Ontology Pre-training for Poison Prediction

Martin Glauer, Fabian Neuhaus, Till Mossakowski +1

Integrating human knowledge into neural networks has the potential to improve their robustness and interpretability. We have developed a novel approach to integrate knowledge from…