I focus on developing new methods to provide provable formal guarantees on neural network behaviour. Specifically, I have interests in geometric robustness and Lipschitz-based local and global robustness for neural network verification.
While a standalone PhD is a great opportunity to investigate a very specific area of research, being part of the CDT has provided an additional roundedness to my training that it would be hard to otherwise obtain. The CDT activities not only round out your academic training, they also afford the chance to work with stakeholders, external academics, and public engagement centres – all this serves to provide you with a great foundation for your career, whether that be in academia or industry. It is for these reasons I applied to do my PhD at Imperial as part of the STAI CDT.
Masters Qualification: MPhys – Physics – University of Bath