Putting Pieces Together: Understanding Patent Abstracts

Lebowitz, Michael

One aspect of the development of RESEARCHER (Lebowitz 83a), an intelligent information system that reads, remembers and learns from patent abstracts, is the use of strongly semantic-based text understanding methods. We show in this paper how patent abstracts can be processed by using only very simple syntactic rules to identify "pieces" of the ultimate representation and then "putting the pieces together." An example of RESEARCHER processing a sample abstract is shown.


More About This Work

Academic Units
Computer Science
Department of Computer Science, Columbia University
Columbia University Computer Science Technical Reports, CUCS-098-84
Published Here
October 26, 2011