KBLRN: End-to-End Learning of Knowledge Base Representations with Latent, Relational, and Numerical Features

We present KBLRN, a framework for end-to-end learning of knowledge base from latent, relational, and numerical features. KBLRN feature types with a novel combination of neural representation and probabilistic product of experts models. To the best of our, KBLRN is the first approach that learns representations of knowledge by integrating latent, relational, and numerical features. We show that of KBLRN outperform existing methods on a range of knowledge base tasks. We contribute a novel data sets enriching commonly used base completion benchmarks with numerical features. The data sets are under a permissive BSD-3 license. We also investigate the impact features have on the KB completion performance of KBLRN.