1. The hierarchical organisation and dynamics of complex networks
- Author
-
McDonald, David
- Subjects
QA75 Electronic computers. Computer science - Abstract
Complex networks offer flexible representations of complex heterogeneous real-world systems. They are often weighted, attributed, directed and/or dynamic. As such, gaining an overall understanding of information flow through these systems remains a challenging problem in the machine learning community. This thesis provides a comprehensive examination of the hierarchy inherent to many complex networks, with the following contributions: • The first algorithm to learn low dimensional non-Euclidean representations of attributed nodes in a weighted complex network. • The first algorithm to learn low dimensional non-Euclidean representations of attributed nodes in a directed complex network. • A framework to explore the multi-scale organization of meso-scopic architectures in signalling networks, allowing for the identification of statistically significant drug-able targets. Through these contributions, the work proposed in this thesis contributes towards a greater understanding of the hierarchy in the organization and dynamics of complex real-world systems.
- Published
- 2021