Homomorphic encryption (HE)-based deep neural network (DNN) inference protects data and model privacy but suffers from significant computation overhead. We observe transforming the DNN weights into circulant matrices converts general matrix-vector multiplications into HE-friendly 1-dimensional convolutions, drastically reducing the HE computation cost. Hence, in this paper, we propose PrivCirNet, a protocol/network co-optimization framework based on block circulant transformation. At the...

Fixed point arithmetic (FPA) is essential to enable practical Privacy-Preserving Machine Learning. When multiplying two fixed-point numbers, truncation is required to ensure that the product maintains correct precision. While multiple truncation schemes based on Secure Multiparty Computation (MPC) have been proposed, which of the different schemes offers the best trade-off between accuracy and efficiency on common PPML datasets and models has remained underexplored. In this work, we...

Fully Homomorphic Encryption (FHE) enables privacy-preserving computation but imposes significant computational and communication overhead on the client for the public-key encryption. To alleviate this burden, previous works have introduced the Hybrid Homomorphic Encryption (HHE) paradigm, which combines symmetric encryption with homomorphic decryption to enhance performance for the FHE client. While early HHE schemes focused on binary data, modern versions now support integer prime fields,...

As privacy concerns have arisen in machine learning, privacy-preserving machine learning (PPML) has received significant attention. Fully homomorphic encryption (FHE) and secure multi-party computation (MPC) are representative building blocks for PPML. However, in PPML protocols based on FHE and MPC, interaction between the client (who provides encrypted input data) and the evaluator (who performs the computation) is essential to obtain the final result in plaintext. Functional encryption...

\textit{Federated Learning} (FL) is a distributed machine learning paradigm that allows multiple clients to train models collaboratively without sharing local data. Numerous works have explored security and privacy protection in FL, as well as its integration with blockchain technology. However, existing FL works still face critical issues. \romannumeral1) It is difficult to achieving \textit{poisoning robustness} and \textit{data privacy} while ensuring high \textit{model accuracy}....

We present $\textit{Rhombus}$, a new secure matrix-vector multiplication (MVM) protocol in the semi-honest two-party setting, which is able to be seamlessly integrated into existing privacy-preserving machine learning (PPML) frameworks and serve as the basis of secure computation in linear layers. $\textit{Rhombus}$ adopts RLWE-based homomorphic encryption (HE) with coefficient encoding, which allows messages to be chosen from not only a field $\mathbb{F}_p$ but also a ring...

We propose an efficient non-interactive privacy-preserving Transformer inference architecture called Powerformer. Since softmax is a non-algebraic operation, previous studies have attempted to modify it to be HE-friendly, but these methods have encountered issues with accuracy degradation or prolonged execution times due to the use of multiple bootstrappings. We propose replacing softmax with a new ReLU-based function called the \textit{Batch Rectifier-Power max} (BRPmax) function without...

Privacy-Preserving Machine Learning (PPML) is one of the most relevant use cases for Secure Multiparty Computation (MPC). While private training of large neural networks such as VGG-16 or ResNet-50 on state-of-the-art datasets such as ImageNet is still out of reach due to the performance overhead of MPC, GPU-based MPC frameworks are starting to achieve practical runtimes for private inference. However, we show that, in contrast to plaintext machine learning, the usage of GPU acceleration for...

Homomorphic Encryption (HE) is a cutting-edge cryptographic technique that enables computations on encrypted data to be mirrored on the original data. This has quickly attracted substantial interest from the research community due to its extensive practical applications, such as in cloud computing and privacy-preserving machine learning. In addition to confidentiality, the importance of authenticity has emerged to ensure data integrity during transmission and evaluation. To address...

Privacy-preserving machine learning (PPML) enables multiple distrusting parties to jointly train ML models on their private data without revealing any information beyond the final trained models. In this work, we study the client-aided two-server setting where two non-colluding servers jointly train an ML model on the data held by a large number of clients. By involving the clients in the training process, we develop efficient protocols for training algorithms including linear regression,...

Deduplication is a vital preprocessing step that enhances machine learning model performance and saves training time and energy. However, enhancing federated learning through deduplication poses challenges, especially regarding scalability and potential privacy violations if deduplication involves sharing all clients’ data. In this paper, we address the problem of deduplication in a federated setup by introducing a pioneering protocol, Efficient Privacy-Preserving Multi-Party Deduplication...

Model quantization has become a common practice in machine learning (ML) to improve efficiency and reduce computational/communicational overhead. However, adopting quantization in privacy-preserving machine learning (PPML) remains challenging due to the complex internal structure of quantized operators, which leads to inefficient protocols under the existing PPML frameworks. In this work, we propose a new PPML paradigm that is tailor-made for and can benefit from quantized models. Our...

With concerns about data privacy growing in a connected world, cryptography researchers have focused on fully homomorphic encryption (FHE) for promising machine learning as a service solutions. Recent advancements have lowered the computational cost by several orders of magnitude, but the latency of fully homomorphic neural networks remains a barrier to adoption. This work proposes using multi-exit neural networks (MENNs) to accelerate the FHE inference. MENNs are network architectures that...

Decision trees are an important class of supervised learning algorithms. When multiple entities contribute data to train a decision tree (e.g. for fraud detection in the financial sector), data privacy concerns necessitate the use of a privacy-enhancing technology such as secure multi-party computation (MPC) in order to secure the underlying training data. Prior state-of-the-art (Hamada et al.) construct an MPC protocol for decision tree training with a communication of $\mathcal{O}(hmN\log...

We present a variant of Function Secret Sharing (FSS) schemes tailored for point, comparison, and interval functions, featuring compact key sizes at the expense of additional comparison. While existing FSS constructions are primarily geared towards $2$-party scenarios, exceptions such as the work by Boyle et al. (Eurocrypt 2015) and Riposte (S&P 2015) have introduced FSS schemes for $p$-party scenarios ($p \geq 3$). This paper aims to achieve the most compact $p$-party FSS key size to date....

Machine learning is widely used for a range of applications and is increasingly offered as a service by major technology companies. However, the required massive data collection raises privacy concerns during both training and inference. Privacy-preserving machine learning aims to solve this problem. In this setting, a collection of servers secret share their data and use secure multi-party computation to train and evaluate models on the joint data. All prior work focused on the scenario...

Privacy is a major concern in large-scale digital applications, such as cloud-computing, machine learning services, and access control. Users want to protect not only their plain data but also their associated attributes (e.g., age, location, etc). Functional encryption (FE) is a cryptographic tool that allows fine-grained access control over encrypted data. However, existing FE fall short as they are either inefficient and far from reality or they leak sensitive user-specific...

Secure multi-party computation (MPC) in a three-party, honest majority scenario is currently the state-of-the-art for running machine learning algorithms in a privacy-preserving manner. For efficiency reasons, fixed-point arithmetic is widely used to approximate computation over decimal numbers. After multiplication in fixed-point arithmetic, truncation is required to keep the result's precision. In this paper, we present an efficient three-party truncation protocol secure in the presence of...

Secure equality testing and comparison are two important primitives that have been widely used in many secure computation scenarios, such as privacy-preserving machine learning, private set intersection, secure data mining, etc. In this work, we propose new constant-round two-party computation (2PC) protocols for secure equality testing and secure comparison. Our protocols are designed in the online/offline paradigm. Theoretically, for 32-bit integers, the online communication for our...

Secure Aggregation (SA) stands as a crucial component in modern Federated Learning (FL) systems, facilitating collaborative training of a global machine learning model while protecting the privacy of individual clients' local datasets. Many existing SA protocols described in the FL literature operate synchronously, leading to notable runtime slowdowns due to the presence of stragglers (i.e. late-arriving clients). To address this challenge, one common approach is to consider stragglers as...

Privacy-preserving decision tree evaluation (PDTE) allows a client that holds feature vectors to perform inferences against a decision tree model on the server side without revealing feature vectors to the server. Our work focuses on the non-interactive batched setting where the client sends a batch of encrypted feature vectors and then obtains classifications, without any additional interaction. This is useful in privacy-preserving credit scoring, biometric authentication, and many more...

Running machine learning algorithms on encrypted data is a way forward to marry functionality needs common in industry with the important concerns for privacy when working with potentially sensitive data. While there is already a growing field on this topic and a variety of protocols, mostly employing fully homomorphic encryption or performing secure multiparty computation (MPC), we are the first to propose a protocol that makes use of a specialized encryption scheme that allows to do secure...

At S$\&$P 2023, a family of secure three-party computing protocols called Bicoptor was proposed by Zhou et al., which is used to compute non-linear functions in privacy preserving machine learning. In these protocols, two parties $P_0, P_1$ respectively hold the corresponding shares of the secret, while a third party $P_2$ acts as an assistant. The authors claimed that neither party in the Bicoptor can independently compromise the confidentiality of the input, intermediate, or output. In...

With the growing adoption of cloud computing, the ability to store data and delegate computations to powerful and affordable cloud servers have become advantageous for both companies and individual users. However, the security of cloud computing has emerged as a significant concern. Particularly, Cloud Service Providers (CSPs) cannot assure data confidentiality and computations integrity in mission-critical applications. In this paper, we propose a confidential and verifiable delegation...

The Gradient Boosting Decision Tree (GBDT) is a well-known machine learning algorithm, which achieves high performance and outstanding interpretability in real-world scenes such as fraud detection, online marketing and risk management. Meanwhile, two data owners can jointly train a GBDT model without disclosing their private dataset by executing secure Multi-Party Computation (MPC) protocols. In this work, we propose NodeGuard, a highly efficient two party computation (2PC) framework for...

This paper introduces a new method for training decision trees and random forests using CKKS homomorphic encryption (HE) in cloud environments, enhancing data privacy from multiple sources. The innovative Homomorphic Binary Decision Tree (HBDT) method utilizes a modified Gini Impurity index (MGI) for node splitting in encrypted data scenarios. Notably, the proposed training approach operates in a single cloud security domain without the need for decryption, addressing key challenges in...

Classifying images has become a straightforward and accessible task, thanks to the advent of Deep Neural Networks. Nevertheless, not much attention is given to the privacy concerns associated with sensitive data contained in images. In this study, we propose a solution to this issue by exploring an intersection between Machine Learning and cryptography. In particular, Fully Homomorphic Encryption (FHE) emerges as a promising solution, as it enables computations to be performed on encrypted...

Garbled Circuit (GC) is a basic technique for practical secure computation. GC handles Boolean circuits; it consumes significant network bandwidth to transmit encoded gate truth tables, each of which scales with the computational security parameter $\kappa$. GC optimizations that reduce bandwidth consumption are valuable. It is natural to consider a generalization of Boolean two-input one-output gates (represented by $4$-row one-column lookup tables, LUTs) to arbitrary $N$-row...

Fully Homomorphic Encryption (FHE) is a powerful tool for performing privacy-preserving analytics over encrypted data. A promising method for FHE over real and complex numbers is approximate homomorphic encryption, instantiated with the Cheon-Kim-Kim-Song (CKKS) scheme. The CKKS scheme enables efficient evaluation for many privacy-preserving machine learning applications. While the efficiency advantages of CKKS are clear, there is currently a lot of confusion on how to securely instantiate...

Matrix multiplication is a common operation in applications like machine learning and data analytics. To demonstrate the correctness of such an operation in a privacy-preserving manner, we propose zkMatrix, a zero-knowledge proof for the multiplication of committed matrices. Among the succinct non-interactive zero-knowledge protocols that have an $O(\log n)$ transcript size and $O(\log n)$ verifier time, zkMatrix stands out as the first to achieve $O(n^2)$ prover time and $O(n^2)$ RAM usage...

We study a fundamental problem in Multi-Party Computation, which we call the Multiple Millionaires’ Problem (MMP). Given a set of private integer inputs, the problem is to identify the subset of inputs that equal the maximum (or minimum) of that set, without revealing any further information on the inputs beyond what is implied by the desired output. Such a problem is a natural extension of the Millionaires’ Problem, which is the very first Multi- Party Computation problem that was...

This paper introduces LERNA, a new framework for single-server secure aggregation. Our protocols are tailored to the setting where multiple consecutive aggregation phases are performed with the same set of clients, a fraction of which can drop out in some of the phases. We rely on an initial secret sharing setup among the clients which is generated once-and-for-all, and reused in all following aggregation phases. Compared to prior works [Bonawitz et al. CCS’17, Bell et al. CCS’20], the...

In this work, we introduce FANNG-MPC, a versatile secure multi-party computation framework capable to offer active security for privacy preserving machine learning as a service (MLaaS). Derived from the now deprecated SCALE-MAMBA, FANNG is a data-oriented fork, featuring novel set of libraries and instructions for realizing private neural networks, effectively reviving the popular framework. To the best of our knowledge, FANNG is the first MPC framework to offer actively secure MLaaS in the...

Private computation of nonlinear functions, such as Rectified Linear Units (ReLUs) and max-pooling operations, in deep neural networks (DNNs) poses significant challenges in terms of storage, bandwidth, and time consumption. To address these challenges, there has been a growing interest in utilizing privacy-preserving techniques that leverage polynomial activation functions and kernelized convolutions as alternatives to traditional ReLUs. However, these alternative approaches often suffer...

The privacy-preserving machine learning (PPML) has gained growing importance over the last few years. One of the biggest challenges is to improve the efficiency of PPML so that the communication and computation costs of PPML are affordable for large machine learning models such as deep learning. As we know, linear algebra such as matrix multiplication occupies a significant part of the computation in deep learning such as deep convolutional neural networks (CNN). Thus, it is desirable to...

The advent of transformers has brought about significant advancements in traditional machine learning tasks. However, their pervasive deployment has raised concerns about the potential leakage of sensitive information during inference. Existing approaches using secure multiparty computation (MPC) face limitations when applied to transformers due to the extensive model size and resource-intensive matrix-matrix multiplications. In this paper, we present BOLT, a privacy-preserving inference...

Privacy-preserving machine learning (PPML) techniques have gained significant popularity in the past years. Those protocols have been widely adopted in many real-world security-sensitive machine learning scenarios, e.g., medical care and finance. In this work, we introduce $\mathsf{Aegis}$~-- a high-performance PPML platform built on top of a maliciously secure 3-PC framework over ring $\mathbb{Z}_{2^\ell}$. In particular, we propose a novel 2-round secure comparison (a.k.a., sign bit...

Over years of the development of secure multi-party computation (MPC), many sophisticated functionalities have been made pratical and multi-dimensional operations occur more and more frequently in MPC protocols, especially in protocols involving datasets of vector elements, such as privacy-preserving biometric identification and privacy-preserving machine learning. In this paper, we introduce a new kind of correlation, called tensor triples, which is designed to make multi-dimensional MPC...

Explainable AI (XAI) refers to the development of AI systems and machine learning models in a way that humans can understand, interpret and trust the predictions, decisions and outputs of these models. A common approach to explainability is feature importance, that is, determining which input features of the model have the most significant impact on the model prediction. Two major techniques for computing feature importance are LIME (Local Interpretable Model-agnostic Explanations) and...

Secure multi-party computation (MPC) enables (joint) computations on sensitive data while maintaining privacy. In real-world scenarios, asymmetric trust assumptions are often most realistic, where one somewhat trustworthy entity interacts with smaller clients. We generalize previous two-party computation (2PC) protocols like MUSE (USENIX Security'21) and SIMC (USENIX Security'22) to the three-party setting (3PC) with one malicious party, avoiding the performance limitations of...

Computing the distance between two non-normalized vectors $\mathbfit{x}$ and $\mathbfit{y}$, represented by $\Delta(\mathbfit{x},\mathbfit{y})$ and comparing it to a predefined public threshold $\tau$ is an essential functionality used in privacy-sensitive applications such as biometric authentication, identification, machine learning algorithms ({\em e.g.,} linear regression, k-nearest neighbors, etc.), and typo-tolerant password-based authentication. Tackling a widely used distance...

Abstract—Large transformer-based models have realized state- of-the-art performance on lots of real-world tasks such as natural language processing and computer vision. However, with the increasing sensitivity of the data and tasks they handle, privacy has become a major concern during model deployment. In this work, we focus on private inference in two-party settings, where one party holds private inputs and the other holds the model. We introduce BumbleBee, a fast and...

Classic MLaaS solutions suffer from privacy-related risks since the user is required to send unencrypted data to the server hosting the MLaaS. To alleviate this problem, a thriving line of research has emerged called Privacy-Preserving Machine Learning (PPML) or secure MLaaS solutions that use cryptographic techniques to preserve the privacy of both the input of the client and the output of the server. However, these implementations do not take into consideration the possibility of...

Domain Generation Algorithms (DGAs) are used by malware to generate pseudorandom domain names to establish communication between infected bots and Command and Control servers. While DGAs can be detected by machine learning (ML) models with great accuracy, offering DGA detection as a service raises privacy concerns when requiring network administrators to disclose their DNS traffic to the service provider. We propose the first end-to-end framework for privacy-preserving classification as a...

Reducing the size of large dimensional data is a critical task in machine learning (ML) that often involves using principal component analysis (PCA). In privacy-preserving ML, data confidentiality is of utmost importance, and reducing data size is a crucial way to cut overall costs. This work focuses on minimizing the number of normalization processes in the PCA algorithm, which is a costly procedure in encrypted PCA. By modifying Krasulina's algorithm, non-polynomial operations were...

Privacy-Preserving Machine Learning (PPML) provides protocols for learning and statistical analysis of data that may be distributed amongst multiple data owners (e.g., hospitals that own proprietary healthcare data), while preserving data privacy. The PPML literature includes protocols for various learning methods, including ridge regression. Ridge regression controls the $L_2$ norm of the model, but does not aim to strictly reduce the number of non-zero coefficients, namely the $L_0$ norm...

Machine learning is a widely-used tool for analysing large datasets, but increasing public demand for privacy preservation and the corresponding introduction of privacy regulations have severely limited what data can be analysed, even when this analysis is for societal benefit. Homomorphic encryption, which allows computation on encrypted data, is a natural solution to this dilemma, allowing data to be analysed without sacrificing privacy. Because homomorphic encryption is computationally...

Keller and Sun (ICML'22) have found a gap in the accuracy between floating-point deep learning in cleartext and secure quantized deep learning using multi-party computation. We have discovered that this gap is caused by a bug in the implementation of max-pooling. In this note, we present updated figures to support this conclusion. We also add figures for another network on CIFAR-10.

As a fundamental operation in fixed-point arithmetic, truncation can bring the product of two fixed-point integers back to the fixed-point representation. In large-scale applications like privacy-preserving machine learning, it is essential to have faithful truncation that accurately eliminates both big and small errors. In this work, we improve and extend the results of the oblivious transfer based faithful truncation protocols initialized by Cryptflow2 (Rathee et al., CCS 2020)....

Privacy-preserving machine learning (PPML) promises to train machine learning (ML) models by combining data spread across multiple data silos. Theoretically, secure multiparty computation (MPC) allows multiple data owners to train models on their joint data without revealing the data to each other. However, the prior implementations of this secure training using MPC have three limitations: they have only been evaluated on CNNs, and LSTMs have been ignored; fixed point approximations...

Secure Machine Learning as a Service is a viable solution where clients seek secure delegation of the ML computation while protecting their sensitive data. We propose an efficient method to securely evaluate deep standard convolutional neural networks based on CKKS fully homomorphic encryption, in the manner of batch inference. In this paper, we introduce a packing method called Channel-by-Channel Packing that maximizes the slot compactness and single-instruction-multipledata capabilities in...

In today’s interconnected world, data has become a valuable asset, leading to a growing interest in protecting it through techniques such as privacy-preserving computation. Two well-known approaches are multi-party computation and homomorphic encryption with use cases such as privacy-preserving machine learning evaluating or training neural networks. For multi-party computation, one of the fundamental arithmetic operations is the secure multiplication in the malicious security model and by...

The increased popularity of Machine Learning as a Service (MLaaS) makes the privacy of user data and network weights a critical concern. Using Torus FHE (TFHE) offers a solution for privacy-preserving computation in a cloud environment by allowing computation directly over encrypted data. However, software TFHE implementations of cyphertext-cyphertext multiplication needed when both input data and weights are encrypted are either lacking or are too slow. This paper proposes a new way to...

Homomorphic encryption (HE) enables calculating on encrypted data, which makes it possible to perform privacy- preserving neural network inference. One disadvantage of this technique is that it is several orders of magnitudes slower than calculation on unencrypted data. Neural networks are commonly trained using floating-point, while most homomorphic encryption libraries calculate on integers, thus requiring a quantisation of the neural network. A straightforward approach would be to...

The concept of using Lookup Tables (LUTs) instead of Boolean circuits is well-known and been widely applied in a variety of applications, including FPGAs, image processing, and database management systems. In cryptography, using such LUTs instead of conventional gates like AND and XOR results in more compact circuits and has been shown to substantially improve online performance when evaluated with secure multi-party computation. Several recent works on secure floating-point computations and...

Tremendous efforts have been made to improve the efficiency of secure Multi-Party Computation (MPC), which allows n ≥ 2 parties to jointly evaluate a target function without leaking their own private inputs. It has been confirmed by previous research that Three-Party Computation (3PC) and outsourcing computations to GPUs can lead to huge performance improvement of MPC in computationally intensive tasks such as Privacy-Preserving Machine Learning (PPML). A natural question to ask is whether...

Secure 2-party computation (2PC) of floating-point arithmetic is improving in performance and recent work runs deep learning algorithms with it, while being as numerically precise as commonly used machine learning (ML) frameworks like PyTorch. We find that the existing 2PC libraries for floating-point support generic computations and lack specialized support for ML training. Hence, their latency and communication costs for compound operations (e.g., dot products) are high. We provide novel...

In this paper, we propose a non-interactive privacy-preserving naive Bayes classifier from leveled fully homomorphic encryption schemes. The classifier runs on a server that is also the model’s owner (modeler), whose input is the encrypted data from a client. The classifier produces encrypted classification results, which can only be decrypted by the client, while the modelers model is only accessible to the server. Therefore, the classifier does not leak any privacy on either the servers...

Machine learning (ML) has seen a strong rise in popularity in recent years and has become an essential tool for research and industrial applications. Given the large amount of high quality data needed and the often sensitive nature of ML data, privacy-preserving collaborative ML is of increasing importance. In this paper, we introduce new actively secure multiparty computation (MPC) protocols which are specially optimized for privacy-preserving machine learning applications. We concentrate...

Privacy enhancing technologies (PETs) have been proposed as a way to protect the privacy of data while still allowing for data analysis. In this work, we focus on Fully Homomorphic Encryption (FHE), a powerful tool that allows for arbitrary computations to be performed on encrypted data. FHE has received lots of attention in the past few years and has reached realistic execution times and correctness. More precisely, we explain in this paper how we apply FHE to tree-based models and get...

With the development of sequencing technologies, viral strain classification -- which is critical for many applications, including disease monitoring and control -- has become widely deployed. Typically, a lab (client) holds a viral sequence, and requests classification services from a centralized repository of labeled viral sequences (server). However, such ``classification as a service'' raises privacy concerns. In this paper we propose a privacy-preserving viral strain classification...

We propose a novel privacy-preserving, two-party computation of various distance metrics (e.g., Hamming distance, Scalar Product) followed by a comparison with a fixed threshold, which is known as one of the most useful and popular building blocks for many different applications including machine learning, biometric matching, etc. Our solution builds upon recent advances in function secret sharing and makes use of an optimized version of arithmetic secret sharing. Thanks to this combination,...

Data privacy concerns are increasing significantly in the context of Internet of Things, cloud services, edge computing, artificial intelligence applications, and other applications enabled by next generation networks. Homomorphic Encryption addresses privacy challenges by enabling multiple operations to be performed on encrypted messages without decryption. This paper comprehensively addresses homomorphic encryption from both theoretical and practical perspectives. The paper delves into the...

Secure aggregation enables a server to learn the sum of client-held vectors in a privacy-preserving way, and has been successfully applied to distributed statistical analysis and machine learning. In this paper, we both introduce a more efficient secure aggregation construction and extend secure aggregation by enabling input validation, in which the server can check that clients' inputs satisfy required constraints such as $L_0$, $L_2$, and $L_\infty$ bounds. This prevents malicious clients...

Zero-knowledge proof is a critical cryptographic primitive. Its most practical type, called zero-knowledge Succinct Non-interactive ARgument of Knowledge (zkSNARK), has been deployed in various privacy-preserving applications such as cryptocurrencies and verifiable machine learning. Unfortunately, zkSNARK like Groth16 has a high overhead on its proof generation step, which consists of several time-consuming operations, including large-scale matrix-vector multiplication (MUL),...

In the classical notion of multiparty computation (MPC), an honest party learning private inputs of others, either as a part of protocol specification or due to a malicious party's unspecified messages, is not considered a potential breach. Several works in the literature exploit this seemingly minor loophole to achieve the strongest security of guaranteed output delivery via a trusted third party, which nullifies the purpose of MPC. Alon et al. (CRYPTO 2020) presented the notion of Friends...

The overhead of non-linear functions dominates the performance of the secure multiparty computation (MPC) based privacy-preserving machine learning (PPML). This work introduces a family of novel secure three-party computation (3PC) protocols, Bicoptor, which improve the efficiency of evaluating non-linear functions. The basis of Bicopter is a new sign determination protocol, which relies on a clever use of the truncation protocol proposed in SecureML (S\&P 2017). Our 3PC sign...

The offline-online model is a leading paradigm for practical secure multi-party computation (MPC) protocol design that has successfully reduced the overhead for several prevalent privacy-preserving computation functionalities common to diverse application domains. However, the prohibitive overheads associated with secure comparison -- one of these vital functionalities -- often bottlenecks current and envisioned MPC solutions. Indeed, an efficient secure comparison solution has the potential...

We implement training of neural networks in secure multi-party computation (MPC) using quantization commonly used in said setting. We are the first to present an MNIST classifier purely trained in MPC that comes within 0.2 percent of the accuracy of the same convolutional neural network trained via plaintext computation. More concretely, we have trained a network with two convolutional and two dense layers to 99.2% accuracy in 3.5 hours (under one hour for 99% accuracy). We have also...

Secure multi-party computation (MPC) is an essential tool for privacy-preserving machine learning (ML). However, secure training of large-scale ML models currently requires a prohibitively long time to complete. Given that large ML inference and training tasks in the plaintext setting are significantly accelerated by Graphical Processing Units (GPUs), this raises the natural question: can secure MPC leverage GPU acceleration? A few recent works have studied this question in the context of...

Machine learning as a service scenario typically requires the client to trust the server and provide sensitive data in plaintext. However, with the recent improvements in fully homomorphic encryption (FHE) schemes, many such applications can be designed in a privacy preserving way. In this work, we focus on such a problem, private decision tree evaluation (PDTE) --- where a server has a decision tree classification model, and a client wants to use the model to classify her private data...

The growing volumes of data being collected and its analysis to provide better services are creating worries about digital privacy. To address privacy concerns and give practical solutions, the literature has relied on secure multiparty computation. However, recent research has mostly focused on the small-party honest-majority setting of up to four parties, noting efficiency concerns. In this work, we extend the strategies to support a larger number of participants in an honest-majority...

In this paper, we propose a new concept of hierarchical rotation key for homomorphic encryption to reduce the burdens of the clients and the server running on the fully homomorphic encryption schemes such as Cheon-Kim-Kim-Song (CKKS) and Brakerski/Fan-Vercauteran (BFV) schemes. Using this concept, after the client generates and transmits only a small set of rotation keys to the server, the server can generate any required rotation keys from the public key and the smaller set of rotation keys...

Privacy preservation is a sensitive issue in our modern society. It is becoming increasingly important in many applications in this ever-growing and highly connected digital era. Functional encryption is a computation on encrypted data paradigm that allows users to retrieve the evaluation of a function on encrypted data without revealing the data, thus effectively protecting users' privacy. However, existing functional encryption implementations are still very time-consuming for practical...

We present the first algorithm that combines privacy-preserving technologies and state-of-the-art explainable AI to enable privacy-friendly explanations of black-box AI models. We provide a secure algorithm for contrastive explanations of black-box machine learning models that securely trains and uses local foil trees. Our work shows that the quality of these explanations can be upheld whilst ensuring the privacy of both the training data, and the model itself.

We build a library SecFloat for secure 2-party computation (2PC) of 32-bit single-precision floating-point operations and math functions. The existing functionalities used in cryptographic works are imprecise and the precise functionalities used in standard libraries are not crypto-friendly, i.e., they use operations that are cheap on CPUs but have exorbitant cost in 2PC. SecFloat bridges this gap with its novel crypto-friendly precise functionalities. Compared to the prior cryptographic...

Quantum information has the property that measurement is an inherently destructive process. This feature is most apparent in the principle of complementarity, which states that mutually incompatible observables cannot be measured at the same time. Recent work by Broadbent and Islam (TCC 2020) builds on this aspect of quantum mechanics to realize a cryptographic notion called certified deletion. While this remarkable notion enables a classical verifier to be convinced that a (private-key)...

Secure two-party neural network inference (2PC-NN) can offer privacy protection for both the client and the server and is a promising technique in the machine-learning-as-a-service setting. However, the large overhead of the current 2PC-NN in- ference systems is still being a headache, especially when applied to deep neural networks such as ResNet50. In this work, we present Cheetah, a new 2PC-NN inference system that is faster and more communication-efficient than state-of-the-arts. The...

Comparisons or Inequality Tests are an essential building block of Rectified Linear Unit functions (ReLU's), ever more present in Machine Learning, specifically in Neural Networks. Motivated by the increasing interest in privacy-preserving Artificial Intelligence, we explore the current state of the art of privacy preserving comparisons over Multi-Party Computation (MPC). We then introduce constant round variations and combinations, which are compatible with customary fixed point arithmetic...

We address the problem of learning a machine learning model from training data that originates at multiple data owners, while providing formal privacy guarantees regarding the protection of each owner's data. Existing solutions based on Differential Privacy (DP) achieve this at the cost of a drop in accuracy. Solutions based on Secure Multiparty Computation (MPC) do not incur such accuracy loss but leak information when the trained model is made publicly available. We propose an MPC solution...

Recently, the standard ResNet-20 network was successfully implemented on residue number system variant Cheon-Kim-Kim-Song (RNS-CKKS) scheme using bootstrapping, but the implementation lacks practicality due to high latency and low security level. To improve the performance, we first minimize total bootstrapping runtime using multiplexed parallel convolution that collects sparse output data for multiple channels compactly. We also propose the \emph{imaginary-removing bootstrapping} to prevent...

Today's data-intensive applications increasingly suffer from significant performance bottlenecks due to the limited memory bandwidth of the classical von Neumann architecture. Near-Data Processing (NDP) has been proposed to perform computation near memory or data storage to reduce data movement for improving performance and energy consumption. However, the untrusted NDP processing units (PUs) bring in new threats to workloads that are private and sensitive, such as private database queries...

Data, when coupled with state-of-the-art machine learning models, can enable remarkable applications. But, there exists an underlying tension: users wish to keep their data private, and model providers wish to protect their intellectual property. Homomorphic encryption (HE) and multi-party computation (MPC) techniques have been proposed as solutions to this problem; however, both techniques require model providers to fully trust the server performing the machine learning computation. This...

Due to standardization, AES is today’s most widely used block cipher. Its security is well-studied and hardware acceleration is available on a variety of platforms. Following the success of the Intel AES New Instructions (AES-NI), support for Vectorized AES (VAES) has been added in 2018 and already shown to be useful to accelerate many implementations of AES-based algorithms where the order of AES evaluations is fixed a priori. In our work, we focus on using VAES to accelerate the...

Recently, the application of multi-party secure computing schemes based on homomorphic encryption in the field of machine learning attracts attentions across the research fields. Previous studies have demonstrated that secure protocols adopting packed additive homomorphic encryption (PAHE) schemes based on the ring learning with errors (RLWE) problem exhibit significant practical merits, and are particularly promising in enabling efficient secure inference in machine-learning-as-a-service...

Machine learning as a service (MLaaS) has risen to become a prominent technology due to the large development time, amount of data, hardware costs, and level of expertise required to develop a machine learning model. However, privacy concerns prevent the adoption of MLaaS for applications with sensitive data. A promising privacy preserving solution is to use fully homomorphic encryption (FHE) to perform the ML computations. Recent advancements have lowered computational costs by several...

The increasing adoption of machine learning inference in applications has led to a corresponding increase in concerns surrounding the privacy guarantees offered by existing mechanisms for inference. Such concerns have motivated the construction of efficient secure inference protocols that allow parties to perform inference without revealing their sensitive information. Recently, there has been a proliferation of such proposals, rapidly improving efficiency. However, most of these protocols...

In more detail, we propose Soteria, a system for distributed privacy-preserving ML that leverages Trusted Execution Environments (e.g., Intel SGX) to run computations over sensitive information in isolated containers (enclaves). Unlike previous work, where all ML-related computation is performed at trusted enclaves, we introduce a hybrid scheme, combining computation done inside and outside these enclaves. The experimental evaluation validates that our approach reduces the runtime of ML...

Secure computation enables two or more parties to jointly evaluate a function without revealing to each other their private input. G-module is an abelian group M, where the group G acts compatibly with the abelian group structure on M. In this work, we present several secure computation protocols for G-module operations in the online/offline mode. We then show how to instantiate those protocols to implement many widely used secure computation primitives in privacy-preserving machine learning...

Secure Multiparty Computation (MPC) is an invaluable tool for training machine learning models when the training data cannot be directly accessed by the model trainer. Unfortunately, complex algorithms, such as deep learning models, have their computational complexities increased by orders of magnitude when performed using MPC protocols. In this contribution, we study how to efficiently train an important class of machine learning problems by using MPC where features are known by one of the...

In many machine learning applications, training data consists of sensitive information from multiple sources. Privacy-preserving machine learning using secure computation enables multiple parties to compute on their joint data without disclosing their inputs to each other. In this work, we focus on clustering, an unsupervised machine learning technique that partitions data into groups. Previous works on privacy-preserving clustering often leak information and focus on the k-means algorithm,...

Clustering is a popular unsupervised machine learning technique that groups similar input elements into clusters. It is used in many areas ranging from business analysis to health care. In many of these applications, sensitive information is clustered that should not be leaked. Moreover, nowadays it is often required to combine data from multiple sources to increase the quality of the analysis as well as to outsource complex computation to powerful cloud servers. This calls for efficient...

Fully homomorphic encryption (FHE) is one of the prospective tools for privacy-preserving machine learning (PPML), and several PPML models have been proposed based on various FHE schemes and approaches. Although the FHE schemes are known as suitable tools to implement PPML models, previous PPML models on FHE such as CryptoNet, SEALion, and CryptoDL are limited to only simple and non-standard types of machine learning models. These non-standard machine learning models are not proven efficient...

Secure aggregation is a critical component in federated learning, which enables the server to learn the aggregate model of the users without observing their local models. Conventionally, secure aggregation algorithms focus only on ensuring the privacy of individual users in a single training round. We contend that such designs can lead to significant privacy leakages over multiple training rounds, due to partial user selection/participation at each round of federated learning. In fact, we...

In the era of cloud computing and machine learning, data has become a highly valuable resource. Recent history has shown that the benefits brought forth by this data driven culture come at a cost of potential data leakage. Such breaches have a devastating impact on individuals and industry, and lead the community to seek privacy preserving solutions. A promising approach is to utilize Fully Homomorphic Encryption (FHE) to enable machine learning over encrypted data, thus providing resiliency...

Mixing arithmetic and boolean circuits to perform privacy-preserving machine learning has become increasingly popular. Towards this, we propose a framework for the case of four parties with at most one active corruption called Tetrad. Tetrad works over rings and supports two levels of security, fairness and robustness. The fair multiplication protocol costs 5 ring elements, improving over the state-of-the-art Trident (Chaudhari et al. NDSS'20). A key feature of Tetrad is that robustness...

Most existing Secure Multi-Party Computation (MPC) protocols for privacy-preserving training of decision trees over distributed data assume that the features are categorical. In real-life applications, features are often numerical. The standard ``in the clear'' algorithm to grow decision trees on data with continuous values requires sorting of training examples for each feature in the quest for an optimal cut-point in the range of feature values in each node. Sorting is an expensive...

Machine Learning (ML) algorithms, especially deep neural networks (DNN), have proven themselves to be extremely useful tools for data analysis, and are increasingly being deployed in systems operating on sensitive data, such as recommendation systems, banking fraud detection, and healthcare systems. This underscores the need for privacy-preserving ML (PPML) systems, and has inspired a line of research into how such systems can be constructed efficiently. We contribute to this line of...

We present a framework GenoPPML for privacy-preserving machine learning in the context of sensitive genomic data processing. The technology combines secure multiparty computation techniques based on the recently proposed Manticore secure multiparty computation framework for model training and fully homomorphic encryption based on TFHE for model inference. The framework was successfully used to solve breast cancer prediction problems on gene expression datasets coming from distinct private...

Recent progress in interactive zero-knowledge (ZK) proofs has improved the efficiency of proving large-scale computations significantly. Nevertheless, real-life applications (e.g., in the context of private inference using deep neural networks) often involve highly complex computations, and existing ZK protocols lack the expressiveness and scalability to prove results about such computations efficiently. In this paper, we design, develop, and evaluate a ZK system (Mystique) that allows for...