Discovering semantic relations using singular value decomposition based techniques

Understanding language in any form requires understanding connections among words, concepts, phrases and thoughts. Many of the problems faced today in artificial intelligence depend in some manner on understanding this network of relationships that represents the facts that each of us knows about the world and how words relate to one another. Researchers in many areas have looked for ways to discover such relationships automatically, but current automated methods lack the spontaneity that is present in human learning and can miss many basic relationships that are rarely stated directly in corpora. Some of the most essential common sense knowledge is learned at an early age before people can write at all, much less write the text traditionally found in corpora. If this knowledge is found in text it tends to be implicit, rather than explicitly stated. The paradox of this vital unspoken information has led researchers in these fields to manually con-struct corpora of common sense using either experts or volunteers from the Internet. Such projects are major undertakings, using vast amounts of time and resources, but the results are invaluable to the field. To aid in the development of such resources, I present a technique called “blending ” that infers new knowledge from the alignment and inference over disparate resources. The process is very similar to how people make analogies for reasoning