Dr. Andy Way
School of Computing
Dublin City University

Example-Based Machine Translation:
An Overview of Recent Research Projects

Dr. Andy Way will talk about research he is conducting in the area of Example-Based Machine Translation in conjunction with two Computational Linguists at Dublin City University. Both researchers are developing Example-Based MT systems, i.e. new sentences are translated not by rules but by referring to a set of previous example translation pairs in the system's database. On rare occasions, such as translating updates of software manuals, for instance, exact matches of the source language sentences can be found, in which case their translations are grabbed 'for free'.

More likely, however, we will have to construct a translation from a number of close (or 'fuzzy') matches in the system's memory. One of the co-workers has produced a "linguistics-lite" approach to EBMT, while the other has developed a "linguistics-rich" model. First, we will provide an overview of EBMT in general.

We will go on to sketch the architectures of these systems, outline recent research results, and report on their contributions to the field. We will also provide a comparison of the two techniques, and report on avenues for further research