Bio
Dom Danks was a final-year PhD student under the supervision of Christopher Yau. He began his doctoral studies at The Alan Turing Institute and the University of Birmingham in September 2019. His background lies in the fields of physics and machine learning, having studied for an MSci in Theoretical Physics at the University of Birmingham and an MSc in Computational Statistics and Machine Learning at UCL. He is particularly excited by the prospect of improving healthcare outcomes using modern statistical techniques. His doctoral research therefore focuses on developing statistical and machine learning methodologies which have applications within biomedical research.
Research interests
During Dom's PhD, they developed a number of machine learning approaches to certain biomedical problems, in particular pseudotemporal analysis and time-to-event modelling. In the area of pseudotemporal analysis, they introduced a model, BasisDeVAE (ICML 2021), which performs simultaneous learning of pseudotemporal profiles from cross-sectional data and assignment of those pseudotemporal profiles to interpretable clusters. In the area of survival analysis, they developed neural network based approaches (e.g. DeSurv, AISTATS 2022) capable of flexibly modelling survival data directly without the restrictions or assumptions inherent to many current models and without the need to necessarily work with the hazard function.
Selected publications and papers
Dominic Danks & Christopher Yau, BasisDeVAE: Interpretable Simultaneous Dimensionality Reduction and Feature-Level Clustering with Derivative-Based Variational Autoencoders, Proceedings of the 38th International Conference on Machine Learning, PMLR 139:2410-2420 (ICML 2021)
Dominic Danks & Christopher Yau, Derivative-based Neural Modelling of Cumulative Distribution Functions for Survival Analysis, The 25th International Conference on Artificial Intelligence and Statistics (AISTATS 2022)
Fabian Falck, Christopher Williams, Dominic Danks, George Deligiannidis, Christopher Yau, Christopher C. Holmes, Arnaud Doucet & Matthew Willetts, A Multi- Resolution Framework for U-Nets with Applications to Hierarchical VAEs, The Thirty-Sixth Annual Conference on Neural Information Processing Systems (NeurIPS 2022)
Vasilis Stavrinides, Georgios Papageorgiou, Dominic Danks, Francesco Giganti, Nora Pashayan, Bruce Trock, Alex Freeman, Yipeng Hu, Hayley Whitaker, Clare Allen, Alex Kirkham, Shonit Punwani, Geoffrey Sonn, Dean Barratt, Mark Emberton & Caroline M. Moore, Mapping PSA density to outcome of MRI-based active surveillance for prostate cancer through joint longitudinal-survival models, Prostate Cancer and Prostatic Diseases (2021)