25 April 2008

Congratulations to Karolina Owczarzak who successfully defended her thesis and will be awarded the degree of PhD.

The title of Karolina's thesis is "A Novel Dependency-Based Evaluation Metric for Machine Translation".

She completed her PhD in the National Centre for Language Technology (NCLT), and the School of Computing, DCU under the supervision of Professor Josef van Genabith and Professor Andy Way.

Karolina will soon begin her new post as a guest researcher at the National Institute of Standards and Technology (NIST), Gaithersburg, MD, USA, working on evaluation methods for automatic summaries.

Brief description of Project:

In the development of Machine Translation (MT) systems, automatic evaluation metrics allow developers to conduct frequent, fast, and cost-effective evaluations of their translation models. However, most of the metrics rely on a comparison of word strings, measuring only the surface similarity of the candidate and reference translations. A candidate translation expressing the source meaning accurately and fluently will be given a low score if the lexical and syntactic choices it contains, even though legitimate, are not present in at least one of the references. This thesis presents a method that automatically evaluates the quality of translation based on the labelled dependency structure of the sentence, rather than on its surface form. This lets us compare the translation and the reference on the level of shallow semantic relations, and the score obtained in this way reflects human judgment of translation quality better than most existing metrics.

This project was generously funded by Enterprise Ireland (EI) and Microsoft Ireland.

Back to News Headlines