Project: TransBooster

Title

TransBooster: boosting the performance of existing MT by complex sentence reduction

Period

01/10/2003 - 01/10/2006

Sponsor

EnterPrise Ireland

Supervisors

Prof. Josef van Genabith

Dr. Andy Way

Researchers

Bart Mellebeek

Karolina Owczarzak

Abstract

The goal of this project is to improve the quality of current Machine Translation output, not by proposing redevelopment from scratch, but by building on the strengths of existing MT engines while trying to correct their common shortcomings. The motivation for the project lies in the fact that many existing wide-coverage translation systems handle simple or short sentences better than linguistically complex ones. Part of the reason for this is that linguistic analysis is costly to develop and that shallow parses are often preferred to deep ones because of robustness reasons. This, however, makes room for a module that can be used in combination with any of the existing MT systems, independent of their nature, to produce better output translations.
Our goal is to develop a chunking algorithm that, based on a high-quality linguistic analysis (possibly involving PCFG and finite-state techniques), has the potential of improving current state-of-the-art MT output. Our approach will be system-independent and will be implemented with efficiency and wide coverage in mind.
Among the likely outputs is a web-based demonstration system, where users will be able to submit text and retrieve translations.

MT engines

Systran

Logomedia

SDL International

Promt