CUBBITT translation
translation quality comparable to human professionals
Charles University Block-Backtranslation-Improved Transformer Translation
Try the translation service at lindat.cz/cubbitt.
Read the paper in Nature Communications.
Popularization in Czech:
- Úspěch českého překladače - Události ČT + ČT24, September 21, 2020
- Překladač z Matfyzu dohání v kvalitě běžné překladatele - UKforum.cz, October 5, 2020
- Jak se vám líbí překlad (od stroje)? - Vesmír, October 5, 2020
- Trénink překladače na Mechanickém pomeranči u oficiálních dokumentů narazil, směje se jeho autor - ČT24, January 23, 2022
Popularization in English:
- Czech-developed computer translator CUBBITT “reaching quality of human professionals”, Radio Prague International, September 22, 2020
- MT Weekly blog, September 11, 2020
- Automatic translation of Czech interview at ČT24, January 23, 2022
Background on (Neural) Machine Translation:
Training data
How to cite
If you make use of CUBBITT, please cite the following paper:
@article{ biblio:PoToTransformingmachine2020,
journal = {Nature Communications},
title = {Transforming machine translation: a deep learning system reaches news translation quality comparable to human professionals},
author = {Martin Popel and Marketa Tomkova and Jakub Tomek and {\L}ukasz Kaiser and Jakob Uszkoreit and Ond{\v{r}}ej Bojar and Zden{\v{e}}k {\v{Z}}abokrtsk{\'{y}}},
year = {2020},
volume = {11},
number = {4381},
pages = {1--15},
issn = {2041-1723},
doi={10.1038/s41467-020-18073-9},
url={https://www.nature.com/articles/s41467-020-18073-9},
}