Modern Artificial Intelligence and Machine Learning (ML) techniques have revolutionized the field of computer science, finding applications in diverse domains ranging from healthcare to finance. As these technologies evolve, ensuring their robust, ethical, and efficient deployment demands a nuanced understanding of underlying incentives at play in the models themselves as well as the contexts they are deployed in. In this PhD, I will investigate a series of problems in computer science and related fields, applying methods from traditional fields like game theory to examine the incentive structures underpinning these issues. I will then employ modern optimization techniques from ML and Deep Learning (DL), making use of the insight into these incentive schemes. In particular, I intend to examine applications of the Transformer architecture for specific data needs, such as handling sequential time-based data or network graphs.
While a large body of current research uses deep learning models as an end-to-end solution, I believe that combining ML/DL models with traditional techniques such as game theory could produce hybrid systems that leverage the strengths of both approaches. Through this combination of modern and traditional methods, I aim not only to enhance existing solutions, but to develop novel strategies for addressing complex decision-making and optimization problems.
I am generally interested in AI and Machine Learning, in particular modern Deep Learning methods such as the Transformer architecture. I am also interested in more theory-focussed techniques such as Game Theory and traditional optimisation.
MSc Computer Science and Artificial Intelligence, University of Nottingham
Research Assistant (Using AI in Logistics Optimisation), University of Bath
Dr Jie Zhang
Prof Mike Tipping