Timezone: »
Over the past few years, neural networks have been proven vulnerable to adversarial images: targeted but imperceptible image perturbations lead to drastically different predictions. We show that adversarial vulnerability increases with the gradients of the training objective when viewed as a function of the inputs. Surprisingly, vulnerability does not depend on network topology: for most current network architectures, we prove that at initialization, the L1-norm of these gradients grows as the square root of the input dimension, leaving the networks increasingly vulnerable with growing image size. We empirically show that this dimension-dependence persists after either usual or robust training, but gets attenuated with higher regularization.
Author Information
Carl-Johann Simon-Gabriel (Max-Planck-Institute for Intelligent Systems)
Yann Ollivier (Facebook Artificial Intelligence Research)
Leon Bottou (Facebook)
Bernhard Schölkopf (MPI for Intelligent Systems Tübingen, Germany)
Bernhard Scholkopf received degrees in mathematics (London) and physics (Tubingen), and a doctorate in computer science from the Technical University Berlin. He has researched at AT&T Bell Labs, at GMD FIRST, Berlin, at the Australian National University, Canberra, and at Microsoft Research Cambridge (UK). In 2001, he was appointed scientific member of the Max Planck Society and director at the MPI for Biological Cybernetics; in 2010 he founded the Max Planck Institute for Intelligent Systems. For further information, see www.kyb.tuebingen.mpg.de/~bs.
David Lopez-Paz (Facebook AI Research)
Related Events (a corresponding poster, oral, or spotlight)
-
2019 Poster: First-Order Adversarial Vulnerability of Neural Networks and Input Dimension »
Wed Jun 12th 01:30 -- 04:00 AM Room Pacific Ballroom
More from the Same Authors
-
2020 Workshop: Inductive Biases, Invariances and Generalization in Reinforcement Learning »
Anirudh Goyal · Rosemary Nan Ke · Stefan Bauer · Jane Wang · Theophane Weber · Fabio Viola · Bernhard Schölkopf · Stefan Bauer -
2020 Workshop: Workshop on Continual Learning »
Haytham Fayek · Arslan Chaudhry · David Lopez-Paz · Eugene Belilovsky · Jonathan Schwarz · Marc Pickett · Rahaf Aljundi · Sayna Ebrahimi · Razvan Pascanu · Puneet Dokania -
2020 Poster: Weakly-Supervised Disentanglement Without Compromises »
Francesco Locatello · Ben Poole · Gunnar Ratsch · Bernhard Schölkopf · Olivier Bachem · Michael Tschannen -
2019 Poster: Robustly Disentangled Causal Mechanisms: Validating Deep Representations for Interventional Robustness »
Raphael Suter · Djordje Miladinovic · Bernhard Schölkopf · Stefan Bauer -
2019 Oral: Robustly Disentangled Causal Mechanisms: Validating Deep Representations for Interventional Robustness »
Raphael Suter · Djordje Miladinovic · Bernhard Schölkopf · Stefan Bauer -
2019 Poster: Kernel Mean Matching for Content Addressability of GANs »
Wittawat Jitkrittum · Wittawat Jitkrittum · Patsorn Sangkloy · Muhammad Waleed Gondal · Amit Raj · James Hays · Bernhard Schölkopf -
2019 Oral: Kernel Mean Matching for Content Addressability of GANs »
Wittawat Jitkrittum · Wittawat Jitkrittum · Patsorn Sangkloy · Patsorn Sangkloy · Muhammad Waleed Gondal · Muhammad Waleed Gondal · Amit Raj · Amit Raj · James Hays · James Hays · Bernhard Schölkopf · Bernhard Schölkopf -
2019 Poster: Manifold Mixup: Better Representations by Interpolating Hidden States »
Vikas Verma · Alex Lamb · Christopher Beckham · Amir Najafi · Ioannis Mitliagkas · David Lopez-Paz · Yoshua Bengio -
2019 Poster: AdaGrad stepsizes: sharp convergence over nonconvex landscapes »
Rachel Ward · Xiaoxia Wu · Leon Bottou -
2019 Poster: White-box vs Black-box: Bayes Optimal Strategies for Membership Inference »
Alexandre Sablayrolles · Douze Matthijs · Cordelia Schmid · Yann Ollivier · Herve Jegou -
2019 Poster: Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations »
Francesco Locatello · Stefan Bauer · Mario Lucic · Gunnar Ratsch · Sylvain Gelly · Bernhard Schölkopf · Olivier Bachem -
2019 Poster: Making Deep Q-learning methods robust to time discretization »
Corentin Tallec · Leonard Blier · Yann Ollivier -
2019 Poster: Separable value functions across time-scales »
Joshua Romoff · Peter Henderson · Ahmed Touati · Yann Ollivier · Joelle Pineau · Emma Brunskill -
2019 Oral: AdaGrad stepsizes: sharp convergence over nonconvex landscapes »
Rachel Ward · Xiaoxia Wu · Leon Bottou -
2019 Oral: White-box vs Black-box: Bayes Optimal Strategies for Membership Inference »
Alexandre Sablayrolles · Douze Matthijs · Cordelia Schmid · Yann Ollivier · Herve Jegou -
2019 Oral: Separable value functions across time-scales »
Joshua Romoff · Peter Henderson · Ahmed Touati · Yann Ollivier · Joelle Pineau · Emma Brunskill -
2019 Oral: Manifold Mixup: Better Representations by Interpolating Hidden States »
Vikas Verma · Alex Lamb · Christopher Beckham · Amir Najafi · Ioannis Mitliagkas · David Lopez-Paz · Yoshua Bengio -
2019 Oral: Making Deep Q-learning methods robust to time discretization »
Corentin Tallec · Leonard Blier · Yann Ollivier -
2019 Oral: Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations »
Francesco Locatello · Stefan Bauer · Mario Lucic · Gunnar Ratsch · Sylvain Gelly · Bernhard Schölkopf · Olivier Bachem -
2018 Poster: Detecting non-causal artifacts in multivariate linear regression models »
Dominik Janzing · Bernhard Schölkopf -
2018 Poster: On Matching Pursuit and Coordinate Descent »
Francesco Locatello · Anant Raj · Sai Praneeth Reddy Karimireddy · Gunnar Ratsch · Bernhard Schölkopf · Sebastian Stich · Martin Jaggi -
2018 Poster: Mixed batches and symmetric discriminators for GAN training »
Thomas LUCAS · Corentin Tallec · Yann Ollivier · Jakob Verbeek -
2018 Oral: Detecting non-causal artifacts in multivariate linear regression models »
Dominik Janzing · Bernhard Schölkopf -
2018 Oral: On Matching Pursuit and Coordinate Descent »
Francesco Locatello · Anant Raj · Sai Praneeth Reddy Karimireddy · Gunnar Ratsch · Bernhard Schölkopf · Sebastian Stich · Martin Jaggi -
2018 Oral: Mixed batches and symmetric discriminators for GAN training »
Thomas LUCAS · Corentin Tallec · Yann Ollivier · Jakob Verbeek -
2018 Poster: Tempered Adversarial Networks »
Mehdi S. M. Sajjadi · Giambattista Parascandolo · Arash Mehrjou · Bernhard Schölkopf -
2018 Poster: Differentially Private Database Release via Kernel Mean Embeddings »
Matej Balog · Ilya Tolstikhin · Bernhard Schölkopf -
2018 Poster: Optimizing the Latent Space of Generative Networks »
Piotr Bojanowski · Armand Joulin · David Lopez-Paz · Arthur Szlam -
2018 Oral: Differentially Private Database Release via Kernel Mean Embeddings »
Matej Balog · Ilya Tolstikhin · Bernhard Schölkopf -
2018 Oral: Tempered Adversarial Networks »
Mehdi S. M. Sajjadi · Giambattista Parascandolo · Arash Mehrjou · Bernhard Schölkopf -
2018 Oral: Optimizing the Latent Space of Generative Networks »
Piotr Bojanowski · Armand Joulin · David Lopez-Paz · Arthur Szlam -
2018 Poster: Learning Independent Causal Mechanisms »
Giambattista Parascandolo · Niki Kilbertus · Mateo Rojas-Carulla · Bernhard Schölkopf -
2018 Oral: Learning Independent Causal Mechanisms »
Giambattista Parascandolo · Niki Kilbertus · Mateo Rojas-Carulla · Bernhard Schölkopf -
2017 Poster: Wasserstein Generative Adversarial Networks »
Martin Arjovsky · Soumith Chintala · Léon Bottou -
2017 Talk: Wasserstein Generative Adversarial Networks »
Martin Arjovsky · Soumith Chintala · Léon Bottou -
2017 Invited Talk: Causal Learning »
Bernhard Schölkopf