Neural Concept Formation in Knowledge Graphs

AAAI 2021

Abstract: In this work, we investigate how to learn novel concepts in Knowledge Graphs (KGs) in a principled way, and how to effectively exploit them to produce more accurate neural link prediction models. Specifically, we show how concept mem- bership relationships learned via unsupervised clustering of entities can be reified and effectively used to augment a KG. In our experiments we show that neural link predictors trained on these augmented KGs, or in a joint Expectation-Maximization iterative scheme, can generalize better and produce more accu- rate predictions for infrequent relationships while delivering meaningful concept representations. Authors: Agnieszka Dobrowolska, Antonio Vergari and Pasquale Minervini (University College London, University of California, Los Angeles)