T
Taro Watanabe
Researcher at Google
Publications - 116
Citations - 1995
Taro Watanabe is an academic researcher from Google. The author has contributed to research in topics: Machine translation & Rule-based machine translation. The author has an hindex of 26, co-authored 88 publications receiving 1910 citations. Previous affiliations of Taro Watanabe include Nippon Telegraph and Telephone & National Institute of Information and Communications Technology.
Papers
More filters
Proceedings Article
Online Large-Margin Training for Statistical Machine Translation
TL;DR: Experiments on Arabic-toEnglish translation indicated that a model trained with sparse binary features outperformed a conventional SMT system with a small number of features.
Proceedings ArticleDOI
Reordering constraints for phrase-based statistical machine translation
TL;DR: This work investigates different reordering constraints for phrase-based statistical machine translation, namely the IBM constraints and the ITG constraints and presents efficient dynamic programming algorithms for both constraints.
Proceedings Article
An Unsupervised Model for Joint Phrase Alignment and Extraction
TL;DR: An unsupervised model for joint phrase alignment and extraction using non-parametric Bayesian methods and inversion transduction grammars (ITGs) is presented, which matches the accuracy of traditional two-step word alignment/phrase extraction approach while reducing the phrase table to a fraction of the original size.
Proceedings ArticleDOI
Recurrent Neural Networks for Word Alignment Model
TL;DR: A word alignment model based on a recurrent neural network (RNN), in which an unlimited alignment history is represented by recurrently connected hidden layers, which outperforms the feed-forward neural network-based model as well as the IBM Model 4 under Japanese-English and French-English word alignment tasks.
Proceedings Article
Bilingual Lexicon Extraction from Comparable Corpora Using Label Propagation
TL;DR: A novel method for lexicon extraction that extracts translation pairs from comparable corpora by using graph-based label propagation and achieves improved performance by clustering synonyms into the same translation.