Affinity-Based Reinforcement Learning: A New Paradigm for Agent

You may follow the disputation online. Link for registration as an online spectator at the bottom of this page.

Charl Maree of the Faculty of Engineering and Science at the University of Agder has submitted the thesis entitled «Affinity-Based Reinforcement Learning: A New Paradigm for Agent Interpretability» and will defend the thesis for the PhD-degree 10 March 2023.

He has followed the PhD programme of the Faculty of Engineering and Science.

Summary of the thesis by Charl Maree:

The evolution of financial services is driven by continuously growing customer bases and an ever-increasing demand for personalisation; the need for automation has proliferated artificial intelligence (AI) into a ubiquitous tool in finance. While AI algorithms are typically opaque, financial service providers are subject to a fiduciary duty that dictates transparency in all systems that affect people’s lives. Recent efforts therefore pursue explainability and interpretability through advanced methods in explainable AI.

This study shows how to predict customers’ personality traits from their financial transactions and align investment advice accordingly. For instance, highly conscientious individuals might prefer reduced risk, whereas individuals that are highly open to new experiences might prefer more novel and interesting asset types. Affinity-based reinforcement learning (ab-RL) is a new paradigm that changes the way machines learn optimal strategies (which are decomposed according to different personality traits). This novel method imprints desired and undesired behaviours into reinforcement learning agents, making their strategies inherently interpretable. Ab-RL mathematically guarantees convergence to an optimal personal strategy.

This work has been published in several prestigious journals, such as Digital Finance and AI and Ethics (both Springer Nature journals), which demonstrates both novelty and value in real-world applications. The new paradigm is generic and future work will extend it to imprint desired behaviours that are both situation-dependent and time-dependent. Potential future applications include optimal control of windfarms, personalised medical treatment, personal education plans, etc.

What to do as an online audience member:

The disputation is open to the public, but to follow the trial lecture and the public defence online, transmitted via the Zoom conferencing app, you have to register as an audience member on this link: (coming soon)

A Zoom-link will be returned to you. (Here are introductions for how to use Zoom: support.zoom.us if you cannot join by clicking on the link.)

We ask online audience members to join the virtual trial lecture at 09:50 at the earliest and the public defense at 11:50 at the earliest. After these times, you can leave and rejoin the meeting at any time. Further, we ask online audience members to turn off their microphone and camera and keep them turned off throughout the event. You do this at the bottom left of the image when in Zoom. We recommend you use ‘Speaker view’. You select that at the top right corner of the video window when in Zoom.

Opponent ex auditorio:

The chair invites members of the public to pose questions ex auditorio in the introduction to the public defense. Deadline is during the break between the two opponents. The person asking questions should have read the thesis. For online audience the Contact Persons e-mail are available in the chat function during the Public Defense, and questions ex auditorio can be submitted to Emma Elizabeth Horneman.