How necessary is parallel data for improving machine translation quality for low-resource languages? Google researchers scale model with ‘pragmatic approach.’
To be fair, this is the first-ever multilingual model to win the international machine translation contest. But major tech companies have been exploring multilingual models for years.
The most-cited neural machine translation research papers show how NMT came to dominate the field — and how academia and industry’s interests have evolved.
Google study shows human-paraphrased reference translations and new evaluation metric, BLEUP, produce better translations. Findings to be integrated into consumer-facing products.
A group of AI researchers takes the perennial human vs. machine debate to the next level by claiming their system outperforms professional translators...in news, on adequacy (read on for more caveats).
New research suggests biases of QE datasets lead to guesstimates on machine translation quality. Researchers propose modified dataset, debut it at WMT 2020.
As demand for media loc strains suppliers, there is a case for automation. Research finds NMT might be used for subtitles without special training.
Though Amazon announced its open-source, sequence-to-sequence machine learning framework Sockeye would be available to all last July 2017, its researchers waited until December 2017 to file papers into Arxiv so they can test their creation against WMT2017’s best.
In early September, Copenhagen will become the center of the machine translation universe as hundreds of researchers gather in the Danish capital to learn whose system performs best.
Slator Weekly: Join over 15,500 subscribers and get the latest language industry intelligence every Friday
Tool Box: Join 10,000 subscribers for your monthly linguist technology update.
Your information will not be shared with third parties. No Spam.
This will close in 0 seconds
