skip to content
Michael Aerni
About
Now
Publications
Publications
Preprints
2026
Large-scale online deanonymization with LLMs
S. Lermen*, D. Paleka*, J. Swanson,
M. Aerni
, N. Carlini, F. Tramèr
blog
pdf
tweet
press
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
Conferences
ICLR 2026
Modal Aphasia: Can Unified Multimodal Models Describe Images From Memory?
M. Aerni*
, J. Swanson*, K. Nikolić, F. Tramèr
blog
pdf
code
tweet
ICLR 2025
Measuring Non-Adversarial Reproduction of Training Data in Large Language Models
M. Aerni*
, J. Rando*, E. Debenedetti, N. Carlini, D. Ippolito, F. Tramèr
blog
pdf
code
tweet
CCS 2024
Evaluations of Machine Learning Privacy Defenses are Misleading
M. Aerni*
, J. Zhang*, F. Tramèr
blog
pdf
code
tweet
doi
ICLR 2023
Strong inductive biases provably prevent harmless interpolation
M. Aerni*
, M. Milanta*, K. Donhauser, F. Yang
pdf
code
NeurIPS 2021
Interpolation can hurt robust generalization even when there is no noise
K. Donhauser*, A. Țifrea*,
M. Aerni
, R. Heckel, F. Yang
pdf
code
poster
Workshops
DLSP 2025
Membership Inference Attacks on Sequence Models
L. Rossi,
M. Aerni
, J. Zhang, F. Tramèr
pdf
ICML OPPO 2021
Surprising benefits of ridge regularization for noiseless regression
K. Donhauser*, A. Țifrea*,
M. Aerni
, R. Heckel, F. Yang
pdf
code
ICML AML 2021
Maximizing the robust margin provably overfits on noiseless data
K. Donhauser*, A. Țifrea*,
M. Aerni
, R. Heckel, F. Yang
pdf
code