Halvtidsseminarium - Georgios Panayiotou: "Fair and Scalable Multilayer Network Community Detection"

Välkommen till ett halvtidsseminarium presenterat av Georgios Panayiotou. Seminariet hålls på engelska.

Abstract:

Network data are often large and inherently complex, capturing multiple types of entities and relationships across multiple layers, and can encode structural biases. Addressing these challenges requires algorithms that are not only scalable, but also explicitly designed to account for fairness in their outputs.

In this seminar, we focus on the development of fairness-aware methods for community detection in multilayer networks. We specifically discuss challenges in studying large-scale multilayer networks and the construction of a realistic population-scale benchmark network, and explore fairness-aware community detection algorithms that effectively manage the trade-off between partition quality and fairness outcomes, while maintaining scalability.