Exploiting Comparable Corpora with TER and TERp

In this paper we present an extension of a successful simple and effective method for extracting parallel sentences from comparable corpora and we apply it to an Arabic/English NIST system. We experiment with a new TERp filter, along with WER and TER filters. We also report a comparison of our approach with that of (Munteanu and Marcu, 2005) using exactly the same corpora and show performance gain by using much lesser data. Our approach employs an SMT system built from small amounts of parallel texts to translate the source side of the nonparallel corpus. The target side texts are used, along with other corpora, in the language model of this SMT system. We then use information retrieval techniques and simple filters to create parallel data from a comparable news corpora. We evaluate the quality of the extracted data by showing that it significantly improves the performance of an SMT systems.