An Architecture for Content Analysis of Documents and its Use in Information and Knowledge Management Tasks
A modular, reconfigurable architecture (SCOOP) for document content analysis: natural-language components (lexical, morphological, syntactic, semantic) that can be recombined for different document types and depths of analysis. Shown at work on two tasks: turning technical manuals into databases for on-line help, and building topic-rich abstractions of news stories.
From the natural-language research program at Apple’s Advanced Technology Group in Cupertino.