Civile.it
/internet
Osservatorio sul diritto e telecomunicazioni informatiche, a cura del dott. V. Spataro dal 1999, documenti.

Il dizionario e' stato letto volte. Scarica l'app da o da



Segui via: Email - Telegram
  Dal 1999   spieghiamo il diritto di internet  Store  Caffe20  Dizionario  App  Video Demo · Accesso · Iscrizioni    
             

Abbonati:




Scarica il dizionario:


Per cancellarsi dalla precedente newsletter interna

  


Deep Learning 10.12.2018    Pdf    Appunta    Letti    Post successivo  

AI e deep learning: interferenze nell'apprendimento neuronale

Comprehensive Privacy Analysis of Deep Learning:
Stand-alone and Federated Learning under Passive and Active White-box Inference Attacks

 

&

 

Deep neural networks are susceptible to various inference attacks as they remember information about their training data. We perform a comprehensive analysis of white-box privacy inference attacks on deep learning models. We measure the privacy leakage by leveraging the final model parameters as well as the parameter updates during the training and fine-tuning processes. We design the attacks in the stand-alone and federated settings, with respect to passive and active inference attackers, and assuming different adversary prior knowledge. 
We design and evaluate our novel white-box membership inference attacks against deep learning algorithms to measure their training data membership leakage. We show that a straightforward extension of the known black-box attacks to the white-box setting (through analyzing the outputs of activation functions) is ineffective. We therefore design new algorithms tailored to the white-box setting by exploiting the privacy vulnerabilities of the stochastic gradient descent algorithm, widely used to train deep neural networks. We show that even well-generalized models are significantly susceptible to white-box membership inference attacks, by analyzing state-of-the-art pre-trained and publicly available models for the CIFAR dataset. We also show how adversarial participants of a federated learning setting can run active membership inference attacks against other participants, even when the global model achieves high prediction accuracies.
Subjects: Machine Learning (stat.ML); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
Cite as: arXiv:1812.00910 [stat.ML]
  (or arXiv:1812.00910v1 [stat.ML] for this version)

10.12.2018 Spataro
Fonte: spataro

Download Pdf



Segui le novità in materia di Deep Learning su Civile.it via Telegram
oppure via email: (gratis Info privacy)



Dossier:



dallo store:
visita lo store

Altro su Deep Learning:


"Il successo non mai definitivo, il fallimento non mai fatale; il coraggio di continuare che conta." - Churchill


Spieghiamo contratti, ecommerce, privacy e il diritto di internet su Civile.it dal 1999



Tinyletter - I testi sono degli autori e di IusOnDemand srl p.iva 04446030969 - diritti riservati - Privacy - Cookie - Condizioni d'uso - in