Contributed Talk: Ontology-based Interpretable Machine Learning with Learnable Anchors