Project Category: Algorithmic transparency

While Deep neural networks (DNNs) succeed in exploiting non-linear patterns in very large and high-dimensional datasets, they catastrophically fail without warning under dataset shift, i.e., changes in data distribution.  This PhD project will study various ways to resolve this pitfall by characterising, detecting, and generalising against dataset shift. The research will theoretically unify sparse and inconsistent literature, and empirically validate that theory in the application of genomics, with results to inform ways to maximise reliability of learning systems under datasets shifts.  

The project focuses on developing techniques and methodologies grounded in unsolved challenges in computational biology and multi-modal healthcare data.  The project commenced in October 2021 with the recruitment of the Centre’s first PhD Researcher, Sam MacDonald, who is based at The University of Queensland. Sam is investigating the proposed methodologies on real datasets from different healthcare organisations, and is supervised by Chief Investigator Dr Fred Roosta-Khorasani from the School of Mathematics and Physics (UQ) and Dr Quan Nguyen from the Institute for Molecular Bioscience (IMB) at UQ.

The team collaborated with Max Kelsen, a Brisbane based artificial intelligence and software engineering agency, during 2021 to early 2023. Max Kelsen was acquired by Bain & Company in 2023.

A major bottleneck for enterprises adopting AI is the difficulty in applying and interpreting the correct method for a given problem.  

This project will survey available interpretable methods in AI and communicate best practices in both lay and comprehensive terms and explore new theoretical landscapes to extend and innovate interpretable methods in AI, focusing on both uncertainty (aleatoric and epistemic), and causality. Emphasis will be on probabilistic inference, using graphical models, including both neural networks and more general approaches like neural wirings and directed acyclic graphs.  

The project will investigate the proposed methodologies on real datasets from different healthcare organisations. 

This project commenced in April 2023 with the recruitment of PhD researcher Eslam Zaher, who is based at The University of Queensland. Eslam is supervised by Chief Investigator Dr Fred Roosta-Khorasani, and Affiliate Investigators Dr Quan Nguyen, and Dr Maciej Trzaskowski.

The team collaborated with Max Kelsen, a Brisbane based artificial intelligence and software engineering agency, during early 2023. Max Kelsen was acquired by Bain & Company in 2023.