Dr Evangelia (Lina) Kyrimi

Dr Kyrimi's research interests lie in Bayesian modelling and decision support under uncertainty in healthcare. Her research focuses on (1) developing clinical decision support systems using a combination of knowledge and data and (2) translating causal models into explainable AI (XAI) systems that users can trust.

The latter is a part of a 5-year research fellowship funded by the Royal Academy of Engineering (RAEng). Technological breakthroughs have led to the development of sophisticated healthcare systems, but these will only become widely adopted if patients and healthcare professionals have confidence in their recommendations. Without a solution to the problem of user trust and user acceptance of healthcare technologies generally, the undeniable benefits of these systems will never be realised and efforts to develop accurate health-AI will be in vain. The ‘right to explanation’ and regulations on algorithmic decision-making already exist. Therefore, the ExAIDSS project focuses on translating causal AI models into explainable AI systems that users can trust and adopt in healthcare.

The objectives of this prestigious research fellowship are as follows:

1. Investigate the fundamentals of explanation: Explore fundamental questions that have been neglected, such as what makes an explanation of AI “good”.
2. Develop explanation algorithms that incorporate causality: Develop explanation algorithms that produces meaningful causal explanations for various types of reasoning.
3. Create user-specific explanation outputs: Design an explanation that recognises who is interacting with it and the dynamics of clinical decision making.
4. Create an evaluation protocol: Propose a protocol for evaluating different explanations purposes.
5. Integrate the explanation algorithm and representation into existing healthcare digital platforms. 

(more details are available https://exaidss.com/) 

Dr Kyrimi is also interested in understanding and bridging the existing gap between developing accurate clinical decision support systems and implementing them into practice.