FIN 608: Artificial Intelligence in Finance
Contents
The course provides an application-oriented understanding of Artificial Intelligence in finance. A first, methodological part compactly covers the core technical foundations — classical machine learning, deep learning, transformer architectures, and large language models, including prompting, retrieval-augmented generation, agent-based systems, and evaluation methods — consistently with a focus on financial application. The main emphasis is then on concrete applications across central financial domains: machine learning in asset pricing and trading, natural language processing for financial documents and earnings calls, AI in banking, credit, risk, and RegTech, as well as the use of generative AI in capital markets. Two dedicated application-oriented sessions show how AI can be applied specifically to the finance research workflow (AI for Research) and to the development of finance-related ventures (AI for Ventures). Selected guest lectures from industry complement the course. As the final deliverable, each team works in small groups on either an independent empirical research project with extensive use of AI or a deeply developed entrepreneurial project in the fintech space.
Learning outcomes
Upon successful completion of this course, students
- understand the methodological foundations of modern AI techniques — from classical machine learning through deep learning and transformers to large language models and agent-based systems — and can explain their inner workings and limitations;
- are able to map AI methods to central financial application areas (asset pricing, trading, text analytics, credit risk, RegTech) and critically evaluate them;
- can design and execute AI-assisted research and product-development workflows independently, methodologically validating and reproducibly documenting the results;
- have acquired the ability to plan, execute, and professionally present a substantive own project at a high level.
Necessary prerequisites
–
Recommended prerequisites
Prior coursework in finance at bachelor level, basic familiarity with empirical methods, and fundamental knowledge of Python.
| Forms of teaching and learning | Contact hours | Independent study time |
|---|---|---|
| Lecture with integrated exercise | 4 SWS | 13 SWS |
| ECTS credits | 6 |
| Graded | yes |
| Workload | 180h |
| Language | English |
| Form of assessment | Final presentation (approx. 30 min per team, conducted in a single block of approx. 5h at the end of the semester, 30 %) and final submission (70 %). The final submission consists, depending on the chosen track, either of an empirical research paper of 15–20 p., including replication code and AI workflow documentation, or of a functioning finance-related MVP, including a concept document (problem, solution, market, business model, technical architecture). |
| Restricted admission | yes |
| Further information | Registration via Portal2 |
Examiner Performing lecturer |  | Prof. Dr. Erik Theissen Paul Seidel, M.Sc. |
| Frequency of offering | Fall semester |
| Duration of module | 1 semester |
| Range of application | M.Sc. MMM, M.Sc. Bus. Edu., M.Sc. Econ., M.Sc. Bus. Inf., M.Sc. Bus. Math., M.Sc. MMFACT, M.Sc. MMOSCM |
| Preliminary course work | – |
| Program-specific Competency Goals | CG 1, CG 4 |
| Literature | A continuously updated selection, complemented by current materials: Methodological foundations (machine learning, deep learning, LLMs): • Goodfellow, I., Bengio, Y., Courville, A. (2016). Deep Learning. MIT Press (selected chapters). • Vaswani, A. et al. (2017). Attention Is All You Need. NeurIPS (original transformer paper). • Brown, T. et al. (2020). Language Models are Few-Shot Learners (GPT-3 paper). NeurIPS. • Hastie, T., Tibshirani, R., Friedman, J. (2009). The Elements of Statistical Learning. Springer (refresher). • Current technical documentation by Anthropic and OpenAI on LLM agents, retrieval-augmented generation, tool use, and evaluation frameworks. AI in asset pricing and trading: • Gu, S., Kelly, B., Xiu, D. (2020). Empirical Asset Pricing via Machine Learning. Review of Financial Studies. • Cong, L. W., Liang, A., Zhang, X. (2021). AlphaPortfolio. Working Paper. • López de Prado, M. (2018). Advances in Financial Machine Learning. Wiley (esp. backtesting pitfalls). NLP for finance and LLM applications: • Loughran, T., McDonald, B. (2011). When Is a Liability Not a Liability? Textual Analysis, Dictionaries, and 10-Ks. Journal of Finance. • Lopez-Lira, A., Tang, Y. (2024). Can ChatGPT Forecast Stock Price Movements? Return Predictability and Large Language Models. Working Paper. • Eisfeldt, A. L., Schubert, G., Zhang, M. B. (2023). Generative AI and Firm Values. Working Paper. • Yang, Y., Uy, M. C. S., Huang, A. (2020). FinBERT: A Pretrained Language Model for Financial Communications. Additional sources, code repositories, and current industry reports will be distributed at the start of the course. |
| Course outline | The course comprises 13 sessions. Eleven regular lectures cover the methodological and application-oriented foundations, and two follow-up sessions are explicitly application-oriented (AI for Research, AI for Ventures). It is planned to complement the course with sessions featuring guest lectures from industry. Methodological foundations and financial applications • Session 1: Kickoff and AI in Finance Landscape — taxonomy (predictive ML, NLP, generative, agentic), application map, course organisation • Session 2: Machine Learning Foundations for Finance — supervised learning, regression/ • Session 3: Deep Learning and Transformer Architectures — neural networks, embeddings, attention, encoder/ • Session 4: Large Language Models in Practice — prompt engineering, retrieval-augmented generation, function calling, agent-based systems, evaluation methods, tooling landscape • Session 5: Machine Learning in Asset Pricing and Trading — cross-sectional return prediction, factor models with ML, backtesting pitfalls • Session 6: Natural Language Processing for Finance — earnings calls, 10-K filings, news sentiment, FinBERT, LLM-based asset studies • Session 7: AI in Banking, Credit and Risk — credit scoring, fraud detection, agent-based compliance, RegTech • Session 8: Generative AI in Capital Markets — analyst workflows, synthetic data, automation of sell-side and buy-side processes Application-oriented sessions • Session 9: AI for Research — AI-assisted finance research workflow for own research projects (ideation, literature, empirical design, replication, writing) • Session 10: AI for Ventures — AI-assisted venture workflow for fintech projects (problem/solution, MVP patterns, GTM, regulatory environment) |