OP JAK na Malé Straně? Na Matfyzu? Na UFalu? ❌ Tur dorůstájící kohoutkové výšky až 215 cm na chatovém fóru jako original poster? ✅ Operační Program? ✅ Jan Ámos Komenský? Na Ufalu potřebujeme pomoc a proto hledáme 👉 "Specialistu správy projektů OP JAK" na plný úvazek s flexibilní pracovní dobou. 👉 https://lnkd.in/eqH2J5pt 👈 Těšte se na přátelský mezinárodní kolektiv a mnoho lingvistických zajímavostí 😊
Ústav formální a aplikované lingvistiky
Vyšší vzdělávání
Prague 1, Prague 458 sledující uživatelů
Ústav na MFF UK. Vyučujeme. Bádáme. Vydáváme open-source software i data. Publikujeme o jazyce a algoritmech.
O nás
Ústav formální a aplikované lingvistiky (ÚFAL) na Matematicko-fyzikální fakulty Univerzity Karlovy. Ústav vznikl v roce 1990 jako pokračování výzkumné a pedagogické činnosti, kterou od počátku 60. let prováděla bývalá Laboratoř algebraické lingvistiky na Filozofické fakultě a později na Matematicko-fyzikální fakultě UK. Ústav se věnuje především výzkum, kde se zabývá tématy z oblasti počítačové lingvistiky. Ústav je také garantem uceleného výukového programu jak pro magisterské studium (Mgr.), tak pro doktorské studium (Ph.D.) v oboru počítačová lingvistika. Oba programy jsou vyučovány v českém a anglickém jazyce. Ústav je rovněž členem dvouoborového „magisterského programu LCT" EU.
- Web
-
https://ufal.mff.cuni.cz/
Externí odkaz pro organizaci Ústav formální a aplikované lingvistiky
- Obor
- Vyšší vzdělávání
- Velikost společnosti
- 51 - 200 zaměstnanců
- Ústředí
- Prague 1, Prague
- Typ
- Vzdělávací společnost
- Datum založení
- 1990
- Speciality
- Computational Linguistics, Machine Translation, Corpus Linguistics, Dialogue Systems, LLMs, Conversational AI, Teaching a Open-Source
Lokality
-
Primární
Získat instrukce na cestu
Malostranské náměstí
Prague 1, Prague 11800, CZ
Zaměstnanci společnosti Ústav formální a aplikované lingvistiky
-
Ondřej Plátek
LLM&TTS Evaluation researcher | Ph.D. Candidate @ UFAL, Prague | Engineer | Speaker | Father
-
Milan Straka
Research Assistant at Institute of Formal and Applied Linguistics, Charles University in Prague
-
Dima Taji
PhD student at Charles University, Faculty of Mathematics and Physics, Institute of Formal and Applied Linguistics (UFAL).
-
Rudolf Rosa
výzkumník v oblasti komputační lingvistiky a umělé inteligence
Aktualizace
-
Stroj + člověk = ideální překladatel 🤔 ⁉️ 👇 Registrujte se na dnešní přednášku Ondřej Bojar a připojte se na Zoom ⏰ Přednáška je dnes od 18:00 🙏 Díky Česká centra - Czech Centres za její organizaci 👉 https://lnkd.in/e8PbeS-5 👈 Strojový překlad je „královskou disciplínou“ oboru zvaného počítačová lingvistika. Za svou historii prodělal několik zásadních technologických změn, ty nejviditelnější v roce 2014 (nástup hlubokého strojového učení a neuronových sítí), 2017 (vznik modelu Transformer) a 2022 (vznik velkých jazykových modelů).... více info na registrační stránce a přednášce 😉
-
Uživatel Ústav formální a aplikované lingvistiky to přesdílel
Do you think you know a good way of telling if #LLM is telling you the truth? 🌍 Participate in Mu-SHROOM 🍄 – a SemEval 2025 shared task on detecting hallucination spans in multilingual LLM outputs. The task focuses on identifying when large language models generate information that isn’t true, with a multilingual context that includes Czech and regional Czech questions 🇨🇿: tricky questions that only Czechs know (and also someone wrote them on the Czech Wikipedia). + 13 other languages: Modern Standard Arabic, Basque, Catalan, Mandarin Chinese, English, Farsi, Finnish, French, German, Hindi, Italian, Spanish, and Swedish. 🤯 If you are interested in contributing to research on overgeneration errors in #NLP, this is an opportunity to get involved. Learn more and participate here 👉 https://lnkd.in/dF4V-kEF #MuSHROOM #SemEval2025 #NLP #MultilingualAI #HallucinationDetection
Welcome to the SemEval-2024 Task-6 - SHROOM, a Shared-task on Hallucinations and Related Observable Overgeneration Mistakes
helsinki-nlp.github.io
-
Thank you Jan Hajič, jr. for your hard work and enthusiasm! Music, algorithms, and enthusiasm are meant to be spread 😉 Check the syllabus and covered literature. Here is the link to the course 👉 https://lnkd.in/e4KTCFia 👈
So my first semester of teaching Computational Music Processing at Ústav formální a aplikované lingvistiky is a wrap. Basic stats: - 13 lectures at 90 minutes each, 935 slides (would have been 14 lectures, if not for how the Christmas break was timed). Approx. 70 slides per lecture, or slightly less than 1 per minute (with 10 mins. allowance for required reading discussions). - 27 people officially signed up for the course, with approx. 20 attending each lecture (this stabilized by week 3, and the last lectures two lectures had fewer students due to the proximity of Christmas) - Prep time: between 10 and 20 hours per lecture - Total articles referenced directly in the lectures: 97 (I skimmed probably x2-3 of that) Main takeaways: - Teaching is hard! I underestimated the workload to prepare a lecture by about 30-50 %. But also rewarding. - Designing learning outcomes & content for a course that has both bachelors and masters students was a challenge, but when I figured out what the course should be and committed to it, it kept me on track quite well. - The most recent reviews are not necessarily best. Quality of writing in review articles matters a lot; next time I participate in writing a review article I hope this experience will prove helpful. - Preparing for lectures related to my own research topics took as long as other topics (but was less stressful). - Slightly over 50 % of the students read the assigned papers weekly. The percentage of students actively taking part in the discussion grew from 3-4 to a peak of about 10. So, this element of the course was probably not pointless. But it’s really hard to involve remote participants in the discussion. - Detours into adjacent fields (musicology, copyright & music distribution, the realities of a musician’s life…) were well-received. - I would have preferred to have the 14 lectures instead of 13. I had to decide between music cognition and more focus on non-western musical styles. It was possible to weave mentions of non-western music into other topics and select required readings with cross-cultural materials, so in the end the lecture was on music cognition. - Teaching the course fulfilled my own learning outcomes, especially broadening my overview of music computing to things outside my lines of research. I’m now looking forward to the exams, which will in fact be the best feedback, but also to receiving explicit student feedback about different elements of the course: the talks, the slides, required readings and the discussion, the exam itself… I expect next year to still be a lot of work to improve the course, but hopefully it will not be as much work, and the student feedback will put me on more solid ground. If feedback is good, I plan to advertise it a bit more broadly, since it’s probably the only course on ISMIR-ish topics in the country (there is plenty on digital music *production*, but not so much on computational processing of music as the *input*). We’ll see.
-
-
Congratulations to Marie Mikulova, Barbora Štěpánková a Jan Štěpánek! 🥳 Their paper was accepted to Coling 2025, Abu Dhabi 🥳 📖 From Form to Meaning: The Case of Particles within the Prague Dependency Treebank Annotation Scheme Look at their previous research: 👉 Marie https://lnkd.in/eVbuzp9Q 👉 Barbora https://lnkd.in/eSFSz42M 👉 Honza https://lnkd.in/eS8fbsB8
-
-
🥳 Good news! GAČR will fund two of our projects 🥳 👉 Jindřich Libovický proposes better tokenization for #LLMs and machine translation 🇨🇿 🇪🇺 🇩🇪 👉 Veronika Kolářová will study syntactic features of Czech non-verbal predicates 🇨🇿 😶 ✍ 💪 Dominik Macháček receives Postdoc Individual Fellowship! Look at their research: 🔍Jindra https://lnkd.in/ez6pD2DH 🔍Veronika https://lnkd.in/ePY3NF8K 🔍Dominik https://lnkd.in/e2s-arfs
-
-
-
-
-
+ 1
-
-
Uživatel Ústav formální a aplikované lingvistiky to přesdílel
Yesterday, I participated in the Language Data Space Workshop for Czechia, organized by the Institute of Formal and Applied Linguistics, Charles University and led by Jan Hajic. https://lnkd.in/e28uFVUr There were many interesting insights on data sharing and the data itself from both business and academic perspectives. It was great to meet and talk with the panelists (in two consecutive panels): Petr Schwarz, Jirka Hana, Veronika Krejčířová, Naďa Dřizga, and Aleš Tamchyna. Thanks Barbora Vidova Hladka for the picture!
-
-
Uživatel Ústav formální a aplikované lingvistiky to přesdílel
🌐 Join us for a Language Data and Technology Seminar in Prague! 📅 When: December 2, 2024 (9:00–15:30) 📍 Where: Faculty of Mathematics and Physics, Charles University, Malostranské náměstí 25/2, Prague 1 (House for Professed, 1st Floor) ✨ Discover the latest in AI and language technologies at this seminar organized by the European project European Language Data Space and the Institute of Formal and Applied Linguistics, Charles University Industry experts (The MAMA AI, Seznam.cz, a.s., Phrase), academics (Univerzita Karlova) and a member of the European Commission will share their insights, and a representative from Ministerstvo školství, mládeže a tělovýchovy (MŠMT) will discuss funding opportunities for projects in the Czech Republic. Don’t miss this chance to connect with the leaders shaping the future of language technologies! 🔗 Learn more & view the full program: https://lnkd.in/ezASVymf To ensure your spot (and lunch), please register here: https://lnkd.in/eP-ZJ8VP
-
-
📢 Demo & business announcement 📢 See the demo and contact the team! 😉 These guys have won several awards in MT, and real-time interpreting with small latency. 🙏 Your feedback on the demo and the business is much appreciated 👉 Reach out to the team!
We are running a new business -- **AI Interpreting as a service**. We are ready to help event organizers to easily extend language coverage of their events. Please let us know if you are interested, or connect us to those who could be. Why AI Interpreting? E.g.: - Companies or organizations can invite foreign partners to presentations in their local language. - Field experts with less language skills can meet their foreign partners. - Training events and lectures can be held even for very multi-lingual participants at once, even for those who don't use a foreign language actively. - Cross-lingual interactions can be much easier. Journalists can do interviews in a language they don't know but the respondents are most comfortable with. - And much more. We serve 56 source speech languages and 43 target text languages. We use hardware, software and AI models that are completely ours or open-sourced. See demo, description and contact at https://lnkd.in/eBhwG8q8 . Dominik Macháček, Ondřej Bojar Dávid Javorský Sangeet Sagar Peter Polak, Institute of Formal and Applied Linguistics, Charles University
-
-
📸 Najdete sebe nebo své známé na fotkách z přednášky Jindřich Libovický pro Juniorskou Univerzitu Karlovy? 📸 Jindra s populárním tématem velkých jazykových modelů zaplnil celou posluchárnu na MatFyzu na Malostranském Náměstí. Na studenty čekalo seznámení se základními principy AI a strojového učení a mimojiné vysvětlení, jak je možné, že jazykové modely ovládají češtinu. Již tento čtvrtek se můžete těšit na předposlední přednášku v tomto semestru 🎉Hledání léku v kupce molekul, aneb počítačový vývoj léčiv s David Hoksza🎉 👉 Více o přednáškách JUK zde https://lnkd.in/eHGPmtHY 🎥 Podívejte se také na video k novému článku Jindřich Libovický a Jindra Helcl o tokenizaci LLM. Jindrové popularizují i nejnovější výzkum a zvládnou ukázat také kampus MFF v Troji 😉 https://lnkd.in/eHHujyDU #juniorskauniverzitak #LLM #research #education #UFAL #MatFyz #MFF #UK
-