Two types of adaptive MT environments

Authors:
Sergei Nirenburg;David Farwell;Robert Frederking;Yorick Wilks
Affiliations:
Carnegie Mellon University, Pittsburgh, PA;New mexico State University, Las Cruces, NM;Carnegie Mellon University, Pittsburgh, PA;University of Sheffield, UK
Venue:
COLING '94 Proceedings of the 15th conference on Computational linguistics - Volume 1
Year:
1994

Citing 4
Cited 2

A statistical approach to machine translation

Computational Linguistics
Machine Translation: A Knowledge-Based Approach

Machine Translation: A Knowledge-Based Approach
TAUM-AVIATION: its technical features and some experimental results

Computational Linguistics - Special issues on machine translation
Automatic evaluation of computer generated text: a progress report on the TextEval project

HLT '94 Proceedings of the workshop on Human Language Technology

An open distributed architecture for reuse and integration of heterogeneous NLP components

ANLC '97 Proceedings of the fifth conference on Applied natural language processing
Natural Language Processing as a Foundation of the Semantic Web

Foundations and Trends in Web Science

Quantified Score

Hi-index	0.00

Visualization

Abstract

A number of proposal have come up in recent years for hybridization of MT. Current MT projects --- both "pure" and hybrid, both predominantly technology-oriented and scientific (including those currently funded by NSF) are single-engine projects, capable of one particular type of source text analysis, one particular method of finding target language correspondences for source language elements and one prescribed method of generating the target language text. While such projects can be quite useful, we believe that it is time to make the next step in the design of machine translation systems and to move toward adaptive, multiple-engine systems. We describe the architecture of an adaptive multi-engine MT system which uses each of the engines under the circumstances which are most favorable for its success.