Natural-Language Question-Answering Systems
David L. Waltz · Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign) · 1975
In his excellent book, Libraries of the Future, 1 J.C.R. Licklider paints an elaborate picture of what libraries may become by the year 2000. He sees libraries as being accessible through and augmented by digital computer programs and evolving into "procognitive systems," or general aids to think- ing. Many library documents, as well as much text, such as that of computer- typeset books have already been made computer-readable. But how far have we come in devising programs that do this reading automatically? And how close are we to systems that can understand users' questions, comments and commands? These are questions I will attempt to answer in this paper. The systems I will describe all deal primarily with facts rather than documents. I assume that facts are inherently more difficult to deal with, and that documents are a special case of fact.