
DW
Daniel Whitenack and Chris Benson
· 30 Minutes
PodcastPractical AI
SOTA machine translation at Unbabel
Listen
José and Ricardo joined Daniel at EMNLP 2022 to discuss state-of-the-art machine translation, the WMT shared tasks, and quality estimation. Among other things, they talk about Unbabel’s innovations in quality estimation including COMET, a neural framework for training multilingual machine translation (MT) evaluation models.
Featuring:
- Ricardo Rei – X
- José Souza – X
- Daniel Whitenack – Website, GitHub, X
Show Notes:
- Unbabel
- COMET
- The WMT workshop/ conference
- EMNLP
Upcoming Events:
- Register for upcoming webinars here!
Original source
This story was published by Practical AI and written by Daniel Whitenack and Chris Benson. SyncAI.news shows a preview; the complete article is on the publisher's site.
Read the full story on share.transistor.fm

![[AINews] Inkling 2 Small: 17A-594B trained in 68 hours on 12T tokens, beating Muse Spark](/media/images/2026/11/0dfbab5a1d1fb32a.webp)
