From One Point to A Manifold: Knowledge Graph Embedding For Precise Link Prediction

Knowledge graph embedding aims at offering a numerical knowledge paradigm by transforming the entities and relations into vector space. However, existing methods could not characterize the graph in a fine degree to make a precise prediction. There are two: being an ill-posed algebraic system and applying an overstrict form. As precise prediction is critical, we propose an manifold-based principle () which could be treated as a well-posed system that expands the position of golden triples from one point in models to a manifold in ours. Extensive experiments show that the models achieve substantial improvements against the state-of-the-art especially for the precise prediction task, and yet maintain high.