An introduction to statistical methods in machine translation

Marcin Junczys-Dowmunt

Research output: Contribution to journalArticlepeer-review

Abstract

The intention of this article is to provide a concise introduction to the basic mathematical concepts of statistical translation models as they were introduced by Brown et al. (1993) in their groundbreaking work The Mathematics of Statistical Machine Translation: Parameter Estimation. We concentrate on a simplified description of the first two translation models known as IBM Model 1 and 2. It is one major aim of this work to serve as tutoring material for students of computational linguistics, mathematics or computer science and therefore a lot of comments, additional examples and step-by-step explanations are given, augmenting the original formula by Brown et al. (1993). For both discussed models the calculations for a small parallel corpus are described in detail.
Translated title of the contributionAn introduction to statistical methods in machine translation
Original languageOther
Pages (from-to)44-66
Number of pages23
JournalInvestigationes Linguisticae
Volume16
Publication statusPublished - 2008

Fingerprint

Dive into the research topics of 'An introduction to statistical methods in machine translation'. Together they form a unique fingerprint.

Cite this