Multilinear Embeddings Of Knowledge Graphs
Knowledge graphs, which store facts in form of entities and their relationships, have found important applications in areas such as Web search and question answering. Recently, machine learning for knowledge graphs has received considerable attention as it can be used to discover previously unknown facts, to categorize and disambiguate entities, and to support automated knowledge base construction. However, knowledge graphs pose also unique challenges for machine learning, due to their size and their complex, relational structure.
In this talk, I will present a family of latent feature models for knowledge graphs (and graph-structured data in general) that allow to create complex statistical models of graphs with millions of entities and billions of known facts. I will discuss how the proposed approach exploits relational information through its multilinear latent variable structure and how it can be applied to a wide range of tasks including link prediction, entity resolution, and link-based clustering. In addition, I will show briefly how the model can be used to answer complex probabilistic queries on knowledge graphs and how it can be combined with observable variable models to further increase its predictive performance and scalability.
Maximilian Nickel is a postdoctoral fellow at MIT where he is with the McGovern Institute for Brain Research and the Laboratory for Computational and Statistical Learning. In 2013, he received his PhD with summa cum laude from the Ludwig Maximilian University Munich. From 2010 to 2013 he worked as a research assistant at Siemens Corporate Technology. His research interests center around machine learning from relational and graph-structured knowledge representations and its applications in artificial intelligence, automated knowledge base construction, and question answering.