Design of an Adaptive Tutoring System Based on Retrieval-Enhanced Generation and Dynamic Profiling: Promoting Educational Equity

Yongzhen Ju

doi:10.62051/ijcsit.v8n4.08

Authors

Yongzhen Ju

DOI:

https://doi.org/10.62051/ijcsit.v8n4.08

Keywords:

Resource balancing, LLM, Retrieval-Augmented, Profiling modeling, Adaptive education, Corpus construction

Abstract

Against the backdrop of Education Informatization 2.0 and the scarcity of high-quality educational resources in rural areas due to urban-rural disparities in China, this study addresses the challenges faced by education products based on general-purpose large models. These challenges include the unavoidable occurrence of "knowledge hallucinations" and difficulties in accurately matching teaching content. this paper employs RAG algorithms and dynamic student profiling to establish a personalized learning assistance platform. Integrating methods such as OCR and FAISS, it constructs real-time profiles and utilizes a dual-model collaborative operation mode for Q&A. A precise resource recommendation scheme ensures the accuracy of provided learning materials, genuinely aligning with students' individual circumstances to fulfill their personalized educational needs.

Downloads

Download data is not yet available.

References

[1] IFLYTEK Smart Education. (2024) Statistical Report on China's Basic Education Development in 2024. iFLYTEK Co., Ltd., Hefei.

[2] Guo, J., Rong, Q. (2023) Artificial Intelligence Empowering Educational Equity: International Consensus, Practical Obstacles, and Implementation Pathways. Modern Educational Technology, 33: 5–13.

[3] Wang, C., Li, M.H. (2023) Ethical Dilemmas and Pathway Optimization in Promoting Educational Equity through Artificial Intelligence. Educational Research, 44: 95–104.

[4] Ministry of Education. (2018) Education Informatization 2.0 Action Plan.

[5] Lewis, P., Perez, E., Piktus, A., et al. (2020) Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks. In: Advances in Neural Information Processing Systems. pp. 9459–9474.

[6] Zhang, S., Roller, S., Goyal, N., et al. (2023) Retrieval-augmented generation for conversational search. Transactions of the Association for Computational Linguistics, 11: 765–781.

[7] Piech, C., Bassen, J., Huang, J., et al. (2015) Deep knowledge tracing. In: Advances in Neural Information Processing Systems. pp. 505–513.

[8] Zhao, L., Chen, J. (2022) Construction and application of dynamic student profiles based on multi-source data. China Distance Education: 45–52+79.

[9] Chen, M., Wang, L. (2024) Application Research of Retrieval-Augmented Generation Technology in Intelligent Educational Question-Answering Systems. Modern Educational Technology, 34: 78–85.

[10] Reimers, N., Gurevych, I. (2019) Sentence-BERT: Sentence embeddings using siamese BERT-networks. In: Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). pp. 3982–3992.

[11] Liu, J., Wu, M. (2023) Optimized Application of Redis Caching Technology in Intelligent Education Systems. Information Technology Education: 67–70.

[12] Meng, K., Bau, D., Andonian, A., et al. (2022) Locating and editing factual knowledge in GPT. In: International Conference on Machine Learning. PMLR. pp. 15964–15975.

[13] Johnson, J., Douze, M., Jégou, H. (2019) Billion-scale similarity search with GPUs. IEEE Transactions on Big Data, 7: 535–547.

[14] Li, Y., Zhang, L. (2022) Research on AI Solutions for Balanced Allocation of Urban and Rural Educational Resources. Chinese Journal of Educational Technology: 86–92.

[15] Li, H., Zhang, Y., Liu, X. (2021) An optimized OCR-based method for digitizing printed textbooks with mathematical formulas. Computers Education, 171: 104178.