FrameBase: Representing N-Ary Relations Using Semantic Frames

摘要: Large-scale knowledge graphs such as those in the Linked Data cloud are typically represented subject-predicate-object triples. However, many facts about world involve more than two entities. While n-ary relations can be converted to triples a number of ways, unfortunately, structurally different choices made sources significantly impede our ability connect them. They also make it impossible query data concisely and without prior each individual source. We present FrameBase, wide-coverage knowledge-base schema that uses linguistic frames seamlessly represent from other bases, at levels granularity connected by logical entailment. It opens possibilities draw on natural language processing techniques for querying mining.