My PhD project is about improving the robustness and transparency of black box machine learning (ML) models. We are planning to do that by making ML models respect imparted and discovered causal knowledge to then use this as the basis of argumentative explanations. Explainable AI (XAI) is getting a lot of attention because it allows people to understand models and trust their behaviour. That is my aim, to make ML models more robust and trustworthy through causality and argumentative explanations that open these black boxes.
I chose to do a PhD within the CDT because one of the directors sold it to me very well, and I don’t regret it! I wanted to explore safe and trusted AI from an academic perspective after doing work on the subject, but in industry. The breadth of knowledge of different areas of computing makes the CDT a great environment to think about problems and tackle them from different angles with creative interdisciplinary solutions.
Undergraduate Qualification: BSc Economics, University of Rome “Tor Vergata”
Masters Qualification: MSc Applied Statistics, Birkbeck College London
Work Experience: Finance professional with 6 years’ experience in Credit Risk analytics