Machine learning models can detect patterns in large datasets, but understanding why certain observations stand out as unusual or belong to particular groups remains challenging. While many methods identify outliers or clusters of data, they often provide little insight into the reasons behind these patterns.

During her PhD studies, Pernille Matthews developed methods that leverage the internal structure of well-known machine learning algorithms to gain deeper insight into data. By leveraging this structure, the research makes it possible to identify coherent groups of explanations and better understand why certain observations appear as outliers.

The PhD study was completed at the Department of Computer Science, Faculty of Natural Sciences, Aarhus University.

This summary was prepared by the PhD student.

Time: Wednesday 29^th April at 10:00
Place: Building 5510, room 104, Lille Auditorium, Department of Computer Science (INCUBA), Aarhus University, Åbogade 34, 8200 Aarhus N.
Title of PhD thesis: Structure-Guided Explainability: From Regional to Counterfactual Explanations
Contact information: Pernille Matthews, matthews@cs.au.dk, +45 31 31 97 45
Members of the assessment committee:
Professor Panagiotis Papapetrou, Department of Computer and Systems Sciences, Stockholm University, Sweden
Associate Professor Riccardo Guidotti, Department of Computer Science, University of Pisa, Italy
Professor Sophia Yakoubov (chair), Department of Computer Science, Aarhus University, Denmark
Main supervisor: Professor Ira assent, Department of Computer Science, Aarhus University, Denmark
Co-supervisor: Professor Arthur Zimek, Department of Mathematics and Computer Science, University of Southern Denmark, Denmark
Language: The PhD dissertation will be defended in English

The defence is public.
The PhD thesis is available for reading at the Graduate School of Natural Sciences/GSNS, Ny Munkegade 120, building 1521, 8000 Aarhus C