I am currently interested in the testing and verification of machine learning systems, and also the use of machine learning to improve testing and verification in general.
Specifically, I have been investigating whether machine learning can be used to automatically generate realistic test cases. By considering so-called adversarial examples as a kind of test input to neural networks, this work has so far cumulated in an arXiv preprint, Generating Realistic Unrestricted Adversarial Inputs using Dual-Objective GAN Training. In this paper, we describe a method which can be used to train a neural network to generate an realistic-looking image which fools a given target classifier.
In ongoing work, I am looking at whether this approach can be generalised to generate tests for other kinds of system. I am also exploring whether the generated realistic unrestricted adversarial inputs can be used to train a more robust classifier.
After completing my undergraduate degree in computer science at the University of Cambridge in 2016, I spent two years teaching computer science at a school in Birmingham through the Teach First Leadership Development programme. I began studying for a PhD under Profs. Daniel Kroening and Tom Melham in October 2018.