We're releasing a new multilingual model for NameTag 3: nametag3-multilingual-260521. It achieves state-of-the-art performance on 33 test datasets across 23 languages. Try the demo: lindat.mff.cuni.cz/services/nam... Documentation: ufal.mff.cuni.cz/nametag/3

🎉 Congratulations to our colleague Dan Zeman, who together with Marie-Catherine de Marneffe, @chrmanning.bsky.social & Joakim Nivre has won the ACL 2026 Computational Linguistics High Impact Paper Award for "Universal Dependencies"! 👏 aclanthology.org/2021.cl-2.11

#ACL2026 continues with the main conference (main + findings + demos + industry + SRW), and @ufal.mff.cuni.cz folks will present 8️⃣ papers. Stop by and check with our colleagues. All times in PDT.

Congrats! Stop by to chat with @ivankartac.bsky.social at #ACL2026 in San Diego 🇺🇸 He's presenting 2 papers: A modular neuro-symbolic system pairing small LLMs with a theorem prover for syllogistic reasoning, and BOULDER, a new benchmark showing LLM reasoning drops sharply inside multi-turn dialogue.

Ivan Kartáč@ivankartac.bsky.social · last mo.

Heading to San Diego for #ACL2026, where I’ll be presenting two papers (see 🧵). Stop by to chat about evaluating reasoning embedded in task-oriented dialogue, or how to use small LLMs in modular neuro-symbolic approaches to syllogistic reasoning!

#ACL2026 starts with two days of tutorials and workshops. @ufal.mff.cuni.cz folks will be there and present their work, here's our schedule, so don't miss our presentations 👇

If you are at #ACL2026, don't miss @mariemikulova.bsky.social et al. presenting a semantic-pragmatic annotation layer for the 3M+ token Prague Dependency Treebank aclanthology.org/2026.finding...

Marie Mikulová@mariemikulova.bsky.social · last mo.

Semantic-pragmatic annotation in the Prague Dependency Treebank — three times at #ACL2026! 📍 Jul 3, 13:30–14:15 LAW La Jolla, talk 📍 Jul 3, 14:20–15:30 CODI-CRAC Regatta, poster 📍 Jul 6, 11–12:30 Findings Poster, Grand Hall See you in San Diego! @ufal.mff.cuni.cz aclanthology.org/2026.finding...

Input, more input 🤖⚡ Just like Jonny 5 in Short Circuit, our baby model is reading every single token from its pretraining dataset. So far: 10 trillion tokens, 36 languages + code & math as their own "languages" 📚🌍💻 We’re tracking progress & sharing it openly 👇 (1/2)

Ally Sheedy and Johnny 5 in Short Circuit

ALT: Ally Sheedy and Johnny 5 in Short Circuit

static.klipy.com

📣 TokShop 2026 deadline extended! 🗓️ New submission deadline: Friday, June 26, 2026 (AoE) Research papers (up to 9 pages) and extended abstracts (up to 2 pages) are welcome. Submit: openreview.net/group?id=col... More info: tokenization-workshop.github.io

Lukáš Eigler defended his thesis (co-supervised with David Hurych, @valeoai.bsky.social) 🎉 Congrats! #NLP metric validation needs 🐌💰 human judgment data. Our fix: generate synthetic data for metric validation. ✅ Tested on MT, QA, summarization. To appear #ACL2026 SRW: arxiv.org/abs/2603.09403

LLM as a Meta-Judge: Synthetic Data for NLP Evaluation Metric Validation

Validating evaluation metrics for NLG typically relies on expensive and time-consuming human annotations, which predominantly exist only for English datasets. We propose LLM as a Meta-Judge, a scalabl...

arxiv.org

One of a dozen master's theses defended at our institute this season. Thesis season is always a reminder of the talented students passing through here. 🎓

Jindřich Libovický@jlibovicky.bsky.social · 2mo ago

Is "Je to obrovský problém, tvrdí české ministerstvo obrany." Czech or Slovak? Both. Language ID systems must pick one: and that's just one way they break. Our new benchmark CHALIS exposes how fragile they really are. 📄 arxiv.org/abs/2606.06088 🤗 huggingface.co/datasets/michal-tichy/CHALIS

Announcing First Call for Papers: Second Tokenization Workshop 🔡 📣 ▶️ Non-archival submissions of two types: Research papers (up to 9 pages) ▶️ Extended abstracts (up to 2 pages) Submission deadline June 23, 2026 (AoE) Acceptance notification on July 24, 2026 (AoE) tokenization-workshop.github.io

Great discussions, inspiring talks, and lots of interest around the 🌲Prague Dependency Treebank🌲 at #LREC2026. Hope we helped give PDT some well-deserved visibility there! @ufal.mff.cuni.cz ➡️ lrec.elra.info/lrec2026-mai... ➡️ lrec.elra.info/lrec2026-mai... ➡️ lrec.elra.info/lrec2026-mai...

#LREC2026 is over. It had huge @ufal.mff.cuni.cz presence 👯🙆‍♂️🧑‍🤝‍🧑🙋🧓🧒 of 2️⃣6️⃣ people and 1️⃣4️⃣ main conference papers. If you missed our presentations, you can still checkout our #NLProc and #CL papers 👇

TokShop will be at #COLM2026! 🗓️ October 9th, 2026 📍 San Francisco, USA More details and a call for papers coming soon.

All ready to share information about #OpenEuroLLM with the #LREC2026 crowd. Let's talk data, infra, evals and open multilingual LLM models together! Come to booth #5 at the poster area 1, Elyxir Building. #multingualLLMs #openLLMs #diverseLLMs #safeLLMs

#LREC2026 🌴🇪🇸 continues with day 1 of the main conference. @ufal.mff.cuni.cz folks will present 5 papers: From annotating medieval manuscripts to auditing LLM hallucinations: Prague NLP is everywhere 🚀

🔸 🔶 Petr Sgall was a pioneer who saw the future of NLP long before the age of LLMs 🔶 🔸 As we mark the 100th anniversary of his birth, Eva Hajičová and Jarmila Panevová have shared a brilliant tribute to his life and work. 🔗 Read the full tribute here: www.matfyz.cz/clanky/ste-v...

Sté výročí průkopníka počítačové lingvistiky prof. Petra Sgalla

V květnu tohoto roku si připomínáme 100 let od narození emeritního profesora Univerzity Karlovy a dlouholetého pracovníka Matematicko-fyzikální fakulty UK PhDr. Petra Sgalla, Dr. h. c. mult. (27. květ...

matfyz.cz

#LREC2026 starts in a couple of hours and @ufal.mff.cuni.cz folks are there to present their research. 🤩 Don't miss the panel discussion with @hajicjan.bsky.social on the role of symbolic representations in the era of LLMs. And check out our posters and talks 👇

Exciting news for our team! 🚀 The results for @ACL 2026 are out, and we are thrilled to announce that at least 3 papers from ÚFAL (Institute of Formal and Applied Linguistics) will be presented in San Diego in July! 🔗https://2026.aclweb.org/

Let us introduce one of our partners, the Institute of History of the Czech Academy of Sciences (AV ČR) and their Czech Historical Database project (biblio.hiu.cas.cz) 📺 Watch the video: youtube.com/shorts/wngjX... (turn on the English subtitles in the ... menu). . youtube.com/shorts/wngjX...

LINDAT/CLARIAH CZ - MEDAILONEK AVČR

YouTube video by ÚFAL MFF UK

youtube.com

#EACL2026 is over. Thank for representing @ufal.mff.cuni.cz and congrats on the work you have presented! @abyste.bsky.social, @gianlucavico.bsky.social, @patuchen.bsky.social, @jlibovicky.bsky.social, @zdenekkasner.cz, @tuetschek.bsky.social, @namlh201.bsky.social, @jjon19.bsky.social et al.

On the weekend, #EACL2026 continues with workshops and @ufal.mff.cuni.cz folks present their research 👇 Also, don't miss @tuetschek.bsky.social's keynote talk on How (Not) to Find Errors in LLM Outputs at the LowResMT workshop in the morning.

New preprint: arxiv.org/abs/2603.20133 Does performance on reasoning benchmarks transfer to real-world settings such as task-oriented dialogue? Not necessarily: our new benchmark tests LLMs on problems framed in both standalone and dialogue settings, and shows that dialogue makes reasoning harder.

#EACL2026 in Rabat 🇲🇦 starts tomorrow and @ufal.mff.cuni.cz folks will present their research. Don't miss our presentations 👇

🔸 Save the Date: A Special Centenary Seminar 🔸 Event Details: 📅 Date: Monday, April 13, 2026 ⏰ Time: 14:00 📍 Location: S1, MFF UK, Malostranské náměstí 25 🎁 Talks by Marie Mikulova, Dan Zeman and Jindřich Libovický 🔗 Full program and details: ufal.mff.cuni.cz/events/semin...

V rozhovoru pro aktuálně.cz se dozvíte, jak probíhá vytváření evropského velkého jazykového modelu @openeurollm.bsky.social zpravy.aktualne.cz/zahranici/ev...

Evropané vytváří novou AI. V čem bude lepší než americká a čínská?

Evropská unie vytváří projekt velkého jazykového modelu, jehož koordinátorem je český vědec, počítačový lingvista Jan Hajič z z Matematicko-fyzikální fakulty pražské Karlovy univerzity. V čem bude nov...

zpravy.aktualne.cz

🎉 KORPUS TŘICETILETÝ: Nová publikace oddělení Českého Národního korpusu FF UK u příležitosti 30. výročí ČNK 🎉 🔗 www.nln.cz/knihy/korpus... ÚFAL se na knize podílel nejen editorsky, ale i třemi klíčovými příspěvky. #NLP hashtag#Lingvistika hashtag#UFAL hashtag#PDT30 hashtag#CorpusLinguistics

New work from our team on morphology-aware tokenization evaluation, accepted to EACL 2026 Findings.

Jindřich Libovický@jlibovicky.bsky.social · 6mo ago

We (= mostly @abyste.bsky.social) developed a way to evaluate how morphological a #tokenization is w/o gold segmentation labels. arxiv.org/abs/2601.18536 The key: align subword tokens with morphological features from UniMorph using IBM Model 1. To appear in EACL 2026 Findings.

If you think labeling text spans with LLMs is easy, you probably have not tried it yourself (we have! 🙃). Any method you can think of – be it tagging, matching, or indexing – has flaws. In our new preprint, we tested them all 💪We also proposed how to improve one of them. arxiv.org/abs/2601.16946

Ondrej Bojar on prompt dilution and the role of red team in deep learning models, processes essential for understanding why AI models sometimes bypass safety guardrails, and how to mprove their robustness. Watch here (in Czech): 🔗 www.ceskatelevize.cz/porady/10969...