Lukas Struppek (@lukasstruppek) 's Twitter Profile
Lukas Struppek

@lukasstruppek

Researcher (PhD) working on Trustworthy Generative AI | @TUDarmstadt | @DFKI | @KITKarlsruhe | ✨ Open for Researcher Positions ✨

ID: 1450123005150076931

linkhttps://lukasstruppek.github.io/ calendar_today18-10-2021 15:34:32

129 Tweet

189 Followers

218 Following

Lukas Struppek (@lukasstruppek) 's Twitter Profile Photo

πŸ“’Check out our @ecai2024 paper, where we leverage backdoor attacks from adversarial ML research to erase private information from multi-modal models like CLIP and Stable Diffusion and protect our privacy. πŸ‘‡

Kristian Kersting (@kerstingaiml) 's Twitter Profile Photo

πŸŽ‰ Our paper has been accepted at "AI and Ethics" πŸŽ‰ We present FAIR DIFFUSION as a strategy to attenuate biases during the deployment of generative text-to-image models. πŸ™ felfri Dominik Hintersdorf Lukas Struppek Manuel Brack Patrick Schramowski Sasha Luccioni, PhD πŸ¦‹πŸŒŽβœ¨πŸ€— πŸ‘‰ link.springer.com/article/10.100…

πŸŽ‰ Our paper has been accepted at "AI and Ethics" πŸŽ‰ We present FAIR DIFFUSION as a strategy to attenuate biases during the deployment of generative text-to-image models. πŸ™ <a href="/felix_friedri/">felfri</a> <a href="/d_hintersdorf/">Dominik Hintersdorf</a> <a href="/LukasStruppek/">Lukas Struppek</a> <a href="/MBrack_AIML/">Manuel Brack</a> <a href="/schrame90/">Patrick Schramowski</a> <a href="/SashaMTL/">Sasha Luccioni, PhD πŸ¦‹πŸŒŽβœ¨πŸ€—</a>  πŸ‘‰ link.springer.com/article/10.100…
Kristian Kersting (@kerstingaiml) 's Twitter Profile Photo

Super proud of my team πŸ’ 5 cutting-edge #NeurIPS 2024 papers will showcase important facets of our "Reasonable AI" (RAI) agenda like #genAI, #GNNs, #neurosymbolic #AI, and #memorization πŸ™ Teamwork makes the dream work! πŸ™‡ to everyone

Super proud of my team πŸ’  5 cutting-edge #NeurIPS 2024 papers will showcase important facets of our "Reasonable AI" (RAI) agenda like #genAI, #GNNs, #neurosymbolic #AI, and #memorization πŸ™ Teamwork makes the dream work! πŸ™‡ to everyone
Lukas Struppek (@lukasstruppek) 's Twitter Profile Photo

Check out our #NeurIPS2024 paper "Finding NeMo: Localizing Neurons Responsible For Memorization in Diffusion Models" ⏰ Wed 11 Dec | 4:30 pm πŸ“East Hall A-C #2808 πŸ–₯️ arxiv.org/abs/2406.02366 With Dominik Hintersdorf Kristian Kersting @fraboeni Adam Dziedzic Happy to have a chat πŸ˜ƒ

Check out our #NeurIPS2024 paper "Finding NeMo: Localizing Neurons Responsible For Memorization in Diffusion Models"

⏰ Wed 11 Dec | 4:30 pm
πŸ“East Hall A-C #2808
πŸ–₯️ arxiv.org/abs/2406.02366

With <a href="/d_hintersdorf/">Dominik Hintersdorf</a> <a href="/kerstingAIML/">Kristian Kersting</a> <a href="/fraboeni/">@fraboeni</a> <a href="/adam_dziedzic/">Adam Dziedzic</a>

Happy to have a chat πŸ˜ƒ
Lukas Struppek (@lukasstruppek) 's Twitter Profile Photo

πŸ“’We have a new pre-print available on the connection between shortcut learning, spurious correlations, and confounders β€” phenomenons that also relate to adversarial machine learning concepts, such as data poisoning.

Lukas Struppek (@lukasstruppek) 's Twitter Profile Photo

πŸ“’ Happy to announce that our workshop "The Impact of Memorization on Trustworthy Foundation Models" has been accepted at #icml2025 (ICML Conference ). Stay tuned for the CfP!

Lukas Struppek (@lukasstruppek) 's Twitter Profile Photo

Don't miss our ICML Conference Workshop on the Impact of Memorization on Foundation Models and its great bunch of speakers from industry and academia. ⏰08:25 πŸ“West Meeting Room 223/224 🌎 icml2025memfm.github.io