Parsing the Voyager Domain Using Pearl

Abstract

This paper describes a natural language parsing algorithm for unrestricted text which uses a probability-based scoring function to select the "best" parse of a sentence according to a given grammar. The parser, Pearl, is a time-asynchronous bottom-up chart parser with Earley-type lop-down prediction which pursues the highest-scoring theory in the chart, where the score of a theory represents the extent to which the context of the sentence predicts that interpretation. This parser differs from previous attempts at stochastic parsers in that it uses a richer form of conditional probabilities based on context to predict likelihood. Pearl also provides a framework for incorporating the results of previous work in part-of-speech assignment, unknown word models, and other probabilistic models of linguistic features in one parsing tool, interleaving these techniques instead of using the traditional pipeline architecture. In tests performed on the Voyager direction-finding domain, Pearl has been successful at resolving part-of-speech ambiguity, determining categories for unknown words, and selecting correct parses first using a very loosely fitting covering grammar.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 1991
Accession Number
ADA460709

Entities

People

  • David M. Magerman
  • Mitchell P. Marcus

Organizations

  • University of Pennsylvania

Tags

Communities of Interest

  • Air Platforms
  • Energy and Power Technologies
  • Human Systems

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Ambiguity
  • Diptera
  • Direction Finding
  • Drosophila
  • Grammars
  • Information Operations
  • Information Science
  • Language
  • Models
  • Natural Language Processing
  • Natural Language Understanding
  • Natural Languages
  • Probabilistic Models
  • Probability
  • Statistics

Fields of Study

  • Computer science

Readers

  • Computational Linguistics
  • Geodesy