Statistical phrase-based translation

Abstract

We propose a new phrase-based translation model and decoding algorithm that enables us to evaluate and compare several, previously proposed phrase-based translation models. Within our framework, we carry out a large number of experiments to understand better and explain why phrase-based models outperform word-based models. Our empirical results, which hold for all examined language pairs, suggest that the highest levels of performance can be obtained through relatively simple means: heuristic learning of phrase translations from word-based alignments and lexical weighting of phrase translations. Surprisingly, learning phrases longer than three words and learning phrases from high-accuracy wordlevel alignment models does not have a strong impact on performance. Learning only syntactically motivated phrases degrades the performance of our systems.

A Statistical Machine Translation Model with Forest-to-Tree Algorithm for Semantic Parsing

“…We assume the alignment to be n-to-1, where each word is linked to at most one MRL production. Basically, a reduction-based λ-SCFG grammar rule and a phrase alignment (Koehn et al, 2003) can be extracted from an λ-hybrid tree where logical variables are explicitly bound by λ-operators. And these grammar rules are extracted in a bottomup manner, starting with MRL productions at the leaves of the λ-hybrid tree.…”

WMT 2016 Multimodal Translation System Description based on Bidirectional Recurrent Neural Networks with Double-Embeddings

“…Similarly, machine translation approaches based on neural networks (Sutskever et al, 2014; are competing with standard phrase-based systems (Koehn et al, 2003). Neural machine translation uses an encoder-decoder structure .…”

Historical Documents Modernization

“…These models rely on a log-linear combination of different models : namely, phrase-based alignment models, reordering models and language models; among others (Zens et al, 2002;Koehn et al, 2003). However, in the last few years, neural machine translation (Sutskever et al, 2014;Bahdanau et al, 2015) has had a great impact.…”