Project: Machine Learning (ML) algorithms are capable of assessing patterns in complex datasets by analysing and drawing inferences. The recent advancements in ML focus on data-driven approaches for modelling and analysing real-world dynamical systems. The glucoregulatory system in our body is such a complex dynamical system which is impaired in Type 1 Diabetes (T1D). To treat T1D, insulin is administered to the body to manage the glucose levels. However, exercise and other daily activities, together with their associated uncertainty, increase the complexity of the system and treatment further. Physiological signals (e.g., heart rate, galvanic skin response) captured by wearables devices provide additional information to help capture and understand these complexities. In this project we use ML methods and information theoretic approaches to explore biomarkers, understand the effects of physiological signals and their relationships related to the glucoregulatory system to improve treatment systems, focusing on a latest advance dataset whose use for the purposes of this project has already been approved by the ANU human research ethics committee.
Timeline:
- Background & Related Work. Familiarize about the glucoregulatory system as a real-world dynamical system and learn about its technical characteristics (uncertainty, disturbances, delays) and related ML/DL applications. As a starting point, see our demonstration system called CAPSML at https://capsml.com/ .Then write a literature survey to explore the literature associated with the application of ML/DL in T1D, to identify existing state-of-the-art benchmarks and algorithms.
- Experiments & Validation. Develop novel supervised/unsupervised ML/DL and Information Theory approaches towards identifying and analyzing target physiological signals captured from wearables and their relationships with the glucoregulatory system. Perform their comparative evaluations or even a validation study.
- Bonus Item. Design and implement visualizations of the developed techniques to be showcased in the CAPSML tool (https://capsml.com/), for clinicians, people with T1D, and researchers to try out the proposed methods.
- Conclusion. Thesis writing, potential publication writing, and code documentation.
Supervisors:
- [Dr. Chirath Hettiarachchi](/people/chirath-hettiarachchi/ (Research Fellow, School of Computing), a computer scientist working on ML and Reinforcement Learning algorithms to develop closed-loop treatment for clinical applications.
- Prof. Chris Nolan (Associate Director Research, ANU School of Medicine and Psychology) is a clinician/scientist/teacher/policy advisor in diabetes aiming to improve the care of people with T1D using a personalised medicine approach.
- Prof. Hanna Suominen (Associate Director of JEI, Neuroinformatics), a computer scientist with 20 years’ of experience at forefront of bringing ML/DL algorithms, document analysis methods, and personalized medicine technologies to bear.
Requirements: To undertake substantial work on this project, students must have a solid background and very good skills in programming and ML/DL. To apply for this opportunity please forward your CV and transcript to Dr. Chirath Hettiarachchi (chirath.hettiarachchi@anu.edu.au). Successful candidates will be invited for a discussion to learn more about the project.