QGen: generation module for the register restricted InBASE system

  • Authors:
  • Michael V. Boldasov;Elena G. Sokolova

  • Affiliations:
  • Moscow State University, Computational Mathematic and Cybernetic department, Moscow, Russia;Russian Research Institute for Artificial Intelligence, Moscow, Russia

  • Venue:
  • CICLing'03 Proceedings of the 4th international conference on Computational linguistics and intelligent text processing
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper we present our investigations of constructing of a generation module for the InBASE system - a commercially oriented system for understanding of natural language queries to a Data Base. The initial prototype of the module re-generates the user query from the internal OQL representation into a natural language text presented in the form of extended nominal group. We discuss the main principles and methods of the organization of the generation module and peculiarities of the approaches we use for the knowledge representation as well as at planning and realization phases of generation. The initial prototype demonstrates direct transition from the OQL register specific representation to morphologically marked up structured representation of the query text. Directions of the further investigations are also discussed in the article.