pith. sign in

Deepfool: a simple and accurate method to fool deep neural networks

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

citation-role summary

background 1

citation-polarity summary

fields

stat.ML 1

years

2017 1

verdicts

ACCEPT 1

roles

background 1

polarities

background 1

representative citing papers

Towards Deep Learning Models Resistant to Adversarial Attacks

stat.ML · 2017-06-19 · accept · novelty 7.0

Adversarial training via projected gradient descent on the inner maximization problem produces neural networks with substantially improved resistance to a wide range of attacks and establishes security against first-order adversaries as a concrete guarantee.

citing papers explorer

Showing 1 of 1 citing paper.

  • Towards Deep Learning Models Resistant to Adversarial Attacks stat.ML · 2017-06-19 · accept · none · ref 21

    Adversarial training via projected gradient descent on the inner maximization problem produces neural networks with substantially improved resistance to a wide range of attacks and establishes security against first-order adversaries as a concrete guarantee.