Here is a formal impossibility result for XAI: Informative Post-Hoc Explanations Only Exist for Simple Functions. I'll give an online presentation about this work next tuesday in @timvanerven.nl 's Theory of Interpretable AI Seminar: arxiv.org/abs/2508.11441 tverven.github.io/tiai-seminar/

Working on the interface of Machine Learning with Law, Philosophy, Ethics, or Social Science? Come join our new Lab for Norms and Practices in Machine Learning for Science. uni-tuebingen.de/en/128980#c2

Tübingen Conference for AI and Law starting on wednesday! Keynotes by Rediet Abebe, Solon Borocas, Sylvie Delacroix, Lilian Edwards, Christoph Engel, Michal Gal, Philipp Hacker, Christoph Kern, Christoph Sorge. ailawinstitute.de/conference-f...

Looking forward to the Winter School on Causality and Explainable AI in Paris next week! xai-winter-school.github.io

Time to figure out which provable guarantees one can(not) give on XAI! Workshop "Theory of Explainable Machine Learning", Dec 2 in Copenhagen as part of the Ellis Unconference/EurIPS. Submission deadline: Oct 15. sites.google.com/view/theory-... eurips.cc/ellis/

I am hiring PhD students and/or Postdocs, to work on the theory of explainable machine learning. Please apply through Ellis or IMPRS, deadlines end october/mid november. In particular: Women, where are you? Our community needs you!!! imprs.is.mpg.de/application ellis.eu/news/ellis-p...

Our Machine Learning for Science Cluster in Tübingen @ml4science.bsky.social is extending its collaboration with the African Institute for Mathematical Sciences #AIMS: We hire a joint research group leader, based in Kigali. Application deadline: August 31. tinyurl.com/38z4e6dy

I love the #eurips initiative! But to live up to its potential, it should be accepted an official #neurips conference location (similar to Mexico City) and not just an addon! Then we would save C02, rather than adding to it!!! What can we do to achieve this?

How to disentangle feature importance scores into contributions of individual features, feature interactions in the function, and feature dependencies in the data: #AISTATS with @gunnark und Eric Guenther: Paper: arxiv.org/pdf/2410.23772 Code: github.com/gcskoenig/dipd Video: youtu.be/7MrMjabTbuM

Our #ICML position paper: #XAI is similar to applied statistics: it uses summary statistics in an attempt to answer real world questions. But authors need to state how concretely (!) their XAI statistics contributes to answer which concrete (!) question! arxiv.org/abs/2402.02870

Cool #ICML paper by my postdoc @sbordt.bsky.social . Very careful experiments to find out how often you can contaminate training data yet contents are being forgotten - or not, depending on the size of your model. arxiv.org/abs/2410.03249

One more day to the deadline for YOUR contribution to the Tübingen AI and Law Conference in November! ailawinstitute.de/conference-f...

Super-happy that @mariokrenn.bsky.social is joining our Machine Learning for Science cluster @ml4science.bsky.social‬ in Tübingen as a Full Professor of "Machine Learning for Science" next week. Welcome Mario!

Tübingen Conference for AI and Law, Nov 5-6, 2005, abstract deadline: June 3! Keynote presentations by Rediet Abebe, Solon Barocas, Sylvie Delacroix, Christoph Engel, Michal Gal, Philipp Hacker, Christoph Kern, Christoph Sorge. Please repost :-) ailawinstitute.de/conference-f...

Our cluster Machine Learning for Science is up for 7 years more funding!

Want to present at the Tübingen Conference for AI and Law? Abstract deadline is June 3! We target an academic audience that is willing to think beyond the boundaries of their own discipline. Also check out our international keynote speakers: ailawinstitute.de/conference-f...

Ever aggregated SHAP values across sample points? Our #COLT2025 paper proves that this might be safe when your goal is to discard unimportant features - but only if you add one extra line of code that reshuffles your data! With Robi Bhattacharjee and Karolin Frohnapfel arxiv.org/abs/2503.23111

Finally made it to bluesky as well ...