Fichier de travail (INPUT) : ./DUMP-TEXT/2-16.txt
Encodage utilisé (INPUT) : utf-8
Forme recherchée : translation|traduction|机器
_________________________________________________________________________________________________
Ligne n°99 : ... The Google Neural Machine Translation system 'surpasses' the results of- Ligne n°100 : all other machine-translation solutions currently available, with GNMT
Ligne n°101 : now being used for Chinese-to-English translations. ...
Ligne n°100 : ... all other machine-translation solutions currently available, with GNMT- Ligne n°101 : now being used for Chinese-to-English translations.
Ligne n°102 : Corinne Reichert ...- Ligne n°113 : Google has announced a neural machine translation (NMT) system that it
Ligne n°114 : says will reduce translation errors across its Google Translate service ...
Ligne n°113 : ... Google has announced a neural machine translation (NMT) system that it- Ligne n°114 : says will reduce translation errors across its Google Translate service
Ligne n°115 : by between 55 percent and 85 percent, calling the achievement by its ...
Ligne n°121 : ... systems, but had previously found it "challenging" to improve machine- Ligne n°122 : translation until now.
Ligne n°137 : ... (GNMT), which utilizes state-of-the-art training techniques to achieve- Ligne n°138 : the largest improvements to date for machine translation quality."
- Ligne n°140 : Unlike the currently used phrase-based machine translation (PBMT)
Ligne n°141 : system -- which translates words and phrases independently within a ...
Ligne n°141 : ... system -- which translates words and phrases independently within a- Ligne n°142 : sentence, and is notorious for its mistranslations -- neural machine
Ligne n°143 : translation considers the entire sentence as one unit to be translated. ...
Ligne n°142 : ... sentence, and is notorious for its mistranslations -- neural machine- Ligne n°143 : translation considers the entire sentence as one unit to be translated.
- Ligne n°145 : Researchers have been working on improving neural machine translation
Ligne n°146 : over the past few years, with Google's scientists finding a way to make ...
Ligne n°151 : ... Wu, Zhifeng Chen, and Mohammad Norouzi et al, said the system is "state- Ligne n°152 : of the art" for English-to-French and English-to-German translations in
Ligne n°153 : particular, reducing errors by 60 percent on average. ...- Ligne n°162 : "To accelerate the final translation speed, we employ low-precision
Ligne n°163 : arithmetic during inference computations. To improve handling of rare ...
Ligne n°166 : ... balance between the flexibility of 'character'-delimited models and the- Ligne n°167 : efficiency of 'word'-delimited models, naturally handles translation of
Ligne n°168 : rare words, and ultimately improves the overall accuracy of the ...
Ligne n°172 : ... English:- Ligne n°173 : google-neural-machine-translation-system-2.png (Image: Google)
Ligne n°183 : ... "Using human-rated side-by-side comparison as a metric, the GNMT system- Ligne n°184 : produces translations that are vastly improved compared to the previous
Ligne n°185 : phrase-based production system." ...
Ligne n°187 : ... Google has launched GNMT in production across Google Translate on web- Ligne n°188 : and mobile for the Chinese-to-English translation pair, accounting for
Ligne n°189 : around 18 million translations daily. It will roll out GNMT with more ...
Ligne n°188 : ... and mobile for the Chinese-to-English translation pair, accounting for- Ligne n°189 : around 18 million translations daily. It will roll out GNMT with more
Ligne n°190 : languages over the coming months. ...
Ligne n°190 : ... languages over the coming months.- Ligne n°191 : google-neural-machine-translation-system.png
Ligne n°192 : google-neural-machine-translation-system.png (Image: Google) ...
Ligne n°191 : ... google-neural-machine-translation-system.png- Ligne n°192 : google-neural-machine-translation-system.png (Image: Google)
Ligne n°195 : ... vocabularies and the challenge of morphologically rich languages for- Ligne n°196 : translation quality and inference speed; that a combination of model
Ligne n°197 : and data parallelism can be used to efficiently train state-of-the-art ...
Ligne n°198 : ... sequence-to-sequence NMT models in roughly a week; that model- Ligne n°199 : quantization drastically accelerates translation inference, allowing
Ligne n°200 : the use of these large models in a deployed production environment; and ...