AI in Finance

This course is designed to equip finance professionals with the practical tools and techniques to integrate AI into their daily workflows without needing deep data science expertise. The focus is on hands-on learning, utilizing AutoML, explainability tools, and generative AI to solve real-world financial problems.

Participants will work on four practical challenges, covering AI applications in trading strategies, risk management, and AI copilots for automated financial insights.

By the end of the course, attendees will be able to immediately apply AI-driven techniques in their workplace, leveraging AI models for decision-making, automation, and analysis.  

• Finance professionals (traders, analysts, risk managers, portfolio managers) looking to leverage AI without becoming data scientists
• Professionals working in asset management, banking, and corporate finance who want to automate financial tasks, improve efficiency, and enhance decision-making
• Teams interested in AutoML, explainability, and AI-driven analytics to optimise their existing processes
• Anyone who wants to apply AI tools right away without needing deep coding or mathematical knowledge

• Build a complete bank capital stress test model, encompassing both econometric and fundamental models of retail and corporate credit risk, market risk and operational risk
• Learn how to apply the model for any of Internal Capital Adequacy Assessment Process (ICAAP), external supervisor-driven stress tests or investor-driven stress tests
• Review the various approaches taken by different banks and supervisors in their capital stress testing, from a range of European, US and Asian banks

• Basic Python knowledge (e.g., understanding variables, loops, functions, and basic data manipulation with Pandas)
• No prior machine learning or AI experience is required – the course is structured to introduce and apply AI concepts in a user-friendly manner
• Familiarity with Excel, financial data, and basic statistics is helpful but not mandatory

Welcome & Overview

• Introduction to the course structure and objectives
• Overview of AI applications in finance

AI Fundamentals for Finance

• Basics of AI/ML concepts (supervised, unsupervised learning, regression, classification)
• Understanding data types and the importance of data preparation in finance
• <

Practical 1: Data Preparation for AI Models

• Hands-on session: Participants will work with a dataset (e.g., stock prices or macroeconomic indicators)
• Using Python libraries like pandas for cleaning and preprocessing
• Feature engineering and dealing with missing values

Introduction to AutoML Tools

• Overview of AutoML and its relevance for finance professionals
• Introduction to popular AutoML tools (e.g., H2O AutoML, Azure Machine Learning)
• Demo: Building a model using AutoML without deep coding

Practical 2: Predicting Stock Prices with AutoML

• Hands-on session: Using an AutoML tool (Azure) to predict stock price movements
• Step-by-step guidance on setting up a project and evaluating model performance
• Discussion on the results and how to interpret them

Explainability in AI Models

• Importance of explainability in finance
• Overview of techniques like SHAP, LIME, and model interpretability
• Real-world examples of how explainability aids in regulatory compliance and decision-making

Practical 3: Model Explainability with SHAP

• Hands-on session: Using SHAP to interpret model predictions
• Application to a financial dataset (e.g., credit risk or loan default prediction)
• Understanding which features drive predictions and how to communicate results

Generative AI and Copilot Creation

• Overview of Generative AI (e.g., GPT models) and its applications in finance
• Building a simple financial Copilot (like Microsoft Copilot) using a pre-trained model
• Use cases in finance: generating reports, answering financial queries, and automating document generation
• Demo: Developing a Copilot for financial Q&A or report generation

Practical 4: Developing a Financial Copilot

• Hands-on session: Participants will create their own basic financial Copilot using a pre-trained LLM (e.g., GPT-3.5 or GPT-4)
• Integration with financial data for real-time query handling
• Discuss how this type of generative AI can be applied to various financial tasks

Recap

• Review of key concepts and practical sessions
• Q&A and feedback session