Mathias Müller

Minimum Bayes Risk decoding has recently gained momentum in #NLProc. However, a generic implementation for @huggingface Transformers was missing. Check out our new 𝗺𝗯𝗿 🔥repo, which is designed to work with any model and metric on the Hub: github.com/ZurichNLP/mbr

We're looking for interns interested in doing research on MT / speech translation in Aachen: jobs.apple.com/en-us/details/…

Check out our paper on segmentation for sign languages! We are one step closer to being able to provide basic NLP tools for sign languages. Great work by Amit and Zifan!

@kchonyc People at the agency should have read this: arxiv.org/abs/1902.10186 😉

noooooo ... from <Reflection paper on the use of Artificial Intelligence (AI) in 6 the medicinal product lifecycle> ema.europa.eu/en/documents/s…

It has been shown that MBR decoding significantly outperforms MAP decoding, but its high cost makes it impractical for most applications. Can we harvest the quality gains of MBR at train time w/o sacrificing inference-time efficiency? TLDR: yes! arxiv.org/abs/2309.10966…

🌟 Just published "SignBank+: Multilingual Sign Language Translation Dataset" (w/ @Jiang_Zifan). Just $500 in dirty dataset cleanup with ChatGPT, leads to a 20-30 BLEU score boost—no changes to the models required. Check out the paper! arxiv.org/abs/2309.11566

@ZJAyres I would read such a book!

@MattNiessner So essentially what you're saying is: we need to advocate for more stable and better-paid jobs in academia.

PhD graduates in AI mostly take boring jobs at big tech companies due to short-term monetary incentives. While understandable to some degree, it's also quite sad to see so many great researchers 'disappear' and give up their talent - join or do your own startup instead!

@sarahookr AYA sounds like a great initiative! Does it include sign languages, or are there plans to include them?

Are you fluent in German, Russian, Ukrainian, Dutch, Romanian, or Hungarian? As of now, these languages are not represented in AYa. Join us to collaborate on this research project. AYA is the largest open science initiative to-date opensourcing multilingual data and models. 🌎

@shortstein I would have liked a third option: arXiv and random conference both tell me nothing about trustworthyness

I'm hiring a post-doctoral researcher to work on multitask/multilingual/multimodal NLP at the University of Zurich. Fully funded until July 2025. Apply by September 15! jobs.uzh.ch/offene-stellen…

Can highly recommend the EASIER Autumn School in Hamburg! Basic idea: teaching sign language to NLP people, and NLP to sign language people. You can apply for fiancial support. sign-lang.uni-hamburg.de/easier/easier-…

Two #ACL2023 Outstanding Paper Awards in our group! Congratulations to @bricksdont, @Jiang_Zifan, @amitmoryossef, @arios272, @ebling_sarah, @RicoSennrich, and all co-authors! aclanthology.org/2023.acl-short… aclanthology.org/2023.acl-long.…

And heartfelt congratulations to my fine co-authors @Jiang_Zifan, @amitmoryossef, @arios272, @ebling_sarah

🎉 To learn more about this work catch Amit *today* at #ACL2023, poster session, 15:00-16:30 UTC, Frontenac Ballroom and Queen's Quay! virtual2023.aclweb.org/paper_P2033.ht…

🥳

This shared task is made possible by @ebling_sarah @lefterav Cristina España-Bonet @DShterionov @amitmoryossef @Jiang_Zifan Oscar Koller @ncihancamgoz Roman Grundkiewicz @DavyVL Annelies Braffort Richard Bowden Anne Goehring @arios272 @malihealikhani Mert Inan

We are excited to see what you can build this year. wmt-slt.com/home

Want to work on a truly unsolved NLP problem, something that ChatGPT most definitely cannot do? Participate in the the Second WMT shared task on sign language translation! 1/3