The objective of this project for Bachelor or Master students is to lay the foundation for a deep learning based decision support system utilizing explainable artificial intelligence methods like prototype-based approaches. So far, prototype-based methods have mostly been used on data sets with clearly distinguishable objects like birds and cars. However, there are images in advertising data sets - such as those produced by Hussain et al. (2017) - that are more diverse and represent more abstract concepts, making classification and explanation efforts both more difficult and more insightful. In this project, students have the opportunity to cooperate closely with the chair team, pursue individual solutions, and apply explainable machine learning techniques.

Learn more about this student project and how to apply:

ProtoXAI - Can AI tell advertisements apart and how?

Contact Person: Luca Gemballa