State-of-the-art black box AI algorithms have revolutionised our ability to construct algorithms for use with complex data sets. These approaches have been of tremendous utility across a number of sectors but have presented some limitations in the biomedical sciences where algorithms that are able to produce “explainable” outputs are highly desirable or critical for some tasks. One particular problem is the need to integrate many types of biomedical data sources (e.g. medical images, genetics and electronic health records) in order to combine these to produce outputs but it remains an open question as to how such integration should be done to produce explainable outputs. This project is associated with an UKRI Turing AI Acceleration Fellowship awarded to the primary supervisor (Professor Christopher Yau). It will seek to develop explainable probabilistic machine learning methods for multimodal data integration for biomedical applications. We will consider a number of theoretical and practical issues with data integration and explainable AI and develop robust statistically sound solutions to these research problems. The candidate will gain a strong foundation in the development of novel machine learning methodologies using concepts such as Bayesian modelling, decision theory and deep learning. The supervision team has a wealth of experience in methodological development and applied data science providing the student with a fully immersive training and learning environment. The candidate should have a strong interest in developing a career in artificial intelligence research and possess a strong quantitative background obtained from a first degree in mathematics, physics, engineering or computer science. The candidate will be required to train and acquire skills in advanced statistical programming in a modern environment (e.g. PyTorch, TensorFlow, JAX). It is expected that the candidate will develop research outputs that will be publishable in the internationally leading machine learning outlets (including NeurIPS, ICML, AISTATS, UAI). UK/Overseas research internships with both academic and industrial partners will be embedded within the PhD study period to give the candidate exposure to the international AI environment. Informal enquiries should be directed to Professor Christopher Yau ([Email Address Removed])
http://cwcyau.github.io
Entry Requirements:
Applications are invited from UK nationals only. Applicants must have obtained, or be about to obtain, at least an upper second class honours degree (or equivalent) in a relevant subject.
To be considered for this project you MUST submit a formal online application form. For information on how to apply for this project, please visit the Faculty of Biology, Medicine and Health Doctoral Academy website (https://www.bmh.manchester.ac.uk/study/research/apply/)