Toward an Architecture for Never-Ending Language Learning.pdfVIP

Toward an Architecture for Never-Ending Language Learning Andrew Carlson , Justin Betteridge , Bryan Kisiel , Burr Settles , Estevam R. Hruschka Jr. , and Tom M. Mitchell School of Computer Science, Carnegie Mellon University, Pittsburgh, PA 15213, USA ˜ ˜ Federal University of Sao Carlos, Sao Carlos, SP, Brazil Abstract supervised learning methods, an ensemble of varied knowl- edge extraction methods, and a flexible knowledge base rep- We consider here the problem of building a never-ending lan- resentation that allows the integration of the outputs of those guage learner; that is, an intelligent computer agent that runs methods. We also discuss design principles for implement- forever and that each day must (1) extract, or read, informa- tion from the web to populate a growing structured knowl- ing this approach. edge base, and (2) learn to perform this task better than on We then describe a prototype implementation of our ap- the previous day. In particular, we propose an approach and proach, called Never-Ending Language Learner (NELL). At a set of design principles for such an agent, describe a par- present, NELL acquires two types of knowledge: (1) knowl- tial implementation