Insights into Unsupervised Learning Models with Explainable AI

Speaker: Grégoire Montavon (BIFOLD/TU)

Unsupervised learning has gained prominence as a way to learn models that are not too task-specific or dependent on user annotations. These models are particularly needed in the context of exploratory data analysis, where human knowledge about the task is often limited or difficult to access. In this talk, I will present how unsupervised models can be enriched with Explainable AI to gain deeper insight into their predictions. The benefits of these Explainable AI-enhanced unsupervised models will be illustrated with use cases in biomedicine and digital humanities.

MDC-BIMSB, Hannoversche Str. 28, 10115 Berlin, Conference room, Ground floor

Please register here for online participation.