New video presentation available on YouTube! You can now watch Gregory Polyakov, the first author of "Interpretability Analysis of Arithmetic In-Context Learning in Large Language Models", explain their research question, methodology, and results.
You can watch the video here:
New video alert! The video presentation of the paper "A Survey on LLM-Assisted Clinical Trial Recruitment" by Dr. Shrestha Ghosh is now live on YouTube.
🎞️ Check out the video here: www.youtube.com/watch?v=fE_3...
And don't forget to subscribe to our YouTube channel for more videos!
With this milestone achieved, she will continue her research as a postdoctoral researcher starting in March 2026.
You can explore her body of work here: dblp.org/pid/322/7058...
Join us in celebrating Dr. Michal Golovanevsky for her persistence, creativity, and contributions to the field of CS👏
Michal has a rich portfolio of publications across top venues, and her most recent paper, "Mechanisms of Prompt-Induced Hallucination in Vision-Language Models", currently under review at ACL 2026, continues this trajectory of high-impact research.
In her PhD years, she focused on understanding the internal workings of VLMs, with particular attention to scalability, interpretability, and control. Her work blends rigorous theory with practical insights, pushing forward how attention mechanisms are designed and understood in multimodal contexts.
Heartfelt congratulations to Dr. Michal Golovanevsky! 🎉
On January 30, 2026, our PhD student at Brown University successfully defended her thesis, "Advancing Attention Mechanisms in Multimodal Deep Learning Models", marking the culmination of years of research excellence and intellectual growth.
Kay highlighted what it really means to translate research into clinical practice, outlined common regulatory pathways, and shared typical mistakes teams make when compliance is treated as an afterthought. The Q&A that followed showed how relevant these challenges are to many projects.
Prof. Carsten Eickhoff introducing Kay Brosien from x-cardiac GmbH.
Thank you to everyone who joined our invited talk last Friday. We were very happy to welcome Kay Brosien from x-cardiac GmbH and to learn from his practical insights on AI regulation in medicine.
Reminder for the talk tomorrow: see you at the Hörsaal of Maria-von-Linden-Straße 6 at 11:00 A.M.
✒️ Gregory Polyakov, Catherine Chen, Carsten Eickhoff
📃 dl.acm.org/doi/10.1145/...
</> github.com/polgrisha/be...
In “Towards Best Practices of Axiomatic Activation Patching in Information Retrieval”, we analyze these pitfalls and propose concrete best practices to make activation patching a more reliable diagnostic tool for neural rankers, paving the way toward more interpretable and trustworthy IR models.
However, applying activation patching in IR is far from straightforward: dataset construction choices, term rareness, small score differences, and other experimental factors can strongly bias the results and lead to misleading conclusions.
Like many areas of machine learning, information retrieval has increasingly adopted large neural models, making mechanistic interpretability more important than ever. A key technique in this space is activation patching, which aims to localize where and how models encode relevance signals.
The talk concludes with an open discussion and micro-consulting: bring your specific blockers, “is this even legal?” questions, or grant application drafts.
📅Friday, February 13th, 2026
📍Maria-von-Linden-Straße 6, Tübingen – Hörsaal
If you’re working on AI in the medical field, don't miss this talk!
The session will also dive into the three classic mistakes teams make along the way: ignoring regulation until it’s too late, treating consultants like plumbers, and buying expensive tools for problems a spreadsheet could solve.
The talk outlines three main paths researchers typically face: Spin-offs, Article 5.5, In-House Exemptions, and research grants, and explains why regulatory systems are often fundamentally misaligned with scientific innovation.
Whether you’re answering the dreaded “How do you intend to translate this?” in a grant application, planning a spin-off, or trying to use your tool “in-house” at a clinic, at some point, you will hit the regulatory wall.
Writing excellent code is one thing. Getting it to a patient without going to jail is another. In this 20-minute talk, Kay will share a practical perspective on translating AI research into clinical practice.
Next Friday, we’re looking forward to welcoming Kay Brosien, COO of x-cardiac GmbH, to Tübingen for a talk on one of the less glamorous but absolutely critical parts of AI research in medicine: regulation⚖️
@twiml.bsky.social are organizing their 4th workshop to bring together ML researchers and enthusiasts in Tübingen. It’s a wonderful opportunity to meet peers, exchange ideas, discuss research, and share your academic journey. Follow their page and stay tuned for more details about the workshop.
🚀 #JobAlert!
Permanent Administrative Staff Member position open at the Tübingen AI Center, a vibrant AI research hub at @unituebingen.bsky.social.
You’ll support our Central Office in financial and HR administration and help run everyday office operations.
🔗 tuebingen.ai/careers/admi...
Did you know...? 🤔 Health NLP Lab also has a YouTube channel!
We are reviving our YouTube channel as a platform for paper presentations and science communication. We hope this will make our research more accessible to a broader audience.
🔗 Subscribe here: www.youtube.com/@health-nlp
Visit her website (sites.google.com/view/miriam-...) or follow her on X (x.com/miriamrateike) to learn more about her work.
Miriam's research focuses on #trustworthiness in machine learning, particularly #fairness and #interpretability, with a growing emphasis on challenges emerging in the era of large language models.
Miriam holds a Master's degree in Industrial Engineering and Management from the Karlsruhe Institute of Technology (KIT). She is currently based in Nairobi, where she works at IBM Research Africa, while pursuing her PhD at the University of Tübingen and her law studies at FernUniversität in Hagen.
Meet Miriam Rateike, and join us in welcoming Health NLP Lab's newest member!
Maik is currently conducting research at the Chair for Databases and Information Systems, part of the Webis Group (@webis.de). His research interests include Information Retrieval, Machine Learning, and Big Data. Learn more about his work and publications here: maik-froebe.de
Last week, we had the pleasure of hosting Maik Fröbe @maik-froebe.bsky.social, from Friedrich Schiller University Jena, who was invited to give a guest lecture and a talk to our group about his recent work on evaluation in information retrieval.
Catherine is in the final phase of her PhD and is seeking opportunities for the next stage of her research career. Visit her website to learn more about her background and research interests: catherineschen.github.io
2. Axiomatic Causal Interventions for Reverse Engineering Relevance Computation in Neural Retrieval Models
✒️ Catherine Chen, Jack Merullo, and Carsten Eickhoff
📃 health-nlp.com/files/pubs/s...
🎥 www.youtube.com/watch?v=gbjX...
