Artificial Intelligence for finance

ECTS : 3

Description du contenu de l'enseignement :

Course Description

This course introduces Generative AI (Gen AI) through its main models, algorithms and applications, then explores its applications in the financial sector. 

The Gen AI architectures and frameworks that will studied are Transformers, Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs).

Large Language Models (LLMs) being the most popular generative model in Gen AI, they will be introduced and deeply analyzed. 

By the end of this course, students will have a good understanding of these architectures, frameworks, and models, and be able to use them to solve real-world financial problems.

Session 1: An Introduction to Generative AI

• Topics:      
  • An       overview of Generative AI (Definitions, models, applications).
  • Neural       architectures for Generative AI : from Recurrent neural networks to       Transformers (RNN àLSTM       & GRU à       Encoder-Decoder à       Attention mechanism à       Transformers). 

Session 2: Large Model Languages (LLMs)

• Topics:      
  • What       is a generative model?
  • What       is a language model? 
  • From       statistical to neural LMs.
  • Training       in LLMs.
  • Important       examples of LLMs: Bert, GPT, LLMA.

Session 3: Regulatory Compliance, Risk Management, Fraud Detection and Reporting

• Objective:      Understand the application of generative AI in ensuring regulatory      compliance and automating reporting tasks.
• Topics      Covered:
  • AI-driven       automation of regulatory reporting
  • Anomaly       detection using AI and LLMs
  • Interpretation       of complex legal texts and regulations using LLMs
  • Compliance       monitoring and auditing with AI

 Session 4: Financial Forecasting, Economic Modeling and Portfolio Management

• Objective:      Explore the use of generative AI for economic modeling and financial      forecasting.

 Session 5: ESG Sentiment Analysis and Market Research

• Objective:      Delve into how LLMs are used for ESG sentiment analysis and market      research to inform financial decisions.

 Session 6: Final Project and Case Studies

• Objective:      Apply knowledge through a final project and review case studies of AI      applications in finance.

 Conclusion:

• Recap:      Overview of the applications covered and potential future developments.
• Discussion:      Ethical and practical considerations in the deployment of AI in

  Compétence à acquérir :

  Course Objectives

  • Understand      the theoretical foundations of generative AI models.
  • Apply      generative models to financial data analysis and prediction.
  • Evaluate      ethical considerations and challenges of AI in finance.

  Mode de contrôle des connaissances :

  A comprehensive project applying Gen AI to a financial problem.

  Bibliographie, lectures recommandées :

    

  1. "Deep      Learning" by Ian Goodfellow, Yoshua Bengio, and Aaron Courville
  2. "      GANs in Action: Deep learning with Generative Adversarial Networks "      by Jakub Langr and Vladimir Bok
  3. "Reinforcement      Learning: An Introduction" by Richard S. Sutton and Andrew G. Barto
  4. G. Neubig, Neural      Machine Translation and Sequence-to-sequence Models : A Tutorial (2017). arXiv:      1703.01619. https://doi.org/10.48550/arXiv.1703.01619
  5. A. Graves,      Generating Sequences With Recurrent Neural Networks (2014). arXiv:      1308.0850.      https://doi.org/10.48550/arXiv.1308.0850
  6. Understanding VAE:      https://towardsdatascience.com/understanding-variational-autoencoders-vaes-f70510919f73
  7. Understanding      Attention in Deep Learning https://towardsdatascience.com/attaining-attention-in-deep-learning-a712f93bdb1e
  8. Transformers: https://towardsdatascience.com/transformers-89034557de14.
  9. Selected      research articles and case studies