UKP Lab

Can we measure the peer review crisis? 📉 Everyone complains about declining review quality. But can it actually be quantified? A new #EMNLP2026 paper by Ilia Kuznetsov, Rohan Nayak, Alla Rozovskaya and @igurevych.bsky.social takes a closer look 🧵 #NLProc #PeerReview

Want to know more or discuss your ideas? Visit the poster: Monday 16:45, Session E, Hall H! Not at EMNLP? Preprint 👉 arxiv.org/abs/2601.15172

Can we measure the peer review crisis? 📉 Everyone complains about declining review quality. But can it actually be quantified? A new #EMNLP2026 paper by Ilia Kuznetsov, Rohan Nayak, Alla Rozovskaya and @igurevych.bsky.social takes a closer look 🧵 #NLProc #PeerReview

Finding: validated against two independent sets of human quality ratings, the measure shows no consistent decline in median review quality over the years.

Can we measure the peer review crisis? 📉 Everyone complains about declining review quality. But can it actually be quantified? A new #EMNLP2026 paper by Ilia Kuznetsov, Rohan Nayak, Alla Rozovskaya and @igurevych.bsky.social takes a closer look 🧵 #NLProc #PeerReview

Ahead of #EMNLP2026, researchers from Columbia, Edinburgh, Toronto, MBZUAI, Bocconi and many more meet at @tuda.bsky.social on 22 and 23 October. How are large language models changing research, and where does human judgement remain essential?

Programme and speakers: future-of-science.netlify.app #NLP #AIforScience

Ahead of #EMNLP2026, researchers from Columbia, Edinburgh, Toronto, MBZUAI, Bocconi and many more meet at @tuda.bsky.social on 22 and 23 October. How are large language models changing research, and where does human judgement remain essential?

“The Future of Science: AI-First, Trustworthy, Human-Steered” is organised by the UKP Lab, @zuseschooleliza.bsky.social and the @ellis.eu Unit Darmstadt, and hosted by Prof. @igurevych.bsky.social

🛡️ Welcome to the UKP Lab, Petros Raptopoulos! Petros joins us as a PhD student on the HAICC project (Human-AI Collaboration for Cybersecurity) at #ATHENE. His thesis explores how AI agents in security-critical applications can be safely deployed under human oversight. Welcome, Petros! 👋

🎓 𝗨𝗞𝗣 𝗟𝗮𝗯 𝘄𝗲𝗹𝗰𝗼𝗺𝗲𝘀 𝗠𝗼𝗵𝗮𝗺𝗺𝗮𝗱 𝗥𝗮𝘄𝗵𝗮𝗻𝗶! Our new PhD student, Mohammad, brings a research focus on 𝗡𝗮𝘁𝘂𝗿𝗮𝗹 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗣𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴, 𝗹𝗮𝗿𝗴𝗲 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗺𝗼𝗱𝗲𝗹𝘀, and 𝗺𝗮𝗰𝗵𝗶𝗻𝗲 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗳𝗼𝗿 𝗹𝗼𝘄-𝗿𝗲𝘀𝗼𝘂𝗿𝗰𝗲 𝘀𝗲𝘁𝘁𝗶𝗻𝗴𝘀. His previous academic work centered on 𝗱𝗼𝗺𝗮𝗶𝗻 𝗮𝗱𝗮𝗽𝘁𝗮𝘁𝗶𝗼𝗻 𝘂𝘀𝗶𝗻𝗴 𝗽𝗮𝗿𝗮𝗺𝗲𝘁𝗲𝗿-𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴 𝗺𝗲𝘁𝗵𝗼𝗱𝘀.

We're looking forward to seeing where your research takes you, Mohammad. Great to have you with us! 👋 #UKPLab #TUDarmstadt #NLP #LLMs #LowResourceNLP #MachineLearning

🎓 𝗨𝗞𝗣 𝗟𝗮𝗯 𝘄𝗲𝗹𝗰𝗼𝗺𝗲𝘀 𝗠𝗼𝗵𝗮𝗺𝗺𝗮𝗱 𝗥𝗮𝘄𝗵𝗮𝗻𝗶! Our new PhD student, Mohammad, brings a research focus on 𝗡𝗮𝘁𝘂𝗿𝗮𝗹 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗣𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴, 𝗹𝗮𝗿𝗴𝗲 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗺𝗼𝗱𝗲𝗹𝘀, and 𝗺𝗮𝗰𝗵𝗶𝗻𝗲 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗳𝗼𝗿 𝗹𝗼𝘄-𝗿𝗲𝘀𝗼𝘂𝗿𝗰𝗲 𝘀𝗲𝘁𝘁𝗶𝗻𝗴𝘀. His previous academic work centered on 𝗱𝗼𝗺𝗮𝗶𝗻 𝗮𝗱𝗮𝗽𝘁𝗮𝘁𝗶𝗼𝗻 𝘂𝘀𝗶𝗻𝗴 𝗽𝗮𝗿𝗮𝗺𝗲𝘁𝗲𝗿-𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴 𝗺𝗲𝘁𝗵𝗼𝗱𝘀.

🏆 𝗕𝗲𝘀𝘁 𝗣𝗮𝗽𝗲𝗿 𝗔𝘄𝗮𝗿𝗱 𝗮𝘁 𝗦𝗲𝗺𝗘𝘃𝗮𝗹-𝟮𝟬𝟮𝟲 𝗦𝗵𝗮𝗿𝗲𝗱 𝗧𝗮𝘀𝗸 𝟮! We're proud to share that work from UKP Lab has received the 𝗕𝗲𝘀𝘁 𝗣𝗮𝗽𝗲𝗿 𝗔𝘄𝗮𝗿𝗱 𝗼𝗳 𝗦𝗲𝗺𝗘𝘃𝗮𝗹-𝟮𝟬𝟮𝟲 𝗦𝗵𝗮𝗿𝗲𝗱 𝗧𝗮𝘀𝗸 𝟮, recognizing the scientific contribution behind a system that also topped the leaderboard with 𝟭𝘀𝘁 𝗽𝗹𝗮𝗰𝗲 𝗶𝗻 𝗦𝘂𝗯𝘁𝗮𝘀𝗸𝘀 𝟭 𝗮𝗻𝗱 𝟮𝗔 😲

The team has presented the work at the 𝗦𝗲𝗺𝗘𝘃𝗮𝗹-𝟮𝟬𝟮𝟲 𝘄𝗼𝗿𝗸𝘀𝗵𝗼𝗽 𝗮𝘁 𝗔𝗖𝗟 𝟮𝟬𝟮𝟲 𝗶𝗻 𝗦𝗮𝗻 𝗗𝗶𝗲𝗴𝗼 on 𝗦𝗮𝘁𝘂𝗿𝗱𝗮𝘆, 𝗝𝘂𝗹𝘆 𝟰𝘁𝗵. 🚀 📄 Preprint: arxiv.org/html/2604.21... 💻 Code: github.com/UKPLab/semev... #SemEval2026 #ACL2026 #NLP #UKPLab #Research #AI

🏆 𝗕𝗲𝘀𝘁 𝗣𝗮𝗽𝗲𝗿 𝗔𝘄𝗮𝗿𝗱 𝗮𝘁 𝗦𝗲𝗺𝗘𝘃𝗮𝗹-𝟮𝟬𝟮𝟲 𝗦𝗵𝗮𝗿𝗲𝗱 𝗧𝗮𝘀𝗸 𝟮! We're proud to share that work from UKP Lab has received the 𝗕𝗲𝘀𝘁 𝗣𝗮𝗽𝗲𝗿 𝗔𝘄𝗮𝗿𝗱 𝗼𝗳 𝗦𝗲𝗺𝗘𝘃𝗮𝗹-𝟮𝟬𝟮𝟲 𝗦𝗵𝗮𝗿𝗲𝗱 𝗧𝗮𝘀𝗸 𝟮, recognizing the scientific contribution behind a system that also topped the leaderboard with 𝟭𝘀𝘁 𝗽𝗹𝗮𝗰𝗲 𝗶𝗻 𝗦𝘂𝗯𝘁𝗮𝘀𝗸𝘀 𝟭 𝗮𝗻𝗱 𝟮𝗔 😲

The paper, 𝗨𝗞𝗣_𝗣𝘀𝘆𝗰𝗼𝗻𝘁𝗿𝗼𝗹 𝗮𝘁 𝗦𝗲𝗺𝗘𝘃𝗮𝗹-𝟮𝟬𝟮𝟲 𝗧𝗮𝘀𝗸 𝟮: 𝗠𝗼𝗱𝗲𝗹𝗶𝗻𝗴 𝗩𝗮𝗹𝗲𝗻𝗰𝗲 𝗮𝗻𝗱 𝗔𝗿𝗼𝘂𝘀𝗮𝗹 𝗗𝘆𝗻𝗮𝗺𝗶𝗰𝘀 𝗳𝗿𝗼𝗺 𝗧𝗲𝘅𝘁, tackles a deceptively hard question: can we predict not just what someone feels from their writing, but how their emotions evolve over time?

🏆 𝗕𝗲𝘀𝘁 𝗣𝗮𝗽𝗲𝗿 𝗔𝘄𝗮𝗿𝗱 𝗮𝘁 𝗦𝗲𝗺𝗘𝘃𝗮𝗹-𝟮𝟬𝟮𝟲 𝗦𝗵𝗮𝗿𝗲𝗱 𝗧𝗮𝘀𝗸 𝟮! We're proud to share that work from UKP Lab has received the 𝗕𝗲𝘀𝘁 𝗣𝗮𝗽𝗲𝗿 𝗔𝘄𝗮𝗿𝗱 𝗼𝗳 𝗦𝗲𝗺𝗘𝘃𝗮𝗹-𝟮𝟬𝟮𝟲 𝗦𝗵𝗮𝗿𝗲𝗱 𝗧𝗮𝘀𝗸 𝟮, recognizing the scientific contribution behind a system that also topped the leaderboard with 𝟭𝘀𝘁 𝗽𝗹𝗮𝗰𝗲 𝗶𝗻 𝗦𝘂𝗯𝘁𝗮𝘀𝗸𝘀 𝟭 𝗮𝗻𝗱 𝟮𝗔 😲

🎓 𝗚𝘂𝗲𝘀𝘁 𝗧𝗮𝗹𝗸 𝗮𝘁 𝘁𝗵𝗲 𝗨𝗞𝗣 𝗟𝗮𝗯 We were thrilled to welcome our long-term collaborator, co-supervisor of PhD students and close colleague Prof. Preslav Nakov, Professor and Department Chair for #NLP at the #MBZUAI in Abu Dhabi, for an invited talk at the UKP Lab in #TUDarmstadt.

We thank Prof. Nakov for his exciting talk and the lively exchange with UKP members on trustworthy, evidence-grounded, and human-centered NLP. #UKPLab #NLProc #LLMs #TrustworthyAI #GuestTalk #FinancialAI #AITextDetection

🎓 𝗚𝘂𝗲𝘀𝘁 𝗧𝗮𝗹𝗸 𝗮𝘁 𝘁𝗵𝗲 𝗨𝗞𝗣 𝗟𝗮𝗯 We were thrilled to welcome our long-term collaborator, co-supervisor of PhD students and close colleague Prof. Preslav Nakov, Professor and Department Chair for #NLP at the #MBZUAI in Abu Dhabi, for an invited talk at the UKP Lab in #TUDarmstadt.

🔍 In the first part, 𝐻𝑢𝑚𝑎𝑛 𝑜𝑟 𝐴𝐼?, he examined the blurred boundary between human and machine-generated text. He introduced ExaGPT, an interpretable detector grounded in concrete text spans, and shared a multilingual study showing that human-like text is not necessarily the text people prefer

𝗙𝗿𝗼𝗺 𝗗𝗮𝗿𝗺𝘀𝘁𝗮𝗱𝘁 𝘁𝗼 𝗔𝗖𝗟 𝟮𝟬𝟮𝟲 𝗶𝗻 𝗦𝗮𝗻 𝗗𝗶𝗲𝗴𝗼 Four UKP Lab papers got an early audience this week. @timbmg.bsky.social, @ibigoulaeva.bsky.social, and @tongletj.bsky.social presented their work at the Post-CVPR Pre-ICML Poster Session, [...]

Heading to San Diego as well? Come find us 🌵 And to everyone who stopped by in Darmstadt: thanks for the great conversations! 🙌 #NLProc #LLMs #NLP #MachineLearning #Multimodal #ACL #Ellis #Zuse

𝗙𝗿𝗼𝗺 𝗗𝗮𝗿𝗺𝘀𝘁𝗮𝗱𝘁 𝘁𝗼 𝗔𝗖𝗟 𝟮𝟬𝟮𝟲 𝗶𝗻 𝗦𝗮𝗻 𝗗𝗶𝗲𝗴𝗼 Four UKP Lab papers got an early audience this week. @timbmg.bsky.social, @ibigoulaeva.bsky.social, and @tongletj.bsky.social presented their work at the Post-CVPR Pre-ICML Poster Session, [...]

▸ 𝗜𝘀 𝘁𝗵𝗶𝘀 𝗰𝗵𝗮𝗿𝘁 𝗹𝘆𝗶𝗻𝗴 𝘁𝗼 𝗺𝗲? 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗻𝗴 𝘁𝗵𝗲 𝗱𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻 𝗼𝗳 𝗺𝗶𝘀𝗹𝗲𝗮𝗱𝗶𝗻𝗴 𝘃𝗶𝘀𝘂𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻𝘀 Jonathan Tonglet, Jan Zimny, Tinne Tuytelaars, Iryna Gurevych arxiv.org/abs/2508.21675

𝗙𝗿𝗼𝗺 𝗗𝗮𝗿𝗺𝘀𝘁𝗮𝗱𝘁 𝘁𝗼 𝗔𝗖𝗟 𝟮𝟬𝟮𝟲 𝗶𝗻 𝗦𝗮𝗻 𝗗𝗶𝗲𝗴𝗼 Four UKP Lab papers got an early audience this week. @timbmg.bsky.social, @ibigoulaeva.bsky.social, and @tongletj.bsky.social presented their work at the Post-CVPR Pre-ICML Poster Session, [...]

▸ 𝗦𝗰𝗶𝗖𝗼𝗤𝗔: 𝗤𝘂𝗮𝗹𝗶𝘁𝘆 𝗔𝘀𝘀𝘂𝗿𝗮𝗻𝗰𝗲 𝗳𝗼𝗿 𝗦𝗰𝗶𝗲𝗻𝘁𝗶𝗳𝗶𝗰 𝗣𝗮𝗽𝗲𝗿-𝗖𝗼𝗱𝗲 𝗔𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁 Tim Baumgärtner, Iryna Gurevych arxiv.org/abs/2601.12910

𝗙𝗿𝗼𝗺 𝗗𝗮𝗿𝗺𝘀𝘁𝗮𝗱𝘁 𝘁𝗼 𝗔𝗖𝗟 𝟮𝟬𝟮𝟲 𝗶𝗻 𝗦𝗮𝗻 𝗗𝗶𝗲𝗴𝗼 Four UKP Lab papers got an early audience this week. @timbmg.bsky.social, @ibigoulaeva.bsky.social, and @tongletj.bsky.social presented their work at the Post-CVPR Pre-ICML Poster Session, [...]

👋 𝗡𝗲𝘄 𝗳𝗮𝗰𝗲 𝗮𝘁 𝘁𝗵𝗲 𝗨𝗞𝗣 𝗟𝗮𝗯: 𝗔 𝘄𝗮𝗿𝗺 𝘄𝗲𝗹𝗰𝗼𝗺𝗲 𝘁𝗼 𝗔𝗻𝗱𝗿𝗲𝗮! @andreamoleri.bsky.social joins us as a doctoral researcher, jointly supervised by Prof. Dr. @igurevych.bsky.social and Prof. Dr. Marcus Rohrbach.

Great to have you with us, Andrea, and wishing you a strong start to your PhD journey! 🚀 #UKPLab #TUDarmstadt #LLMs #AISafety #NLP #TrustworthyAI #NLProc

👋 𝗡𝗲𝘄 𝗳𝗮𝗰𝗲 𝗮𝘁 𝘁𝗵𝗲 𝗨𝗞𝗣 𝗟𝗮𝗯: 𝗔 𝘄𝗮𝗿𝗺 𝘄𝗲𝗹𝗰𝗼𝗺𝗲 𝘁𝗼 𝗔𝗻𝗱𝗿𝗲𝗮! @andreamoleri.bsky.social joins us as a doctoral researcher, jointly supervised by Prof. Dr. @igurevych.bsky.social and Prof. Dr. Marcus Rohrbach.

His PhD focuses on 𝗮𝗹𝗲𝗮𝘁𝗼𝗿𝗶𝗰 𝘂𝗻𝗱𝗲𝗿𝘀𝗽𝗲𝗰𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻 in large language models, with the goal of developing methods to detect and resolve it, so that LLM behavior stays bounded and safe in security-critical settings.

🚀 𝗟𝗮𝘁𝗲𝘀𝘁 𝗡𝗟𝗣𝗲𝗲𝗿 𝗥𝗲𝗹𝗲𝗮𝘀𝗲 𝗳𝗿𝗼𝗺 𝗔𝗥𝗥 𝟮𝟬𝟮𝟲 𝗶𝘀 𝗼𝘂𝘁! Grab your Peer Review data: tudatalib.ulb.tu-darmstadt.de/handle/tudat... 📊 𝗘𝗔𝗖𝗟 𝟮𝟲 𝗔𝗰𝗰𝗲𝗽𝘁𝗲𝗱 𝗣𝗮𝗽𝗲𝗿 𝗗𝗮𝘁𝗮: ✅ 647 papers ✅ 942 reviews ✅ 329 meta-reviews ✅ 550 papers with rebuttals

🔗 𝗥𝗲𝗹𝗮𝘁𝗲𝗱 𝗥𝗲𝘀𝗼𝘂𝗿𝗰𝗲𝘀 ARR Data Collection: arr-data.aclweb.org Dagstuhl Seminar 2024 on Peer Review: www.dagstuhl.de/en/seminars/... #NLP #NLProc #PeerReview #MachineLearning #ArtificialIntelligence #NLPeer #ARR #EACL2026 #OpenScience #OpenData #ResearchData #AIResearch #UKPLab #TUDarmstadt

🚀 𝗟𝗮𝘁𝗲𝘀𝘁 𝗡𝗟𝗣𝗲𝗲𝗿 𝗥𝗲𝗹𝗲𝗮𝘀𝗲 𝗳𝗿𝗼𝗺 𝗔𝗥𝗥 𝟮𝟬𝟮𝟲 𝗶𝘀 𝗼𝘂𝘁! Grab your Peer Review data: tudatalib.ulb.tu-darmstadt.de/handle/tudat... 📊 𝗘𝗔𝗖𝗟 𝟮𝟲 𝗔𝗰𝗰𝗲𝗽𝘁𝗲𝗱 𝗣𝗮𝗽𝗲𝗿 𝗗𝗮𝘁𝗮: ✅ 647 papers ✅ 942 reviews ✅ 329 meta-reviews ✅ 550 papers with rebuttals

🔓 𝗢𝗽𝗲𝗻 & 𝗧𝗿𝗮𝗻𝘀𝗽𝗮𝗿𝗲𝗻𝘁 𝗟𝗶𝗰𝗲𝗻𝘀𝗶𝗻𝗴 All data is released under open and transparent licenses to enable large-scale NLP and AI research on AI-based peer review assistance.

🚀 𝗟𝗮𝘁𝗲𝘀𝘁 𝗡𝗟𝗣𝗲𝗲𝗿 𝗥𝗲𝗹𝗲𝗮𝘀𝗲 𝗳𝗿𝗼𝗺 𝗔𝗥𝗥 𝟮𝟬𝟮𝟲 𝗶𝘀 𝗼𝘂𝘁! Grab your Peer Review data: tudatalib.ulb.tu-darmstadt.de/handle/tudat... 📊 𝗘𝗔𝗖𝗟 𝟮𝟲 𝗔𝗰𝗰𝗲𝗽𝘁𝗲𝗱 𝗣𝗮𝗽𝗲𝗿 𝗗𝗮𝘁𝗮: ✅ 647 papers ✅ 942 reviews ✅ 329 meta-reviews ✅ 550 papers with rebuttals