SecureML: A System for Scalable Privacy-Preserving Machine Learning

By Payman Mohassel, Yupeng Zhang. 2017

2,156 citations recorded by the source corpus.

Subfields: Artificial Intelligence, ML Systems & Hardware, Privacy & Security

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In ML Systems & Hardware

In Privacy & Security

Citations in the corpus

Advances and Open Problems in Federated Learning (2019, background): “Similar protections can be achieved under the MPC model [302, 14].”
How To Backdoor Federated Learning (2018, background): “Secure multi-party computation can help train models while protecting privacy of the training data [30], but it does not protect model integrity.”
Deep Models Under the GAN: Information Leakage from Collaborative Deep Learning (2017, background): “The work [59] provides a solution for training neural networks while preserving the privacy of the participants.”
Privacy in Deep Learning: A Survey (2020, methodology): “SecureML uses secret sharing to achieve privacy during training.”
Secure Evaluation of Quantized Neural Networks (2019, methodology): “y in some cases [62, 64]. We consider this an orthogonal work to secure inference using MPC techniques. Research in the area of secure evaluation of CNNs has been rich during the last couple of years [24, 55, 56, 47, 50, 38, 63, 54, 44]. The main goal of this prior research has been to reduce the pe”
Learning Privately over Distributed Features: An ADMM Sharing Approach (2019, background): “Another approach to privacy-preserving machine learning is through encryption Gilad-Bachrach et al. (2016); Takabi et al. (2016); Kikuchi et al. (2018) or secret sharing Mohassel and Zhang (2017); Wan et al. (2007); Bonte and Vercauteren (2018), so that models are trained on encrypted data.”
QUOTIENT: Two-Party Secure Neural Network Training and Prediction (2019, neutral): “The work most similar to ours is SecureML [41].”
Federated Machine Learning (2019, methodology): “may be considered acceptable if security guarantees are provided. It is possible to build a security model with SMC under lower security requirement in exchange for efficiency [16]. Recently, studies [46] used SMC framework for training machine learning models with two servers and semi-honest assump”
SecureBoost: A Lossless Federated Learning Framework (2019, background): “In the past, several attempts have been made to address the user-privacy problem while exchanging data (Hardy et al. 2017; Mohassel and Zhang 2017).”
Slalom: Fast, Verifiable and Private Execution of Neural Networks in Trusted Hardware (2018, methodology): “, in the latter two where the computation is performed locally by the data owner), the need for integrity is a common denominator, and this is the property we primarily focus on in this paper (and for which we obtain the best results!) Cryptography can address some of these problems [17, 32, 3, 38, ”
Contamination Attacks and Mitigation in Multi-Party Machine Learning (2018, neutral): “Secure MPC can be enabled with cryptographic techniques [6, 11, 18, 29, 32], and systems based on trusted processors such as Intel SGX [3, 5, 33].”
Biscotti: A Ledger for Private and Secure Peer-to-Peer Machine Learning (2018, methodology): “Another solution uses two party computation [36] and encrypts both the model parameters and the training data when performing multi-party ML.”
DeepSecure: Scalable Provably-Secure Deep Learning (2017, methodology): “SecureML [6] is another secure DL framework that proposes to use additive secret-sharing and the GC protocol.”
Machine Learning with Membership Privacy using Adversarial Regularization (2018, methodology): “ganduseofmachinelearningmodels. These methods leverage various techniques including homomorphicencryption, garbledcircuits,and securemulti-partycomputationfor privatemachine learning onencrypted data [8, 20, 33, 37], as well as private computation using trusted hardware (e.g., Intel SGX)[26,39].Alth”
CryptoDL: Deep Neural Networks over Encrypted Data (2017, methodology): “In recent years, several studies have investigated the privacy protection of this sensitive data in different machine learning algorithms such as linear regression [39], linear classifiers [5, 13], decision trees [5, 44] or neural networks [9, 49].”