Learning three-way affinity embeddings for knowledge base completion

摘要: Knowledge bases are an extremely important database for knowledge management, which is very useful question answering, query expansion and other related tasks. However, it often suffers from incompleteness. In this paper, we propose a Three-Way Affinity Embeddings model (TWAE) to map both the entity relationship into two vectors consider any of them direct interaction, then predict possible truth additional facts. The basic idea that confidence predicted fact determined by three-way affinities in triplet using latent representation each item. Experiments show our performs excellent.