Reranking Machine Translation Hypotheses with Structured and Web-based Language Models

Abstract: In this paper, we investigate the use of linguistically motivated and computationally efficient structured LLMs for reranking N-best hypotheses in a statistical machine translation system. These LLMs, developed from Constraint Dependency Grammar parses, tightly integrate knowledge of words, morphological and lexical features, and syntactic dependency constraints. Two structured LLMs are applied for N-best rescoring, one is an almost-parsing LLM, and the other utilizes more syntactic features by explicitly modeling syntactic dependencies between words. We also investigate effective and efficient language modeling methods to use N-grams extracted from up to 1 teraword of web documents. We apply all these LLMs for N-best re-ranking on the NIST and DARPA GALE program 2006 and 2007 machine translation evaluation tasks and find that the combination of these LLMs increases the BLEU score up to 1.6% absolutely on blind test sets.