Vilém Zouhar

zouharvi

AI & ML interests

NLP-oriented human-computer interaction (trust, confidence) ||| Non-mainstream machine translation (quality estimation) ||| NLP evaluation (model-human communication, metrics)

Recent Activity

upvoted a paper about 21 hours ago

We Can't Understand AI Using our Existing Vocabulary

upvoted a paper about 21 hours ago

Early-Exit and Instant Confidence Translation Quality Estimation

commented on a paper about 22 hours ago

Early-Exit and Instant Confidence Translation Quality Estimation

View all activity

Organizations

zouharvi's activity

upvoted 2 papers about 21 hours ago

We Can't Understand AI Using our Existing Vocabulary

Paper • 2502.07586 • Published 15 days ago • 10

Early-Exit and Instant Confidence Translation Quality Estimation

Paper • 2502.14429 • Published 6 days ago • 2

upvoted a collection 1 day ago

PreCOMET

Collection

COMET-like models for MT evaluation that predict some scores given only the source segment. https://github.com/zouharvi/subset2evaluate • 8 items • Updated about 23 hours ago • 2

upvoted a paper 7 days ago

How to Select Datapoints for Efficient Human Evaluation of NLG Models?

Paper • 2501.18251 • Published 27 days ago • 2