Statistical phrase-based translation

DMCA

Statistical phrase-based translation (2003)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Franz Josef Och , Daniel Marcu

Citations:

944 - 11 self

BibTeX

@INPROCEEDINGS{Och03statisticalphrase-based,
    author = {Franz Josef Och and Daniel Marcu},
    title = {Statistical phrase-based translation},
    booktitle = {},
    year = {2003},
    pages = {127--133}
}

OpenURL

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. 1

Keyphrases

statistical phrase-based translation phrase translation high-accuracy wordlevel alignment model heuristic learning word-based alignment phrase-based model empirical result strong impact examined language pair large number phrase-based translation model word-based model simple mean new phrase-based translation model lexical weighting

DMCA