Skip to main content

Concordia researcher applies machine learning to financial risk management

PhD candidate Alexandre Carbonneau says AI could lead to improved stability among banks and insurance companies
December 16, 2019
Alexandre Carbonneau: “I highly encourage students interested in machine learning to dive right in.”
Alexandre Carbonneau: “I highly encourage students interested in machine learning to dive right in.”

Financial institutions such as banks and insurance companies face multiple risk factors, whether operational, credit or market-related.

Managing them is key to maintaining their overall stability — but even with the most sophisticated financial instruments, those working in the field don’t always get it right.

Concordia researcher Alexandre Carbonneau says machine learning might be able to improve upon existing methods. The PhD candidate in financial mathematics is studying the use of reinforcement learning methods to optimize risk management decisions made by banks and insurers.

I hope that my approach will have practical applications in the financial sector

How does this specific image (top left) relate to your research at Concordia?

Alexandre Carbonneau: My main area of research is in the application of machine learning methods in the fields of finance and actuarial science. More precisely, I’m interested in developing methods to help banks and insurers manage the different risks they face.

What is the hoped-for result of your project?

AC: With my project, I aim to show that machine learning methods can have a significant impact on the effectiveness of risk management techniques and can thus improve upon existing methods. Furthermore, I hope that my approach will have practical applications in the financial sector.

What impact could you see it having on people’s lives?

AC: Banks and insurance companies are subject to multiple sources of risks that are growing in complexity. Managing all these risks simultaneously is extremely important for the financial stability of these institutions.

What are some of the major challenges you face in your research?

AC: My main challenge ironically appeals to my greatest source of motivation. It consists of keeping up-to-date with recent developments in multiple areas of research.

Within the field of machine learning, I’m mostly interested in deep reinforcement learning (RL) methods — training an artificial agent to interact in an environment in order to optimize some objectives.

For example, these methods have been used to defeat a human world champion at the game of Go and to learn how to play the classic Atari 2600 games.

At first glance, these problems seem to be completely unrelated to my main research topic, but the underlying problem is similar. In the game of Go, the agent has to pick sequences of movements to optimize the probability of winning the game. In my research, I’m using RL methods to optimize the sequence of decisions made by banks and insurers to manage the different risks they are exposed to.

I need to know the literature related to risk management in my particular field of application as well as be up-to-date with some areas of RL as I will possibly be able to use them in my research.

What first inspired you to study this subject?

AC: My interest in risk management started during my bachelor’s degree in actuarial science. I realized that it could have an important impact on the financial stability of insurers.

I discovered during my master’s degree in financial mathematics that financial risk management was related to several areas of mathematics in which I was particularly interested. These included dynamic optimization, computational finance and applied probability and statistics.

What advice would you give interested students to get involved in this line of research?

AC: I would suggest students develop the ability to learn new technical skills that are outside their specific field of application. I believe this is especially true in this day and age where developments in technology can happen extremely quickly.

Also, I would highly encourage students interested in machine learning that don’t believe they have the required skills to dive right in. In its essence, this field is closely related to multiple subjects many are already familiar with, such as applied statistics, optimization and algorithms.

What do you like best about being at Concordia?

AC: Concordia has a large group of faculty members in the domain of actuarial science and financial mathematics. I also really appreciate the guidance my supervisor Frédéric Godin provides me on a regular basis with my research.

Are there any partners, agencies or other funding/support attached to your research?

AC: I’d like to thank the Fonds de recherche du Québec - Nature et technologies, the Institut des sciences mathématiques, the Fonds pour l’éducation et la saine gouvernance de l’Autorité des marchés financiers as well as the internal funding of Concordia for the financial support during my PhD.

Additionally, I’m a member of NSERC’s Fin-ML CREATE program, which is offered by multiple Canadian universities to train students and researchers in the field of machine learning in quantitative finance and financial business analytics.

Find out more about
Concordia’s Department of Mathematics and Statistics.



Back to top

© Concordia University