Improving Translation Quality by Discarding Most of the Phrasetable

It is possible to reduce the bulk of phrasetables for Statistical Machine Translation using a technique based on the significance testing of phrase pair co-occurrence in the parallel corpus. The savings can be quite substantial (up to 90%) and cause no reduction in BLEU score. In some cases, an improvement in BLEU is obtained at the same time although the effect is less pronounced if state-of-the-art phrasetable smoothing is employed.

• Type: Conference
• Year: 2007
• Where: EMNLP
• Authors: Howard Johnson, Joel D. Martin, George F. Foster, Roland Kuhn
• Keywords: EMNLP 2007, Natural Language Processing, Phrase Pair Co-occurrence, State-of-the-art Phrasetable Smoothing, Statistical Machine Translation