A Tandem Semantic Interpreter for Incremental Parse Selection

Abstract

This report describes TINSEL (Tandem INterpreter and SELection) software, a clause-local semantic interpreter and selection component for use with a natural language parser. As the parser incrementally regularizes candidate phrasal constituents into syntactic operator-operand forms, it submits each form to a user-defined selection process for the application of sublanguage constraints on co-occurrence patterns of semantic word classes. Only those analyses that pass selection are added to the parser's working space for involvement in further search, which can often reduce total processing time considerably. TINSEL treats selection as the enforcement of predicate domain constraints during semantic interpretation, the composing of a data structure that represents the meaning of a natural language phrase in some suitable formalism. The interpreter runs in tandem with the parser, transforming each semantically valid operator-operand form submitted to into a predicate-argument form based on a set of declarative predicate models provided by the user. The parser's regularization component then composes partial interpretations of parent nodes from the interpretations of their children. TINSEL also operates top-down and so can be applied postparse to the final output of the parser rather than interleaved with it.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 28, 1990
Accession Number
ADA227743

Entities

People

  • Kenneth Wauchope

Organizations

  • United States Naval Research Laboratory

Tags

Communities of Interest

  • Ground and Sea Platforms
  • Human Systems
  • Weapons Technologies

DTIC Thesaurus Topics

  • Aircrafts
  • Antisubmarine Aircraft
  • Artificial Intelligence
  • Artificial Intelligence Software
  • Automated Text Summarization
  • Computational Science
  • Computer Languages
  • Detection
  • Grammars
  • Helicopters
  • Information Systems
  • Language
  • Linguistics
  • Lisp Programming Language
  • Natural Language Processing
  • Natural Language Understanding
  • Natural Languages

Fields of Study

  • Computer science

Readers

  • Computational Linguistics

Technology Areas

  • Space