Difference between revisions of "Uighur and Turkish/Paper"

From Apertium
Jump to navigation Jump to search
Line 16: Line 16:
 
* [https://ieeexplore.ieee.org/abstract/document/5666183/ Chinese-uyghur statistical machine translation: The initial explorations]
 
* [https://ieeexplore.ieee.org/abstract/document/5666183/ Chinese-uyghur statistical machine translation: The initial explorations]
 
* [http://en.cnki.com.cn/Article_en/CJFDTOTAL-JYRJ201203060.htm DESIGNING RELATED TRANSFORMATION AND MATCHING RULES FOR CHINESE-UYGHUR MACHINE TRANSLATION]
 
* [http://en.cnki.com.cn/Article_en/CJFDTOTAL-JYRJ201203060.htm DESIGNING RELATED TRANSFORMATION AND MATCHING RULES FOR CHINESE-UYGHUR MACHINE TRANSLATION]
  +
* [http://www.apsipa.org/proceedings/2017/CONTENTS/papers2017/14DecThursday/TP-04/TP-04.1.pdf Memory-augmented Chinese-Uyghur Neural Machine Translation] - This looks interesting
 
 
   
 
==== Japanese-Uyghur ====
 
==== Japanese-Uyghur ====

Revision as of 10:52, 4 July 2018

Info related to paper on uig-tur RBMT.

TO DO

  • Fill out latex template
  • Evaluate system on corpora

To write

  • Intro
  • Evals

Images and Diagrams

Related Work

Uyghur-specific

Chinese-Uyghur

Japanese-Uyghur

Non-english Research

Japanese: Muhsut Muhtar seems to have worked a lot on japanese-uighur translation in japanese universities.

Turkish:

Other

Turkic

Fun

Evaluations

Since there seem to have been a number of prototypes, mostly for Chinese-Uyghur, it might make sense to compare these values with any evaluation those systems may have, and also the Turkic Apertium pairs crh-tur, kaz-tur, kaz-tat and gag-tur.

  • Adalı &co seem to have worked a good bit on Uyghur analysis and disambiguation, so that would also be a good thing to compare.

WER

BLEU

Examples