My PhD focuses on combining temporal logic with causality to improve the verification of AI systems. This will allow us to express interventional (“what will happen if…?”) and counterfactual (“what would have happened if…?”) queries as logical formulae, allowing us to reason about the effect of switching system configurations (e.g., policies in reinforcement learning) on correctness and safety.
I chose the CDT because of its cohort-based approach and diverse (both technical and non-technical) training opportunities. These allow me to engage with people working in many different fields within safe and trusted AI, enhancing my overall PhD experience.
–
Masters Qualification: MEng Computing (with Artificial Intelligence and Machine Learning), Imperial College London
LinkedIn: https://www.linkedin.com/in/jess-lally/
