CUNI Submissions in WMT18
By Tom Kocmi, Roman Sudarikov, Ondrej Bojar. 2018
10 citations recorded by the source corpus.
We participated in the WMT 2018 shared news translation task in three language pairs: English-Estonian, English-Finnish, and English-Czech. Our main focus was the low-resource language pair of Estonian and English for which we utilized Finnish parallel data in a simple method. We first train a “parent model” for the high-resource language pair followed by adaptation on the related low-resource language pair. This approach brings a substantial performance boost over the baseline system trained only on Estonian-English parallel data. Our systems are based on the Transformer architecture. For the English to Czech translation, we have evaluated our last year models of hybrid phrase-based approach and neural machine translation mainly for comparison purposes.
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
- Ondrej Bojar
- Philipp Koehn
- Markus Freitag
- Felix Stahlberg
- Tom Kocmi
- Yvette Graham
- Christian Federmann
- Barry Haddow
Citations in the corpus
- Towards Continual Learning for Multilingual Machine Translation via Vocabulary Substitution (2021, methodology): “Of these baselines, only mBART and (Kocmi et al., 2018) use sacreBLEU which inhibits proper comparison with the rest of the models.”
- Findings of the 2018 Conference on Machine Translation (WMT18) (2018, methodology): “2.4.4 CUNI-KOCMI (Kocmi et al., 2018) The CUNI-KOCMI submission focuses on the low-resource language neural machine translation (NMT).”
- CUNI Submission for Low-Resource Languages in WMT News 2019 (2019, methodology): “Based on our previous experiments (Kocmi et al., 2018) we set the maximum number of sub-words in a sentence to 100, which drops less than 0.1 percent of training sentences.”
- The University of Maryland’s Kazakh-English Neural Machine Translation System at WMT19 (2019, methodology): “Transfer learning has been found effective in submissions to WMT in previous years: Kocmi et al. (2018) reported improvements of +2 .”