Explainable AI: Finding Correlation and Causation in Knowledge Graphs

Franz is involved in a number of Knowledge Graphs that capture real world events. In each domain we work to understand the correlation and co-occurrence of certain events so, given an event X, we can predict the chance of event Y. We all were taught that correlation does not equal causation but following Judea Pearl’s famous book "The Book of Why" we developed several techniques that add the arrow of time to our data analysis so we can start talking about 'causal' relationships in real world Knowledge Graphs. This method rises above the typical machine learning techniques in that it creates explainable AI that can be understood via a dynamic graph visualizer.


Interested in this talk?

Register for SEMANTiCS conference