Neuro-Symbolic Ontology-Mediated Query Answering
Recently, low-dimensional vector space representations of Knowledge Graphs (KGs) have been applied to find answers to logical queries over incomplete KGs. However, the current methods only focus on inductive reasoning, i.e. answering such queries by predicting facts based on patterns learned from the data, and lack the ability of deductive reasoning, the task of computing logical entailments using expert domain knowledge. To address this shortcoming, we investigate how existing embedding models for query answering over incomplete KGs can be adapted to incorporate domain knowledge in the form of ontologies. We propose two novel datasets, based on LUBM and NELL KGs, as well as various training strategies to integrate domain knowledge into prominent representatives of embedding models for query answering. Our strategies involve (1) different ontology-driven data augmentation techniques and (2) adaptation of the loss function using query-rewriting methods. The achieved improvements in the settings that require both inductive and deductive reasoning, are from 20% to 50% in HITS@3.
PDF Abstract
NELL
