Lecture Series “Robust AI”: Rianne M. Schouten

The next lecture in out lecture series will be given by Rianne M. Schouten (Eindhoven University of Technology) on the topic of “Exceptional Model Mining for Hierarchical Data”.

When & where: June 5, 2:15 pm, online, link: https://uni-bielefeld.zoom-x.de/j/67097106624?pwd=MbEH9jKzpIvYo43dP0rWBQ6sVrqwzv.1

Abstract: People differ, not only in physical appearance but in personality, cultural background, abilities, and interests, as well as in cognitive, emotional and social behavior. In this talk, I present our work on analyzing variation in human behavior using a Local Pattern Mining (LPM) framework called Exceptional Model Mining (EMM). EMM aims to discover subgroups in a population that somehow behave exceptionally. These subgroups are described using an interpretable language of conjunctions of attribute-value pairs. Traditionally, EMM is applied to tabular data, but we observe that real-world data is often hierarchically structured, and measurements cannot be considered independent. For instance, repeated measurements are nested in patients or students are nested in classes. In this talk, I present EMM methodologies for hierarchical data and demonstrate real-world impact with use cases from diabetes care, public health and learning analytics (education).

Rianne M. Schouten is a postdoc at Eindhoven University of Technology. Learn more about her work at: https://research.tue.nl/en/persons/rianne-m-schouten