|Talks|

From Hubs to Hypergraphs: Nonparametric Inference for Network Data with the MDL Principle

Visiting speaker
Virtual
Past Talk
Alec Kirkley
Assistant Professor at the University of Hong Kong
Wed, May 8, 2024
1:00 PM UTC
Wed, May 8, 2024
1:00 PM UTC
In-person
4 Thomas More St
London E1W 1YW, UK
The Roux Institute
Room
100 Fore Street
Portland, ME 04101
Network Science Institute
2nd floor
Network Science Institute
11th floor
177 Huntington Ave
Boston, MA 02115
Room
58 St Katharine's Way
London E1W 1LP, UK

Talk recording

Networks pose novel challenges for inference and learning due to their discrete, high-dimensional nature. This inherent complexity necessitates the development of statistically principled unsupervised learning objectives that steer clear of ad hoc heuristics to distinguish meaningful structure from noise in real networks. In this talk I will discuss how to develop principled unsupervised learning methods that parsimoniously summarize structural and dynamical regularities in network data. These methods are unified under the Minimum Description Length principle from information theory, which readily permits fully nonparametric inference while explicitly highlighting particular regularities of interest in discrete datasets. I will discuss the motivation for this family of methods as well as a general procedure for applying this framework to problems in network inference. I will then cover a few examples of recent work in this area where I looked at identifying hub nodes in networks and extracting meaningful cohesive hypergraph structures from temporal datasets using the MDL principle.

About the speaker
Alec Kirkley is an Assistant Professor at the University of Hong Kong (HKU), jointly appointed in the Institute of Data Science and Department of Urban Planning and Design. He is a physicist interested in the theory of complex networks as well as their applications to urban and social systems. His research in network theory aims to develop statistically principled unsupervised learning methods for noisy network data and improve the efficiency and interpretability of network model fitting and evaluation. His research draws on ideas from a range of disciplines including statistical physics, information theory, Bayesian inference, scientific computing, machine learning, urban science, and geography, and he is always interested in forming new interdisciplinary collaborations within these areas.
Share this page:
May 08, 2024