Preconference Day: Summits & WorkshopsMonday, 18 November 2024 - GMT (Greenwich Mean Time, GMTZ)

How has the use of AI in quantitative solutions developed in the past year? How are businesses using different solutions (LLMs vs GenAI) across the industry? What are the key standouts?

Use cases, regulation and supervisory topics

AI is changing the financial industry, and indeed, the world, at an ever accelerating pace. The current generation of AI models is different from the previous state of the art in important ways. Large transformer models are broadly capable – exhibiting state-of-the-art, and even human-level performance, on many tasks without problem-specific training – and uniquely accessible, allowing non-experts to interact with them using natural language.

Nonetheless, they also have important limitations. What is real, and what is hype? What are they capable of, and what can’t they do? In this talk, we will explore the current state of the art in AI applications, discuss the challenges of AI adoption in the enterprise, and talk about the risks and future directions of AI development.

This paper introduces a new quantitative toolkit for (reverse) stress testing of the Banking Book -- as required by EBA IRRBB regulations and the Basel Framework. Our toolkit combines classic yield curve modelling and valuation tools with classical Machine Learning clustering techniques to systematically identify blind spots in a bank’s balance sheet. I illustrate the model’s use by applying it to realistic balance sheets of two hypothetical banks and draw implications for risk management and policy making.

In this session, we'll dive into the transformative potential of Large Language Models (LLMs) in finance, moving beyond basic chatbot applications. The focus is on actionable workflows and strategies for leveraging LLMs to automate complex processes, enhance decision-making, and develop data-driven financial solutions. Attendees will gain insights into implementing LLMs to streamline operations, reduce manual effort, and boost efficiency in real-world financial settings.

• News data sourcing and processing.
• Prompt-based investment signals.
• LLM-based factor reduction to identify successful investment themes.
• Comparison to standard investment factors.

In the present work, we introduce and compare state-of-the-art algorithms, that are now classified under the name of machine learning, to price Asian and look-back products with early-termination features. These include randomized feed-forward neural networks, randomized recurrent neural networks, and a novel method based on signatures of the underlying price process. Additionally, we explore potential applications on callable certificates. Furthermore, we present an innovative approach for calculating sensitivities, specifically Delta and Gamma, leveraging Chebyshev interpolation techniques.

Graph Neural Networks (GNNs) have recently shown significant promise in modeling complex relational data structures. In the realm of finance, where assets are often interconnected, capturing these relationships offers potential for improving time series prediction accuracy and explainability. This talk explores the application of GNNs to financial time series forecasting by representing assets and their interdependencies as dynamic graphs. We introduce novel architectures that integrate temporal dynamics with graph-based learning to model both the temporal evolution and the relational structure of financial data.