UKP Lab
@ukplab.bsky.social
1.1K followers 340 following 990 posts
The Ubiquitous Knowledge Processing Lab researches Natural Language Processing (#NLProc) with a strong emphasis on Large Language Models, Conversational AI & Question Answering | @cs-tudarmstadt.bsky.social · @TUDa.bsky.social https://www.ukp.tu-darmstadt
Posts Media Videos Starter Packs
ukplab.bsky.social
🔗 𝗥𝗲𝗹𝗮𝘁𝗲𝗱 𝗿𝗲𝘀𝗼𝘂𝗿𝗰𝗲𝘀
𝗔𝗥𝗥 𝗗𝗮𝘁𝗮 𝗖𝗼𝗹𝗹𝗲𝗰𝘁𝗶𝗼𝗻: arr-data.aclweb.org

𝗗𝗮𝗴𝘀𝘁𝘂𝗵𝗹 𝗦𝗲𝗺𝗶𝗻𝗮𝗿 𝗼𝗻 𝗣𝗲𝗲𝗿 𝗥𝗲𝘃𝗶𝗲𝘄: www.dagstuhl.de/en/seminars/...

(5/5)
ACL Rolling Review Data Collection (ARR-DC)
Collecting and curating a large-scale dataset of peer reviews and associated metadata from the ACL community.
arr-data.aclweb.org
ukplab.bsky.social
🙌 𝗛𝘂𝗴𝗲 𝘁𝗵𝗮𝗻𝗸𝘀 𝘁𝗼 𝗼𝘂𝗿 𝗰𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗼𝗿𝘀!
This release wouldn’t be possible without the contributions of: Sheng Lu, Nils Dycke, @atnafu.bsky.social, Thamar Solorio, Xiaodan Zhu, Koen Dercksen, Lizhen Qu, Margot Mieskes, @dirkhovy.bsky.social and @igurevych.bsky.social.

(4/🧵)
ukplab.bsky.social
🔓 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.

(3/🧵)
ukplab.bsky.social
This release adds 𝗮 𝘀𝘂𝗯𝘀𝘁𝗮𝗻𝘁𝗶𝗮𝗹 𝗮𝗺𝗼𝘂𝗻𝘁 𝗼𝗳 𝗻𝗲𝘄 𝗱𝗮𝘁𝗮. It again highlights 𝗻𝗲𝘄 𝗱𝗮𝘁𝗮 𝘁𝘆𝗽𝗲𝘀 (𝗺𝗲𝘁𝗮-𝗿𝗲𝘃𝗶𝗲𝘄𝘀, 𝗿𝗲𝗯𝘂𝘁𝘁𝗮𝗹𝘀) that enable novel research, and 𝗻𝗼𝗻-𝘀𝗰𝗿𝗮𝗽𝗮𝗯𝗹𝗲 𝘀𝗼𝘂𝗿𝗰𝗲𝘀 for uncontaminated evals!

(2/🧵)
ukplab.bsky.social
🚀 𝗟𝗮𝘁𝗲𝘀𝘁 𝗣𝗲𝗲𝗿 𝗥𝗲𝘃𝗶𝗲𝘄 𝗗𝗮𝘁𝗮𝘀𝗲𝘁 𝗥𝗲𝗹𝗲𝗮𝘀𝗲 𝗳𝗿𝗼𝗺 𝗔𝗥𝗥 𝟮𝟬𝟮𝟱!
tudatalib.ulb.tu-darmstadt.de/handle/tudat...

📊 𝗡𝗲𝘄𝗹𝘆 𝗮𝗱𝗱𝗲𝗱 𝗔𝗖𝗟 𝟮𝟬𝟮𝟱 𝗱𝗮𝘁𝗮:
✅ 𝟮𝗸 papers
✅ 𝟮𝗸 reviews
✅ 𝟴𝟰𝟵 meta-reviews
✅ 𝟭.𝟱𝗸 papers with rebuttals

(1/🧵)
Illustration of a woman wearing a graduation cap and a lab coat, holding a magnifying glass and examining a document. A mechanical parrot with gears and circuits is perched on her shoulder.
ukplab.bsky.social
The conference brought together talented researchers at the intersection of quantum computing and #NLP, sparking ideas and discussions that push the field forward.

➡️ Check out the event: qnlp.ai

(3/3)

#NLProc #UKPLab #TUDarmstadt #QuantumComputing
Quantum AI and NLP Conference 2025
Quantum AI and NLP Conference 2025 Website. The conference will be held from the 6th to the 8th of August 2025 in Bloomington, Indiana at Indiana University.
qnlp.ai
ukplab.bsky.social
This work was made possible through a great collaboration with

👩‍🔬 Anna Schroeder & Mariami Gachechiladze (Quantum Computing Group, @tuda.bsky.social)

👨‍🔬 Yue Zhang (Westlake University)

👩‍💻 @igurevych.bsky.social (@ukplab.bsky.social, @tuda.bsky.social)

(2/🧵)
Quantum AI and NLP Conference 2025
Quantum AI and NLP Conference 2025 Website. The conference will be held from the 6th to the 8th of August 2025 in Bloomington, Indiana at Indiana University.
qnlp.ai
ukplab.bsky.social
✨ Exciting news from the #QNLP Conference 2025 in Bloomington, Indiana!

@akatief.bsky.social gave a talk on “An Efficient Quantum Classifier Based on Hamiltonian Representations” — exploring how quantum computing can open new frontiers for natural language processing.

(1/🧵)
Federico Tiblias from the UKP Lab gives a presentation at the Quantum AI and NLP 2025 conference. He stands at a podium in front of a projection screen displaying a slide titled Hamiltonian Classifier with formulas and diagrams. An inset portrait of Federico Tiblias appears in the upper right corner. The UKP Lab logo and text “UKP Lab @ Quantum AI and NLP 2025” are shown at the bottom. Audience members are seated in the foreground.
ukplab.bsky.social
Consider following the authors @jitaishik.bsky.social, Prottay Kumar Adhikary, Hiba Arnaout, @igurevych.bsky.social, and @tanmoy-chak.bsky.social

(3/3)

#NLProc #NLP4MentalHealth #NLPsych
ukplab.bsky.social
We expose key gaps and chart how richer, more inclusive datasets can unlock the full potential of Mental Health AI.

📃Paper: arxiv.org/pdf/2508.09809

(2/🧵)
arxiv.org
ukplab.bsky.social
👉 Interested in working on AI for mental health? This is a 𝗺𝘂𝘀𝘁-𝗿𝗲𝗮𝗱 𝘀𝗶𝗻𝗰𝗲 𝘁𝗵𝗲 𝗱𝗮𝘁𝗮 𝗮𝘃𝗮𝗶𝗹𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗶𝘀 𝗮 𝗺𝗮𝗷𝗼𝗿 𝗯𝗼𝘁𝘁𝗹𝗲𝗻𝗲𝗰𝗸!

Our survey maps the mental health dataset landscape, covering a wide range of disorders, tasks, modalities, accessibility, synthetic data, and cultural reach.

(1/🧵)
Flowchart illustrating aspects of clinical mental health datasets. Starting from a dataset icon, branches include:

Disorders: depression, anxiety, PTSD, bipolar disorder, schizophrenia.

Accessibility: public, controlled, private.

Tasks: classification, questionnaire scores, response generation, analysis.

Synthetic Data: artificial data creation.

Modalities: single modality (text, audio, video, brain signals) and multiple modality combinations.

Languages/Cultures: diversity of linguistic and cultural contexts.
ukplab.bsky.social
📚 Discovery Science 2025 is part of the AI for Science Conference in #Ljubljana

🔗 ds2025.ijs.si/speakers/iry...

#DiscoveryScience2025 #AIforScience #AI #UKPLab
Iryna Gurevych | Discovery Science 2025
28th International Conference on Discovery Science 2025
ds2025.ijs.si
ukplab.bsky.social
Her talk “𝘗𝘭𝘦𝘢𝘴𝘦 𝘮𝘦𝘦𝘵 𝘈𝘐, 𝘰𝘶𝘳 𝘥𝘦𝘢𝘳 𝘯𝘦𝘸 𝘤𝘰𝘭𝘭𝘦𝘢𝘨𝘶𝘦. 𝘐𝘯 𝘰𝘵𝘩𝘦𝘳 𝘸𝘰𝘳𝘥𝘴: 𝘤𝘢𝘯 𝘴𝘤𝘪𝘦𝘯𝘵𝘪𝘴𝘵𝘴 𝘢𝘯𝘥 𝘮𝘢𝘤𝘩𝘪𝘯𝘦𝘴 𝘵𝘳𝘶𝘭𝘺 𝘤𝘰𝘰𝘱𝘦𝘳𝘢𝘵𝘦?” addresses how AI and LLMs can support scientists throughout different stages of the research process – from writing and revising manuscripts to peer review and scientific discussions.
Iryna Gurevych | Discovery Science 2025
28th International Conference on Discovery Science 2025
ds2025.ijs.si
ukplab.bsky.social
🚀 𝗞𝗲𝘆𝗻𝗼𝘁𝗲 𝗯𝘆 𝗣𝗿𝗼𝗳. 𝗜𝗿𝘆𝗻𝗮 𝗚𝘂𝗿𝗲𝘃𝘆𝗰𝗵 𝗶𝗻 𝗟𝗷𝘂𝗯𝗹𝗷𝗮𝗻𝗮
Prof. @igurevych.bsky.social, Head of the UKP Lab, is giving a keynote today at the 28th International Conference on Discovery Science 2025 in Ljubljana.
Promotional graphic for the Discovery Science 2025 International Conference keynote. The text reads: “Please meet AI, our dear new colleague. In other words: can scientists and machines truly cooperate?” Keynote by Prof. Iryna Gurevych. On the right is a photo of Prof. Iryna Gurevych standing in front of bookshelves. The logos of UKP Lab and Discovery Science are included at the bottom.
ukplab.bsky.social
Stay tuned as Yang reflects on his research path, current projects, and the lasting impact of his time at UKP Lab 💻💡.

(5/5)

#UKPLab #UKPAlumni #LanguageModels #RLHF #ReinforcementLearning #DeepMind #Gemini #NLProc #AI #TextGeneration #ComputationalLinguistics
ukplab.bsky.social
Before joining Google DeepMind in 2021, Yang held positions at Royal Holloway, University of London and at the Chinese Academy of Sciences, following a PhD at @imperialcollegeldn.bsky.social & and postdoc at @ukplab.bsky.social.

(4/🧵)
ukplab.bsky.social
During his time at UKP, Yang and his collaborators were among the first to explore 𝗥𝗟 𝘄𝗶𝘁𝗵 𝗛𝘂𝗺𝗮𝗻 𝗙𝗲𝗲𝗱𝗯𝗮𝗰𝗸 #RLHF for text generation, pioneering techniques that are foundational to today’s large language models. His work highlighted the importance of 𝗿𝗲𝘄𝗮𝗿𝗱 𝗺𝗼𝗱𝗲𝗹𝗶𝗻𝗴 in RL-based NLP systems.

(3/🧵)
ukplab.bsky.social
Yang joined the UKP Lab as a postdoctoral researcher in 2017, bringing with him a background in 𝗥𝗲𝗶𝗻𝗳𝗼𝗿𝗰𝗲𝗺𝗲𝗻𝘁 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 #RL. His time at UKP marked a turning point: it was here that he began integrating RL with #NLP — an approach that would go on to shape much of his later work.

(2/🧵)
ukplab.bsky.social
🚀 We’re excited to continue our 𝘚𝘱𝘰𝘵𝘭𝘪𝘨𝘩𝘵 𝘰𝘯 𝘜𝘒𝘗 𝘓𝘢𝘣 𝘈𝘭𝘶𝘮𝘯𝘪 series with Yang Gao, now a Research Scientist at #Google #DeepMind, where he works on the Google #Gemini language models.

(1/🧵)
Graphic for the UKP Lab Alumni Spotlight series. It features a circular portrait photo of Yang Gao, smiling and wearing glasses and a jacket, with a blurred building in the background. Text above the image reads: “Spotlight on UKP Lab Alumni.” Below the portrait: “Yang Gao, Research Scientist at Google DeepMind.” The bottom right corner shows the Ubiquitous Knowledge Processing (UKP) Lab logo.
ukplab.bsky.social
🔁 Prof. Iryna Gurevych, Director of the UKP Lab at TU Darmstadt, is featured in this ATHENE video portrait.

🎥 The video (in German) introduces her research on disinformation and cognitive security, where methods from #NLP intersect with questions of #cybersecurity.

⬇️ Watch the video portrait here
athenecenter.bsky.social
👤 Watch our video portrait about ATHENE Principal Investigator and renowned computer scientist Prof. @igurevych.bsky.social @tuda.bsky.social @ukplab.bsky.social!

🎞️To the portrait: www.youtube.com/watch?v=j3Or...

🎞️ The trailer highlights her research on #disinformation and cognitive security:
ukplab.bsky.social
And consider following the authors Furkan Şahinuç, Subhabrata Dutta, @igurevych.bsky.social for more information or an exchange of ideas. (8/8)
ukplab.bsky.social
For further details, check our project page, paper, dataset, and code!

🌐Project webpage: ukplab.github.io/arxiv2025-ex...
📄Paper: arxiv.org/abs/2508.07955
💻Code: github.com/UKPLab/arxiv...
📁 Data: tudatalib.ulb.tu-darmstadt.de/handle/tudat...

(7/🧵)
ukplab.bsky.social
𝗪𝗵𝘆 𝗗𝗼𝗲𝘀 𝗜𝘁 𝗠𝗮𝘁𝘁𝗲𝗿?
GREP bridges a gap in evaluating AI-generated scientific text—moving from superficial scoring to expert-aligned evaluation. Whether you’re interested in RLHF, domain-specific generation, or enhancing scientific writing pipelines, this work offers insight and robust tools.
(6/🧵)
ukplab.bsky.social
𝗥𝗲𝘀𝘂𝗹𝘁 𝘁𝗮𝗸𝗲𝗮𝘄𝗮𝘆𝘀: Strong LLMs often fail to satisfy hard constraints and struggle most with maintaining citation coherence, highlighting the need for iterative, feedback-driven human-AI collaboration to refine outputs. (5/🧵)
ukplab.bsky.social
GREP distinguishes between hard constraints (non-negotiable requirements, such as correct citations) and soft constraints (stylistic preferences and readability). This layered evaluation better reflects how experts truly assess quality in scientific writing. (4/🧵)