ABY3: A Mixed Protocol Framework for Machine Learning

By Payman Mohassel, Peter Rindal. 2018

920 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computational Theory and Mathematics, 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 Computational Theory and Mathematics

In Privacy & Security

Citations in the corpus

Privacy in Deep Learning: A Survey (2020, methodology): “In a more recent work [95], Mohassel et al design protocols for secure three-party training of DNNs with a majority of honest parties.”
Secure Evaluation of Quantized Neural Networks (2019, methodology): “Furthermore, we use the faster comparison proposed by Mohassel and Rindal [50] in the honestmajority setting with rings.”
QUOTIENT: Two-Party Secure Neural Network Training and Prediction (2019, methodology): “In this work, we focus on the twoparty computation setting, which excludes solutions that rely on a majority of honest parties [6, 40].”
Federated Machine Learning (2019, methodology): “rs and semi-honest assumptions. Ref [33] uses MPC protocols for model training and verification without users revealing sensitive data. One of the state-of-the-art SMC framework is Sharemind [8]. Ref [44] proposed a 3PC model [5, 21, 45] with an honest majority and consider security in both semi-hon”
Federated Learning: Challenges, Methods, and Future Directions (2019, background): “s as they incur substantial additional communication and computation costs. Moreover, SMC protocols need to be carefully designed and implemented for each operation in the targeted learning algorithm [25, 79]. We defer interested readers to [13, 97] for a more comprehensive review of the approaches ”
CrypTen: Secure Multi-Party Computation Meets Machine Learning (2021, methodology): “The prospect of such applications of machine learning with rigorous privacy and security guarantees has spurred a number of studies on machine learning via secure MPC [38, 41, 48, 58, 63, 66, 67].”
Private Machine Learning in TensorFlow using Secure Computation (2018, methodology): “The works of Mohassel and Zhang [2017], Juvekar et al. [2018], Wagh et al. [2018], Mohassel and Rindal [2018] focus on adapting secure computation protocols to private machine learning, in some cases using similar optimizations as the protocol presented in Section 2.”
CrypTFlow2: Practical 2-Party Secure Inference (2020, methodology): “to secure matrix multiplication similar to [45, 50, 62].”
SoK: General Purpose Compilers for Secure Multi-Party Computation (2019, methodology): “The EzPC project [30] and ABY(3) project [100], both based on ABY (Section VI-F), attempt to automatically optimize the back-end representation among three protocols.”
An Overview of Federated Deep Learning Privacy Attacks and Defensive Strategies (2020, methodology): “ated learning, this often requires two untrusting servers that connect to some or all the participating clients. Other than two-party computation, three-party communication have been proposed as-well [60]–[63]. These methods rely on at least semi-honest participants or an honest majority. Mohassel e”
XONN: XNOR-based Oblivious Deep Neural Network Inference (2019, background): “meter updates from multiple users. However, none of these approaches [11, 54] study the oblivious inference problem. An overview of related frameworks is provided in [47, 48]. Frameworks such as ABY3 [39] and SecureNN [60] have different computation models and they rely on three (or four) parties du”
Secure SVM Training Over Vertically-Partitioned Datasets Using Consortium Blockchain for Vehicular Social Networks (2020, background): “[19] construct a framework where three servers are necessary to train linear regression, logistic regression and neural network models based on three-party computation.”
DarKnight: A Data Privacy Scheme for Training and Inference of Deep Neural Networks (2020, background): “th signiﬁcant performance penalty [19, 20, 21, 22]. Secure multi-party computing is another approach, where multiple non-colluding servers may use custom data exchange protocols to protect input data [33, 24, 28, 29]. However, this approach requires multiple servers to perform training or inference.”
Falcon: Honest-Majority Maliciously Secure Framework for Private Deep Learning (2020, methodology): “iderable attention over the last few years. Recall that we have summarized some of these works in Table1. Private inference typically relies on one or more of the following techniques: secret sharing [7,8], garbled circuits [9 ,20], homomorphic encryption [6 19] or GoldreichMicali-Wigderson (GMW) [1”
A Scalable Approach for Privacy-Preserving Collaborative Machine Learning (2020, background): “While several more recent works have considered MPC-based learning setups with information-theoretic privacy [37, 27], their constructions are limited to three or four parties.”
New Primitives for Actively-Secure MPC over Rings with Applications to Private Machine Learning (2019, methodology): “For example, this has been done successfully in the ABY framework [15] and subsequent works [16], which convert between arithmetic and binary sharings for applications such as private biometric matching and classiﬁcation using support vector machines, linear/logistic regression and neural networks.”
Cheetah: Optimizing and Accelerating Homomorphic Encryption for Private Inference (2020, background): “, different MPC protocols [23], [41], [44]) and technology (e.”
BatchCrypt: Efficient Homomorphic Encryption for Cross-Silo Federated Learning (2020, background): “Devel-opers have to carefully engineer the ML algorithms and divide the computation among parties to ﬁt the MPC paradigm, which may lower the privacy guarantees for better performance [16,42,43].”