About me

Hello world! My name is Harry Anthony and I am a DPhil (PhD) student at the University of Oxford under the supervision of Professor Konstantinos Kamnitsas. As part of the IBME group, my research focuses on improving the reliability of deep neural networks in the field of medical imaging.

The focus of my DPhil research is on the field of out-of-distribution (OOD) detection. When training a neural network, we have a set of training images and corresponding labels which we call the training data. We use this to train a neural network on a task of interest, such as classifying diseases from an x-ray scan. Once the model is trained, it can be applied to images without labels during inference. Most of these Images will be from the same distribution as the training data, known as in-distribution, however the model may encounter inputs which differ significantly from the training data, known as out-of-distribution. Neural networks cannot be expected to give sensible predictions on OOD inputs, so we want to detect them to prevent erroneous predictions being used. This is a significant issue for AI in medical image analysis, as wrong predictions on OOD inputs could have serious implications for decisions made downstream.

Through my research, I aim to develop novel methods for deep learning algorithms to detect OOD inputs and improve the overall performance of medical imaging systems. I am passionate about applying my expertise to the field of medical imaging and am excited about the potential impact of my work on improving patient care and advancing the field.

Prior to my doctoral studies, I obtained a first class master’s degree in physics from Imperial College London. My master’s project was titled ‘A search for axions at the LHCb experiment’, where I used deep-learning methods to search through billions of data points for signals of axions (a candidate for dark matter) in the decay of B mesons. Previously, I also achieved 4 A* in A-level (for maths, further maths, physics, chemistry) and 11 A* at GCSE.