Event details from: Business Analytics Seminar: Reza Mohammadi (University of Amsterdam)

Date
5 April 2022
Time
12:00 – 13:00

Abstract

Graph-based machine learning approaches have gained incredible success in the domain of deep learning. More recently, graph neural networks have been adapted to leverage the structure and properties of graph/network. Despite their success, these methods do not capture model uncertainty which is crucial in real-world applications, for instance, for the clinical assessment of patients. In contrast, Bayesian methods provide mathematically grounded frameworks to deal with model uncertainty, which come with expensive computations.

Alzheimer's is associated with brain network dysfunction, therefore graph-based machine learning techniques are vital for disease staging and classification. These powerful tools can be applied to facilitate the early diagnosis of Alzheimer’s disease. In this talk, I present a Bayesian graph-based approach to capture changes in the brain network for early diagnosis of Alzheimer's by using imaging (MRI, fMRI, PET) and non-imaging (demographics and genotypes) features. This approach can potentially improve the accuracy as well as the computation of the graph neural network methods. I will also discuss the computational problems followed by several solutions to cope with these issues.

General information

This seminar will be organised via Zoom. If you are interested in joining this seminar, please send an email to the secretariat of Amsterdam Business School at secbs-abs@uva.nl.