The reliability of deep learning algorithms is fundamentally challenged by the existence of adversarial examples, which are incorrectly classified inputs that are extremely close to a correctly classified input. We explore the properties of adversarial examples for deep neural networks with random weights and biases, and prove that for any p≥1, the ell^p distance of any given input from the classification boundary scales as one over the square root of the dimension of the input times the ell^p norm of the input. The results are based on the recently proved equivalence between Gaussian processes and deep neural networks in the limit of infinite width of the hidden layers, and are validated with experiments on both random deep neural networks and deep neural networks trained on the MNIST and CIFAR10 datasets. The results constitute a fundamental advance in the theoretical understanding of adversarial examples, and open the way to a thorough theoretical characterization of the relation between network architecture and robustness to adversarial perturbations.

Adversarial robustness guarantees for Random Deep Neural Networks

Giacomo De Palma
;
2021

Abstract

The reliability of deep learning algorithms is fundamentally challenged by the existence of adversarial examples, which are incorrectly classified inputs that are extremely close to a correctly classified input. We explore the properties of adversarial examples for deep neural networks with random weights and biases, and prove that for any p≥1, the ell^p distance of any given input from the classification boundary scales as one over the square root of the dimension of the input times the ell^p norm of the input. The results are based on the recently proved equivalence between Gaussian processes and deep neural networks in the limit of infinite width of the hidden layers, and are validated with experiments on both random deep neural networks and deep neural networks trained on the MNIST and CIFAR10 datasets. The results constitute a fundamental advance in the theoretical understanding of adversarial examples, and open the way to a thorough theoretical characterization of the relation between network architecture and robustness to adversarial perturbations.
Settore MAT/07 - Fisica Matematica
38th International Conference on Machine Learning
Proceedings of the 38th International Conference on Machine Learning
http://proceedings.mlr.press/v139/de-palma21a.html
File in questo prodotto:
File Dimensione Formato  
Adversarial Robustness Guarantees for Random Deep Neural Networks.pdf

accesso aperto

Tipologia: Published version
Licenza: Accesso gratuito (sola lettura)
Dimensione 952.7 kB
Formato Adobe PDF
952.7 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11384/107629
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? 0
social impact