UKP Lab
@ukplab.bsky.social
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
https://www.ukp.tu-darmstadt
We wish all students the best of luck as they finalize their submissions by the end of this month! ✨
(3/3)
#UKPLab #ThesisDay #NLP #NLProc
(3/3)
#UKPLab #ThesisDay #NLP #NLProc
November 21, 2025 at 2:55 PM
We wish all students the best of luck as they finalize their submissions by the end of this month! ✨
(3/3)
#UKPLab #ThesisDay #NLP #NLProc
(3/3)
#UKPLab #ThesisDay #NLP #NLProc
The day featured well-prepared talks, thoughtful questions, and lively exchanges across topics.
A big thank you to Yongxin Huang, our thesis coordinator, for guiding this cohort through the process, and to all supervisors for their continuous support throughout the semester.
(2/🧵)
A big thank you to Yongxin Huang, our thesis coordinator, for guiding this cohort through the process, and to all supervisors for their continuous support throughout the semester.
(2/🧵)
November 21, 2025 at 2:55 PM
The day featured well-prepared talks, thoughtful questions, and lively exchanges across topics.
A big thank you to Yongxin Huang, our thesis coordinator, for guiding this cohort through the process, and to all supervisors for their continuous support throughout the semester.
(2/🧵)
A big thank you to Yongxin Huang, our thesis coordinator, for guiding this cohort through the process, and to all supervisors for their continuous support throughout the semester.
(2/🧵)
For those interested, the Royal Society has made the event available online:
🎥 royalsociety.org/science-even...
🔗 The presentation shown during the lecture can be accessed separately here:
next.hessenbox.de/index.php/s/...
🎥 royalsociety.org/science-even...
🔗 The presentation shown during the lecture can be accessed separately here:
next.hessenbox.de/index.php/s/...
Milner Prize Lecture: How to spot and debunk misleading content | Royal Society
Royal Society Milner Prize Lecture 2025 delivered by Professor Iryna Gurevych
royalsociety.org
November 18, 2025 at 2:11 PM
For those interested, the Royal Society has made the event available online:
🎥 royalsociety.org/science-even...
🔗 The presentation shown during the lecture can be accessed separately here:
next.hessenbox.de/index.php/s/...
🎥 royalsociety.org/science-even...
🔗 The presentation shown during the lecture can be accessed separately here:
next.hessenbox.de/index.php/s/...
As part of the ceremony at the Royal Society in London, @igurevych.bsky.social delivered the Milner Prize Lecture titled 𝘏𝘰𝘸 𝘵𝘰 𝘴𝘱𝘰𝘵 𝘢𝘯𝘥 𝘥𝘦𝘣𝘶𝘯𝘬 𝘮𝘪𝘴𝘭𝘦𝘢𝘥𝘪𝘯𝘨 𝘤𝘰𝘯𝘵𝘦𝘯𝘵, addressing current challenges in detecting and countering misleading information.
November 18, 2025 at 2:11 PM
As part of the ceremony at the Royal Society in London, @igurevych.bsky.social delivered the Milner Prize Lecture titled 𝘏𝘰𝘸 𝘵𝘰 𝘴𝘱𝘰𝘵 𝘢𝘯𝘥 𝘥𝘦𝘣𝘶𝘯𝘬 𝘮𝘪𝘴𝘭𝘦𝘢𝘥𝘪𝘯𝘨 𝘤𝘰𝘯𝘵𝘦𝘯𝘵, addressing current challenges in detecting and countering misleading information.
The award recognises outstanding contributions to machine learning and natural language processing and highlights the impact of her long-standing research on trustworthy AI and computational linguistics.
November 18, 2025 at 2:11 PM
The award recognises outstanding contributions to machine learning and natural language processing and highlights the impact of her long-standing research on trustworthy AI and computational linguistics.
Following this, Subhabrata Dutta, Jingcheng (Frank) Niu, Simone Balloccu and Jonathan Tonglet will introduce current research directions from the UKP Lab.
November 18, 2025 at 12:15 PM
Following this, Subhabrata Dutta, Jingcheng (Frank) Niu, Simone Balloccu and Jonathan Tonglet will introduce current research directions from the UKP Lab.
Talia Tseriotou – Tempoformer and Personalized Long-Context Generation
Anthony Hills – Multi-stream Data Annotation and Modeling
Mahmud Elahi Akhter – Reasoning in Evidence-based Claim Verification and Next Steps
Yuxiang Zhou & Gunnet Singh Kohli – Temporal Reasoning and Reference-free Evaluation
Anthony Hills – Multi-stream Data Annotation and Modeling
Mahmud Elahi Akhter – Reasoning in Evidence-based Claim Verification and Next Steps
Yuxiang Zhou & Gunnet Singh Kohli – Temporal Reasoning and Reference-free Evaluation
November 18, 2025 at 12:15 PM
Talia Tseriotou – Tempoformer and Personalized Long-Context Generation
Anthony Hills – Multi-stream Data Annotation and Modeling
Mahmud Elahi Akhter – Reasoning in Evidence-based Claim Verification and Next Steps
Yuxiang Zhou & Gunnet Singh Kohli – Temporal Reasoning and Reference-free Evaluation
Anthony Hills – Multi-stream Data Annotation and Modeling
Mahmud Elahi Akhter – Reasoning in Evidence-based Claim Verification and Next Steps
Yuxiang Zhou & Gunnet Singh Kohli – Temporal Reasoning and Reference-free Evaluation
The event opens with contributions from Prof. Maria Liakata’s lab, offering insights into current research:
- Maria Liakata – Adsolve: Addressing Socio-technical Limitations of LLMs for Medical and Social Computing
- Maria Liakata – Adsolve: Addressing Socio-technical Limitations of LLMs for Medical and Social Computing
November 18, 2025 at 12:15 PM
The event opens with contributions from Prof. Maria Liakata’s lab, offering insights into current research:
- Maria Liakata – Adsolve: Addressing Socio-technical Limitations of LLMs for Medical and Social Computing
- Maria Liakata – Adsolve: Addressing Socio-technical Limitations of LLMs for Medical and Social Computing
We are grateful to everyone joining us - onsite or online - to reconnect, share insights and celebrate the past, present and future of the UKP Lab 🎉
(4/4)
(4/4)
November 17, 2025 at 11:54 AM
We are grateful to everyone joining us - onsite or online - to reconnect, share insights and celebrate the past, present and future of the UKP Lab 🎉
(4/4)
(4/4)
The second part of the workshop features extended talks from:
Noam Slonim, Dan Roth, @andreasvlachos.bsky.social and Ido Dagan.
Our thanks go to
Anna Korhonen for opening the workshop and to Frauke Kreuter for concluding the session.
(3/🧵)
Noam Slonim, Dan Roth, @andreasvlachos.bsky.social and Ido Dagan.
Our thanks go to
Anna Korhonen for opening the workshop and to Frauke Kreuter for concluding the session.
(3/🧵)
November 17, 2025 at 11:53 AM
The second part of the workshop features extended talks from:
Noam Slonim, Dan Roth, @andreasvlachos.bsky.social and Ido Dagan.
Our thanks go to
Anna Korhonen for opening the workshop and to Frauke Kreuter for concluding the session.
(3/🧵)
Noam Slonim, Dan Roth, @andreasvlachos.bsky.social and Ido Dagan.
Our thanks go to
Anna Korhonen for opening the workshop and to Frauke Kreuter for concluding the session.
(3/🧵)
[...] He highlights the next breakthroughs needed in interpretability to achieve better understanding and control over these models.
November 17, 2025 at 10:22 AM
[...] He highlights the next breakthroughs needed in interpretability to achieve better understanding and control over these models.
Jingcheng (Frank) Niu, current postdoc at UKP co-leading the lab’s interpretability research, presents the group’s contributions and vision for prying open the LLM black box in his talk, “We Owe Language Models an Explanation.” [...]
November 17, 2025 at 10:22 AM
Jingcheng (Frank) Niu, current postdoc at UKP co-leading the lab’s interpretability research, presents the group’s contributions and vision for prying open the LLM black box in his talk, “We Owe Language Models an Explanation.” [...]
[...] In her talk, “Automated Reasoning for Scientific Knowledge Synthesis,” she explores how NLP can resolve contradictory findings and outlines a broader vision for how NLP can advance scientific discovery.
November 17, 2025 at 10:21 AM
[...] In her talk, “Automated Reasoning for Scientific Knowledge Synthesis,” she explores how NLP can resolve contradictory findings and outlines a broader vision for how NLP can advance scientific discovery.
We now welcome @yufanghou.bsky.social, Professor at @ituaustria.bsky.social and visiting professor at UKP, reflecting on her rich collaboration with the UKP Lab. [...]
November 17, 2025 at 10:21 AM
We now welcome @yufanghou.bsky.social, Professor at @ituaustria.bsky.social and visiting professor at UKP, reflecting on her rich collaboration with the UKP Lab. [...]
[...] In her talk, “Human-Centered LLMs: Empathy, Personalization, Social Intelligence,” she explores how we can make sense of unpredictable human data and what it truly means for an AI system to be human-centered.
November 17, 2025 at 10:12 AM
[...] In her talk, “Human-Centered LLMs: Empathy, Personalization, Social Intelligence,” she explores how we can make sense of unpredictable human data and what it truly means for an AI system to be human-centered.
We’re delighted to have Lucie Flek, Professor at the @unibonn.bsky.social and former UKP member, reflecting on her time at UKP and how openness, freedom and passion yield the best results. [...]
November 17, 2025 at 10:12 AM
We’re delighted to have Lucie Flek, Professor at the @unibonn.bsky.social and former UKP member, reflecting on her time at UKP and how openness, freedom and passion yield the best results. [...]
and offers a tutorial on how preference learning can tackle these challenges and the impressive results it enables. His talk is titled “Preferences, Uncertainty, and Factuality in NLP.”
November 17, 2025 at 10:01 AM
and offers a tutorial on how preference learning can tackle these challenges and the impressive results it enables. His talk is titled “Preferences, Uncertainty, and Factuality in NLP.”
We now welcome Edwin Simpson, Senior Lecturer at the @bristoluni.bsky.social and a former member of UKP. He walks through his work on a range of methods that address messy human data ...
November 17, 2025 at 10:01 AM
We now welcome Edwin Simpson, Senior Lecturer at the @bristoluni.bsky.social and a former member of UKP. He walks through his work on a range of methods that address messy human data ...
... argument summarization into the era of generative LLMs, along with fresh perspectives on how to evaluate these systems. His talk: “Argument Summarization and Its Evaluation in the Era of Large Language Models.”
November 17, 2025 at 9:58 AM
... argument summarization into the era of generative LLMs, along with fresh perspectives on how to evaluate these systems. His talk: “Argument Summarization and Its Evaluation in the Era of Large Language Models.”
We’re excited to welcome another UKP alum: Johannes Daxenberger, Co-Founder and Managing Director at summetix (ArgumenText). His recent work challenges traditional classification-based approaches and brings ...
November 17, 2025 at 9:58 AM
We’re excited to welcome another UKP alum: Johannes Daxenberger, Co-Founder and Managing Director at summetix (ArgumenText). His recent work challenges traditional classification-based approaches and brings ...