Charting Paragraph Topics

Lexikos Corporation

Significance of our Design MODELER formally models topics of English paragraphs and what was said or asked about them.  Its discourse models and your lexicons map each parsed phrase into Java objects in CONTEXT cast as Conceptual Graphs. This design lets client software download, as annotated graphs, the intended contextual meaning of each grammar form in paragraphs semantically charted by our interpreter. Such charts enable man-machine communication, as they are intuitive to people, yet also structured to flexibly control NLP application programs.
Technical Goals Create an easy-to-use desktop interpreter for English that converts any paragraph into semantic metadata that models the things and situations it discusses. Simulate English understanding by transcribing each input sentence into a scoped association of the topics its phrases reference. Work Plan Use a scripting language and open source Java tools to define executable case frames covering standard English. Debug them using our parser; help clients create extension lexicons for APIs and concepts that are application-specific.		Anticipated Applications Man-machine Q&A on modeled Topics Embedded equipment service manuals Analysis of English message traffic Situation-based software systems Multilingual Semantic Lexicons Improved information retrieval Enhanced voice processing Model-based training software Multimedia content indexing UI module for Intelligent Agents To Beta Test, Please Contact Dan Corwin