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

TL;DR

This paper explores the use of structured and web-based language models for reranking machine translation hypotheses, achieving up to 1.6% BLEU score improvements on standard evaluation tasks.

Contribution

It introduces linguistically motivated structured language models based on Constraint Dependency Grammar parses and combines them with web-based N-gram models for improved reranking.

Findings

Structured language models improve translation quality.

Web-based N-gram models enhance reranking effectiveness.

Combination of models yields up to 1.6% BLEU score increase.

Abstract

In this paper, we investigate the use of linguistically motivated and computationally efficient structured language models for reranking N-best hypotheses in a statistical machine translation system. These language models, developed from Constraint Dependency Grammar parses, tightly integrate knowledge of words, morphological and lexical features, and syntactic dependency constraints. Two structured language models are applied for N-best rescoring, one is an almost-parsing language model, 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 language models 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 language models increases the BLEU score up to 1.6% absolutely on blind test sets.