TechRxiv
2022_Frecon_J_t-pami_suap.pdf (541.15 kB)
Download file

Semi-Universal Adversarial Perturbations

Download (541.15 kB)
preprint
posted on 28.03.2022, 04:33 authored by Jordan FreconJordan Frecon, Gilles Gasso, stephane canu
The present work introduces a framework for learning and selecting semi-universal adversarial perturbations. It relies on a joint estimation of multiple universal adversarial perturbations which are chosen in an unsupervised manner depending on the sample to attack. Two algorithmic solutions, with convergence guarantees under Lipschitz continuity assumptions, are proposed to handle either small scale or large scale datasets. Numerical experiments, conducted on benchmark datasets, support its unifying aspect between universal and specific attacks as the number of perturbations grows. In addition, the learned perturbations display strong patterns indicative of the existing similarities between the training instances of different classes.

Funding

A road toward safe artificial intelligence in mobility – Raimo

Agence Nationale de la Recherche

Find out more...

History

Email Address of Submitting Author

jordan.frecon@gmail.com

Submitting Author's Institution

LITIS, INSA Rouen Normandie

Submitting Author's Country

France

Usage metrics

Licence

Exports