←
Hammer Lab ML (inactive)648 posts
Hammer Lab ML (inactive) avatar

Hammer Lab ML (inactive)

@HammerLabML

The official Twitter account of the #MachineLearning lab led by Prof. Barbara Hammer @unibielefeld from 05/2019 - 01/2025. Inactive.

◷ Joined May 2019▧ 211 Photos and videos
David M. Schmidt avatar
⟳ Hammer Lab ML (inactive) retweeted
David M. Schmidt@davidmschmidtde·7 Nov 2024

🚨 Thrilled to announce a new preprint! 📝 Title: Lexicalization Is All You Need: Examining the Impact of Lexical Knowledge in a Compositional QALD System 👩‍🎓👨‍🎓 Authors: David Maria Schmidt, Mohammad Fazleh Elahi, Philipp Cimiano @pcimiano 🔗 Preprint: arxiv.org/abs/2411.03906

Hammer Lab ML (inactive) avatar
Hammer Lab ML (inactive)@HammerLabML·26 Sep 2024

🚀Barbara Hammer enjoyed great discussions at today’s follow-up breakfast from the Research & Innovation Summit @stifterverband on speeding up AI in 🇩🇪&🇪🇺: 💡Unified AI strategies 💡Legal clarity 💡Faster innovation 💡Investment in AI skills linkedin.com/feed/update/urn…

Dr. Ulrike Kuhl avatar
⟳ Hammer Lab ML (inactive) retweeted
Dr. Ulrike Kuhl@DrUlrikeKuhl·9 Aug 2024

Last day @IJCAIconf, and we‘re starting strong with #XAI! Keynote by @fra_toni that bridges 🌉 explainability 🗣️ and contestability 🙅‍♀️, advocating for #XAI that allows humans to challenge and change the AI by computational 💻 argumentation.

Post attachment
Tim Miller avatar
⟳ Hammer Lab ML (inactive) retweeted
Tim Miller@tmiller_uq·5 Aug 2024

Andre Artelt applied DataSHAP to investigate the effect of training data samples on explanation. A paper with lovely theory and a nice case study using this to investigate group fairness. Interesting results.

Post attachment
Hammer Lab ML (inactive) avatar
Hammer Lab ML (inactive)@HammerLabML·5 Aug 2024

#XAI workshop @IJCAIconf: André Artelt showing how training samples impact explanations: ➡️suggesting a quantification of explanation importance ➡️how to find the most relevant samples for a explanation ➡️showing how removal of identified samples may reduce unfairness

Post attachment
Hammer Lab ML (inactive) avatar
Hammer Lab ML (inactive)@HammerLabML·3 Aug 2024

It’s a wrap on the tutorial on #counterfactuals and user studies for #XAI @IJCAIconf. Anyone who’s curious: The materials (slides and code) are available @github: github.com/andreArtelt/IJCAI… For more info about the general program, visit sites.google.com/view/tut-co…

@IJCAIconf, the #XAI tutorial by our own André Artelt and @DrUlrikeKuhl, together with @keanema, is in session! It started with a history of counterfactuals in psychology and first basics of user studies, followed by computational foundations of computing counterfactuals.

Post attachmentPost attachment
Dr. Ulrike Kuhl avatar
⟳ Hammer Lab ML (inactive) retweeted
Dr. Ulrike Kuhl@DrUlrikeKuhl·3 Aug 2024

I‘ve had a great time presenting the basics of how to run effective user studies to explore the usability of XAI! Anyone who’s curious: The materials (slides and code) are available @github: github.com/andreArtelt/IJCAI… ✨Thanks to everyone attending this full day program! ✨