Michael Aerni

3 papers · 2021–2025 · 2 conferences · across top CS/AI conferences

Achievements

🌍 Conference Polyglot (2) 🧭 Keyword Pioneer 🐝 Cross-Pollinator (7)

Conferences

ICLR (2) NIPS (1)

Top co-authors

Konstantin Donhauser (2) Fanny Yang (2) Florian Tramer (1) Daphne Ippolito (1) Nicholas Carlini (1) Edoardo Debenedetti (1) Marco Milanta (1) Javier Rando (1) Reinhard Heckel (1) Alexandru Tifrea (1)

Keywords

robust overfitting (1) robust generalization (1) ridge regularization (1) max-margin classifier (1) min-norm interpolator (1)

Papers

Measuring Non-Adversarial Reproduction of Training Data in Large Language Models ICLR 2025

Strong inductive biases provably prevent harmless interpolation ICLR 2023

Interpolation can hurt robust generalization even when there is no noise NIPS 2021