Visual Analytics for Explainable Graph-Based Machine Learning

Lead Research Organisation: Swansea University
Department Name: College of Science

Abstract

Theme 3 - Novel Mathematical, Physical and Computer Science

A number of techniques in machine learning now use graphs as an underlying structure for learning rather than simpler structures (such as grids). In these approaches, higher level features in the attributes associated with the nodes and edges are learned on the graph. Although attention has been paid to explainable and interpretable machine learning in general, the problem of graph-based approaches remains open. In this thesis, we will explore the development of novel machine learning techniques on graphs, and create visual analytics systems to explain and interpret such approaches. We will work with domain experts to integrate the developed technology into their workflows.

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/S023992/1 01/04/2019 30/09/2027
2265708 Studentship EP/S023992/1 01/10/2019 30/09/2023 Sophie Francesca Sadler