Discriminative Syntactic Reranking for Statistical Machine Translation
Simon Carter and Christof Monz
In: AMTA 2010, 2-3 June, 2010, Colorado, USA.
This paper describes a method that success- fully exploits simple syntactic features for n-best translation candidate reranking using perceptrons. Our approach uses discrimi- native language modelling to rerank the n- best translations generated by a statistical ma- chine translation system. The performance is evaluated for Arabic-to-English translation using NIST’s MT-Eval benchmarks. Whilst parse trees do not consistently help, we show how features extracted from a simple Part-of- Speech annotation layer outperform two com- petitive baselines, leading to significant BLEU improvements on three different test sets.