Prague, Czech Republic
Joined April 2017
Patrícia Schmidtová retweeted
Got disappointing news from ACL today about an evaluation paper you were excited about? You can still commit ARR-reviewed papers to the GEM workshop until Thursday! Link below 👇
Are we facing a measurement crisis? As AI systems become more sophisticated, benchmarks are struggling to keep up. Distinguishing true progress from noise is harder than ever. 📉 The 5th Generation, Evaluation, and Metrics (GEM) Workshop is back at @ACL2026! 🧵
1
4
11
1,513
Patrícia Schmidtová retweeted
Last CFP: Workshop on Confabulation, Hallucinations, & Over-generation in Multilingual & Precision-critical Setting - CHOMPS 2025 Venue: AACL-IJCNLP2025, Mumbai, India Deadline: 29th September AOE More info: chomps2025.github.io – at Nancy, France
🎉 CHOMPS got accepted at #AACL2025 @aaclmeeting. See you in Mumbai ! #NLPProc Stay tuned for more!
2
2
204
Congratulations to @tuetschek 💐🥂🥳🏆👏 for receiving @prg_ai 's AI Award 'AI in Research and Development' for his excellent work on Natural Language Generation and Conversational Systems! @CharlesUniPRG @UniKarlova @marfyz @ERC_Research @tuetschek @vedavyzkum_cz @PavlaHub
3
11
25
1,230
Very thankful to have our work recognized! 😊
This Post is from an account that no longer exists. Learn more
9
299
Patrícia Schmidtová retweeted
Practical D2T @inlgmeeting Our hackathon continues: participants are now brainstorming possible error categories they found in LLM-generated forecasts.
1
1
6
471
Patrícia Schmidtová retweeted
Practical D2T @inlgmeeting We now continue with our panel of experts on the challenges of using LLMs in D2T tasks. With us today we have Claire Gardent (CNRS), C. Maria Keet (University of Cape Town), Michela Lorandi, (ADAPT/DCU), and Craig Thomson (ADAPT/DCU)
1
2
188
Patrícia Schmidtová retweeted
The PracticalD2T @inlgmeeting hackathon is now underway! The participants are now exploring the errors of LLM-generated weather forecasts and annotating them using our tool FactGenie for error span annotation.
1
4
197
Patrícia Schmidtová retweeted
We are starting this week's @inlgmeeting with the second edition of our PracticalD2T workshop. Today, we're about to see: - 2 inspirational keynotes - 3 paper presentations - a wonderful panel discussion - a hands-on hackathon Come join us if you're in Tokyo!
1
1
6
631
I loved the talk of @gfarnadi at @WiMLworkshop on algorithmic fairness. We have much to do as a community to aim for true fairness rather than quick post-processing band-aid hacks.
2
10
568
Patrícia Schmidtová retweeted
Happy to announce the 2nd Workshop on Practical LLM-assisted Data-to-Text Generation, at @inlgmeeting 2024 in Tokyo! 🏯 Website: practicald2t.github.io/ The call for papers is out (deadline: Jul 22 AoE), here is a short thread of what you can expect from our workshop (1/6)
1
9
2
23
2,005
The entire @ufal_cuni crowd at #LRECCOLING2024 😍🤩 In total, there are 15 of our papers at the main conference. Also, we helped the Czech Republic 🇨🇿 having the third highest acceptance rate to the conference (18/29 ≈ 62%). 💪
1
6
32
1,525
Patrícia Schmidtová retweeted
Not bad for my first time at #EACL2024 with @ufal_cuni! Huge thanks to the committee and my coauthors @PSchmidtova @tuetschek @LangoMateusz!
Best Non-publicized Paper Award #EACL2024 presented by @simone
3
2
1
41
2,420
Patrícia Schmidtová retweeted
Thank you for coming at our #EACL2024 poster yesterday! Lots of insightful comments on data contamination in LLMs 😄 If you missed it or want more info, please come to our oral tomorrow at 11! @PSchmidtova @tuetschek
3
35
763
Patrícia Schmidtová retweeted
Going to Malta today 💪🏻 me, @PSchmidtova and @tuetschek will be at EACL so feel free to reach out if you're attending! 😄 And check out our paper on data contamination in closed LLMs!
🚨Happy to share that our paper "Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMs" has been accepted at #eacl2024.🚨 Huge thanks to my co-authors @PSchmidtova, @LangoMateusz and @tuetschek. Link: arxiv.org/abs/2402.03927
2
21
767
Patrícia Schmidtová retweeted
🚨Happy to share that our paper "Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMs" has been accepted at #eacl2024.🚨 Huge thanks to my co-authors @PSchmidtova, @LangoMateusz and @tuetschek. Link: arxiv.org/abs/2402.03927
3
21
6
78
19,739
We also have a strong presence at the workshops. On Thursday, don't miss Niyata Bafna’s presentation at @sigmorphon; @PSchmidtova poster at the Workshop on Narrative Understanding. On Friday: @tomlimi poster at @genderbias. 7/7
2
2
5
Registration for MT Marathon is now open! ufal.mff.cuni.cz/mtm22/regis… 📍 Prague, 📅 September 5-10, 2022 ❓is a week-long gathering of machine translation researchers, developers, students The registration is for free, but there is an option to pre-order onsite lunch. 😋
1
3
4