Fundamentals of AI (EIT CDT)

University of Oxford

Fundamentals of AI (EIT CDT) is a research-based DPhil course focused on foundational AI, machine learning, and computational statistics. Students will help shape the future of AI and Machine Learning with a view to real-world impact.The Ellison Institute of Technology (EIT) Centre for Doctoral Training (CDT) in Fundamentals of AI is dedicated to advancing foundational research in artificial intelligence and machine learning, focusing on theoretical underpinnings and methodological innovation. The CDT's aim is to develop AI technologies with the potential to drive transformative impact across key global challenges aligned with the missions of Ellison Institute of Technology.The course will provide you with training in both cutting-edge AI research methodologies and the development of business and transferable skills. You will work with leading academics at the University of Oxford and will have the opportunity to work closely with project teams at EIT with access to both university and EIT facilities. You will undertake a significant, challenging and original research project, leading to the award of a DPhil.While there can be many definitions of the fundamentals of artificial intelligence (FoAI), within the FoAI CDT, it is defined in three areas that allows a modern, inclusive and diverse interpretation of FoAI.Theory and Foundations: Researchers in this area focus on the foundational mathematical, statistical, and computational principles that underpin AI. This includes research in topics such as learning theory, optimisation, stochastic analysis, complexity theory and formal methods. The aim is to create formal frameworks for the analysis of AI algorithms and systems in order to gain insight into properties, understand behaviours and to develop improved algorithms that could have widespread general use in the field.Applied Fundamentals: At the FoAI CDT, researchers maybe interested in particular applications of AI relevant to EIT’s Humane Themes and Scientific Programmes. Researchers in this area will examine how scientific challenges and the properties of real-world data can guide the reformulation of existing AI algorithms or the design of new algorithms entirely. Topics in this area include physical and process modelling, how to handle missing data, multimodal data integration, decision support, etc.Fundamentals of AI Systems and Engineering: In recent years, there has been an unprecedented emergence of large and complex AI systems, such as Large Language Models. Researchers in this area are interested in the formal frameworks for characterising the design and development of such systems and using these to further understand the properties and behaviours of such systems. They may also be interested in the security, scalability and physical resource requirements of such systems.