Design of a System That Understands Informal Specifications.

Abstract

This paper investigates an artificial intelligence to combining the advantages of both formal and natural languages. The long-term goal is a system which could take as input an English definition of a module, and generate an equivalent formal specification. In addition, the system should generate an English paraphrase of its understanding of the input, so that the user may easily check the system's understanding. The remainder of the paper describes the design decisions made in implementing a prototype to understand English texts defining data structures. Section 2 enumerates some of the reasons we feel are most important for using natural language. Section 3 defines the target specification language and the motivation in selecting it. Section 4 relates our experience in using a parser for texts defining data structures. Section 5 deals with semantic issues such as interpreting spatial metaphors and selecting precise translations of vague English terms. Related work and our conclusions are presented in sections 6 and 7.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 1983
Accession Number
ADA131479

Entities

People

  • Daniel L. Chester
  • Ralph M. Weischedel

Organizations

  • University of Delaware

Tags

Communities of Interest

  • Energy and Power Technologies
  • Human Systems

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Computational Linguistics
  • Computer Languages
  • Computer Programming
  • Computer Programs
  • Computer Science
  • Computers
  • Environment
  • Formal Languages
  • Information Processing
  • Information Science
  • Language
  • Linguistics
  • Natural Language Processing
  • Natural Languages
  • Software Development
  • Specifications

Fields of Study

  • Engineering

Readers

  • Artificial Intelligence
  • Business Analytics
  • Computational Linguistics

Technology Areas

  • AI & ML
  • AI & ML - Machine Translation