain Personalized Learning System Based on RAG and Multi-Agent Collaboration. Journal of Higher Education Teaching. 2026;3(4):148-157. doi:10.62517/jhet.202615423
BibTeX
@article{Geng2026,
author = {Geng, Chao and Fan, Xiaojun and Zheng, Yonghui and Zhang, Ruiheng and Zhao, Yawen and Gong, Mingming},
title = {Research on a Vertical-Domain Personalized Learning System Based on RAG and Multi-Agent Collaboration},
journal = {Journal of Higher Education Teaching},
year = {2026},
volume = {3},
number = {4},
pages = {148--157},
doi = {10.62517/jhet.202615423}
}
EUVIMED