Event Details

DSNE Seminar with Tom Pinder, Lancaster University

This talk will focus on building Gaussian process models for (hyper)graph data with applications to politics and air quality. Modelling data as a network enables us to express the potentially complex dependency structure that exists within our data. Our job is to build regression models whose domain is the network’s vertex set. I will discuss how this can be done with Gaussian processes for both supervised and unsupervised learning problems with applications to street-level modelling of Nitrogen Dioxide levels in London and inferring the party affiliation of politicians given their legislative actions.

Join Microsoft Teams Meeting

Learn more about Teams | Meeting options

By using this service you agree to LU T&Cs - Use the Help Button for tips

Help | Legal