Downloads

https://doi.org/10.62836/iaet.v5i1.0005
Zhao, S., Zhong, Q., Qin, P., Xin, Y., Huo, Z., Zhang, G., Li, H., Xu, M., & Guo, J. (2026). An Industry-Oriented Curriculum Framework for Graduate Education in Quantitative Finance: Integrating Large Language Models and Spatio-Temporal Wavelet Networks. Innovations in Applied Engineering and Technology, 5(1), 0005. https://doi.org/10.62836/iaet.v5i1.0005
Volume 5, Issue 1, 2026

An Industry-Oriented Curriculum Framework for Graduate Education in Quantitative Finance: Integrating Large Language Models and Spatio-Temporal Wavelet Networks

The gap between the standard curriculum of quantitative finance and what is needed in practice has been increasing with the speed of financial artificial intelligence. The graduate students are being asked to be able to perform a single workflow involving structured market data, non-structured financial text, cross-asset dependence, non-stationary signal, reproducible backtesting, and deployable computing environments. This paper develops an industry-based curriculum framework which groups these needs on the basis of competency education and project based learning. The structure of the framework includes four interrelated components that include an industry-task and competency map, a three-tiered modular curriculum, a capstone multimodal forecasting project, and a feedback mechanism of revising the curriculum based on evidence. Factor research can be done using Microsoft Qlib; text representation and sentiment analysis can be performed by means of financial language models; discrete wavelet transform, Wavelet Kolmogorov-Arnold Networks, graph convolutional networks and Stockformer can be used to provide progressively more complex modelling operations; Linux-based deployment will allow for reproducibility, monitoring and engineering practice. It makes the difference between the technical metrics of projects, such as Information Coefficient, ICIR, Sharpe ratio, maximum drawdown, and the demonstration of student competence, such as the provenance of data, comparison of models, reliability of the deployment, reasoning about risk, and technical writing. The proposed curriculum design is reusable in terms of domain knowledge, modeling, engineering, and use of generative AI in responsible manner. Future multi-cohort application and outcome testing should prove its educational efficacy.

graduate education in quantitative finance curriculum reform industry-oriented education large language models spatio-temporal modelling project-based learning AI engineering

References

  1. He Z. The Difference between Academic Degree and Professional Degree. Chinese Graduate Student 2007; 2: 52.
  2. Wyne MF, Zhang L, Farahani A. A Review of Competency Based Education (CBE). In Proceedings of the 2021 International e-Engineering Education Services Conference, Petra, Jordan, 22–23 June 2021; pp. 127–132.
  3. Lavi R, Bathe M, Hosoi A, et al. The NEET Ways of Thinking: Implementing Them at MIT and Assessing Their Efficacy. Advances in Engineering Education 2021.
  4. Liu Y, Xiang G, Wang J. Research on the Integration Model of Industry and Education in Applied Undergraduate Colleges and Its Influencing Factors. China Higher Education Research 2015; 5: 64–68.
  5. Yang X, Liu W, Zhou D, et al. Qlib: An AI-Oriented Quantitative Investment Platform. arXiv 2020. arXiv:2009.11189.
  6. Ke G, Meng Q, Finley T, et al. LightGBM: A Highly Efficient Gradient Boosting Decision Tree. Advances in Neural Information Processing Systems 2017; 30.
  7. Mallat S. A Wavelet Tour of Signal Processing, 2nd ed.; Academic Press: Cambridge, MA, USA, 1999.
  8. Liu Z, Wang Y, Vaidya S, et al. KAN: Kolmogorov-Arnold Networks. International Conference on Learning Representations 2025; 2025: 70367–70413.
  9. Bozorgasl Z, Chen H. Wav-KAN: Wavelet Kolmogorov-Arnold Networks. arXiv 2024. arXiv:2405.12832.
  10. Kipf TN, Welling M. Semi-Supervised Classification with Graph Convolutional Networks. arXiv 2016. arXiv:1609.02907.
  11. Ma B, Xue Y, Lu Y, et al. Stockformer: A Price-Volume Factor Stock Selection Model Based on Wavelet Transform and Multi-Task Self-Attention Networks. Expert Systems with Applications 2025; 273: 126803. https://doi.org/10.1016/j.eswa.2025.126803.
  12. Araci D. FinBERT: Financial Sentiment Analysis with Pre-Trained Language Models. arXiv 2019. arXiv:1908.10063.
  13. Wu S, Irsoy O, Lu S, et al. BloombergGPT: A Large Language Model for Finance. arXiv 2023. arXiv:2303.17564.

Supporting Agencies

  1. Funding: The study was funded by the Henan Province Graduate Education Reform and Quality Enhancement Project, “Machine Learning” (Approval No. YJS2025XQLH10), the Henan Provincial Key Science and Technology Research Program (No. 262102211082), and the Postgraduate Education Reform and Quality Improvement Project of Henan Province (No. YJS2026ZYJC12).