Abstract / Description of output
System combination exploits differences between machine translation systems to form a combined translation from several system outputs. Core to this process are features that reward n-gram matches between a candidate combination and each system output. Systems differ in performance at the n-gram level despite similar overall scores. We therefore advocate a new feature formulation: for each system and each small n, a feature counts n-gram matches between the system and candidate. We show post-evaluation improvement of 6.67 BLEU over the best system on NIST MT09 Arabic-English test data. Compared to a baseline system combination scheme from WMT 2009, we show improvement in the range of 1 BLEU point.
| Original language | English |
|---|---|
| Publication status | Published - 2010 |
| Event | 9th Biennial Conference of the Association for Machine Translation in the Americas, AMTA 2010 - Denver, CO, United States Duration: 31 Oct 2010 → 4 Nov 2010 |
Conference
| Conference | 9th Biennial Conference of the Association for Machine Translation in the Americas, AMTA 2010 |
|---|---|
| Country/Territory | United States |
| City | Denver, CO |
| Period | 31/10/10 → 4/11/10 |