Randomized distributed function computation refers to remote function computation where transmitters send data to receivers which compute function outputs that are randomized functions of the inputs. We study the applications of semantic communications in randomized distributed function computation to illustrate significant reductions in the communication load, with a particular focus on privacy. The semantic communication framework leverages generalized remote source coding methods, where...

Local differential privacy (LDP) is an efficient solution for providing privacy to client's sensitive data while simultaneously releasing aggregate statistics without relying on a trusted central server (aggregator) as in the central model of differential privacy. The shuffle model with LDP provides an additional layer of privacy, by disconnecting the link between clients and the aggregator, further improving the utility of LDP. However, LDP has been shown to be vulnerable to malicious...

There exists a mismatch between the theory and practice of cryptography in the presence of leakage. On the theoretical front, the bounded leakage model, where the adversary learns bounded-length but noiseless information about secret components, and the random probing model, where the adversary learns some internal values of a leaking implementation with some probability, are convenient abstractions to analyze the security of numerous designs. On the practical front, side-channel attacks...

In past years, entire research communities have arisen to address concerns of privacy and fairness in data analysis. At present, however, the public must trust that institutions will re-implement algorithms voluntarily to account for these social concerns. Due to additional cost, widespread adoption is unlikely without effective legal enforcement. A technical challenge for enforcement is that the methods proposed are often probabilistic mechanisms, whose output must be drawn according to...

The shuffle model has recently emerged as a popular setting for differential privacy, where clients can communicate with a central server using anonymous channels or an intermediate message shuffler. This model was also explored in the context of cryptographic tasks such as secure aggregation and private information retrieval (PIR). However, this study was almost entirely restricted to the stringent notion of information-theoretic security. In this work, we study computationally secure...

Provable privacy typically requires involved analysis and is often associated with unacceptable accuracy loss. While many empirical verification or approximation methods, such as Membership Inference Attacks (MIA) and Differential Privacy Auditing (DPA), have been proposed, these do not offer rigorous privacy guarantees. In this paper, we apply recently-proposed Probably Approximately Correct (PAC) Privacy to give formal, mechanized, simulation-based proofs for a range of practical,...

Vertical Federated Learning (VFL) is becoming a standard collaborative learning paradigm with various practical applications. Randomness is essential to enhancing privacy in VFL, but introducing too much external randomness often leads to an intolerable performance loss. Instead, as it was demonstrated for other federated learning settings, leveraging internal randomness —as provided by variational autoencoders (VAEs) —can be beneficial. However, the resulting privacy has never been...

Machine-learning systems continue to advance at a rapid pace, demonstrating remarkable utility in various fields and disciplines. As these systems continue to grow in size and complexity, a nascent industry is emerging which aims to bring machine-learning-as-a-service (MLaaS) to market. Outsourcing the operation and training of these systems to powerful hardware carries numerous advantages, but challenges arise when needing to ensure privacy and the correctness of work carried out by a...

Analyzing user data while protecting the privacy of individuals remains a big challenge. Trusted execution environments (TEEs) are a possible solution as they protect processes and Virtual Machines (VMs) against malicious hosts. However, TEEs can leak access patterns to code and to the data being processed. Furthermore, when data is stored in a TEE database, the data volume required to answer a query is another unwanted side channel that contains sensitive information. Both types of...

We present a construction for secure computation of differentially private sparse histograms that aggregates the inputs from a large number of clients. Each client contributes a value to the aggregate at a specific index. We focus on the case where the set of possible indices is superpolynomially large. Hence, the resulting histogram will be sparse, i.e., most entries will have the value zero. Our construction relies on two non-colluding servers and provides security against malicious...

Banks publish daily a list of available securities/assets (axe list) to selected clients to help them effectively locate Long (buy) or Short (sell) trades at reduced financing rates. This reduces costs for the bank, as the list aggregates the bank's internal firm inventory per asset for all clients of long as well as short trades. However, this is somewhat problematic: (1) the bank's inventory is revealed; (2) trades of clients who contribute to the aggregated list, particularly those deemed...

One of the most basic problems for studying the "price of privacy over time" is the so called private counter problem, introduced by Dwork et al. (2010) and Chan et al. (2010). In this problem, we aim to track the number of events that occur over time, while hiding the existence of every single event. More specifically, in every time step $t\in[T]$ we learn (in an online fashion) that $\Delta_t\geq 0$ new events have occurred, and must respond with an estimate $n_t\approx\sum_{j=1}^t...

In differential-like attacks, the process typically involves extending a distinguisher forward and backward with probability 1 for some rounds and recovering the key involved in the extended part. Particularly in rectangle attacks, a holistic key recovery strategy can be employed to yield the most efficient attacks tailored to a given distinguisher. In this paper, we treat the distinguisher and the extended part as an integrated entity and give a one-step framework for finding rectangle...

Differential privacy is a fundamental concept for protecting individual privacy in databases while enabling data analysis. Conceptually, it is assumed that the adversary has no direct access to the database, and therefore, encryption is not necessary. However, with the emergence of cloud computing and the «on-cloud» storage of vast databases potentially contributed by multiple parties, it is becoming increasingly necessary to consider the possibility of the adversary having (at least...

The rising tide of data breaches targeting large data storage centres and servers has raised serious privacy and security concerns. Homomorphic Encryption schemes offer an effective defence against such attacks, but their adoption has been hindered by substantial computational and communication overheads, particularly on the client's side. The Hybrid Homomorphic Encryption (HEE) protocol was developed to mitigate these issues. However, the susceptibility of HHE to strong attacks,...

In this work we revisit the problem of using general-purpose MPC schemes to emulate the trusted dataholder in differential privacy (DP), to achieve the same accuracy but without the need to trust one single dataholder. In particular, we consider the two-party model where two computational parties (or dataholders), each with their own dataset, wish to compute a canonical DP mechanism on their combined data and to do so with active security. We start by remarking that available definitions of...

Common block ciphers like AES specified by the NIST or KASUMI (A5/3) of GSM are extensively utilized by billions of individuals globally to protect their privacy and maintain confidentiality in daily communications. However, these ciphers lack comprehensive security proofs against the vast majority of known attacks. Currently, security proofs are limited to differential and linear attacks for both AES and KASUMI. For instance, the consensus on the security of AES is not based on formal...

How to achieve distributed differential privacy (DP) without a trusted central party is of great interest in both theory and practice. Recently, the shuffle model has attracted much attention. Unlike the local DP model in which the users send randomized data directly to the data collector/analyzer, in the shuffle model an intermediate untrusted shuffler is introduced to randomly permute the data, which have already been randomized by the users, before they reach the analyzer. The most...

In Crypto 94, Chor, Fiat, and Naor first introduced the traitor tracing (TT) systems, which aim at helping content distributors identify pirates. Since its introduction, many traitor tracing schemes have been proposed. However, we observe until now almost all the traitor tracing systems using probabilistic public key (and secret key) encryption as the the content distribution algorithm, they do not consider this basic fact: the malicious encrypter can plant some trapdoor in the randomness...

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...

While secure multi-party computation (MPC) protects the privacy of inputs and intermediate values of a computation, differential privacy (DP) ensures that the output itself does not reveal too much about individual inputs. For this purpose, MPC can be used to generate noise and add this noise to the output. However, securely generating and adding this noise is a challenge considering real-world implementations on finite-precision computers, since many DP mechanisms guarantee privacy only...

Typically, a decentralized collaborative blockchain decision-making mechanism is realized by remote voting. To date, a number of blockchain voting schemes have been proposed; however, to the best of our knowledge, none of these schemes achieve coercion-resistance. In particular, for most blockchain voting schemes, the randomness used by the voting client can be viewed as a witness/proof of the actual vote, which enables improper behaviors such as coercion and vote-buying. Unfortunately, the...

The min-hash sketch is a well-known technique for low-communication approximation of the Jaccard index between two input sets. Moreover, there is a folklore belief that min-hash sketch based protocols protect the privacy of the inputs. In this paper, we investigate this folklore to quantify the privacy of the min-hash sketch. We begin our investigation by considering the privacy of min-hash in a centralized setting where the hash functions are chosen by the min-hash functionality and...

The proof of stake (PoS) mechanism, which allows stakeholders to issue a block with a probability proportional to their wealth instead of computational power, is believed to be an energy-efficient alternative to the proof of work (PoW). The privacy concern of PoS, however, is more subtle than that of PoW. Recent research has shown that current anonymous PoS (APoS) protocols do not suffice to protect the stakeholder's identity and stake, and the loss of privacy is theoretically inherent for...

We present a new framework for defining information leakage in the setting of US equities trading, and construct methods for deriving trading schedules that stay within specified information leakage bounds. Our approach treats the stock market as an interactive protocol performed in the presence of an adversary, and draws inspiration from the related disciplines of differential privacy as well as quantitative information flow. We apply a linear programming solver using examples from...

In the classical setting of differential privacy, a privacy-preserving query is performed on a private database, after which the query result is released to the analyst; a differentially private query ensures that the presence of a single database entry is protected from the analyst’s view. In this work, we contribute the first definitional framework for differential privacy in the trusted curator setting; clients submit private inputs to the trusted curator, which then computes individual...

In this paper, we pose the problem of defining technical privacy guarantees for anonymizing stories. This is a challenge that arises in practice in contexts like journalism and memoir writing, and can be crucial in determining whether a story gets told and how effectively it is presented. While recent decades have seen leaps and bounds in the development of approaches like differential privacy for numerical and categorical data sets, none of these techniques are directly applicable to...

Given a collection of vectors $\mathbf{x}^{(1)},\dots,\mathbf{x}^{(n)} \in \{0,1\}^d$, the selection problem asks to report the index of an "approximately largest" entry in $\mathbf{x}=\sum_{j=1}^n \mathbf{x}^{(j)}$. Selection abstracts a host of problems; in machine learning it can be used for hyperparameter tuning, feature selection, or to model empirical risk minimization. We study selection under differential privacy, where a released index guarantees privacy for individual vectors....

Addition of $n$ inputs is often the easiest nontrivial function to compute securely. Motivated by several open questions, we ask what can be computed securely given only an oracle that computes the sum. Namely, what functions can be computed in a model where parties can only encode their input locally, then sum up the encodings over some Abelian group $\G$, and decode the result to get the function output. An *additive randomized encoding* (ARE) of a function $f(x_1,\ldots,x_n)$ maps...

Differential obliviousness (DO) is a privacy notion which mandates that the access patterns of a program satisfy differential privacy. Earlier works have shown that in numerous applications, differential obliviousness allows us to circumvent fundamental barriers pertaining to fully oblivious algorithms, resulting in asymptotical (and sometimes even polynomial) performance improvements. Although DO has been applied to various contexts, including the design of algorithms, data structures, and...

Safety, liveness, and privacy are three critical properties for any private proof-of-stake (PoS) blockchain. However, prior work (SP'21) has shown that to obtain safety and liveness, a PoS blockchain must in theory forgo privacy. Specifically, to ensure safety and liveness, PoS blockchains elect parties based on stake proportion, potentially exposing a party's stake even with private transaction processing. In this work, we make two key contributions. First, we present the first stake...

Homomorphic Encryption (HE) is a type of cryptography that allows computing on encrypted data, enabling computation on sensitive data to be outsourced securely. Many popular HE schemes rely on noise for their security. On the other hand, Differential Privacy seeks to guarantee the privacy of data subjects by obscuring any one individual's contribution to an output. Many mechanisms for achieving Differential Privacy involve adding appropriate noise. In this work, we investigate the extent to...

Computing the quantile of a massive data stream has been a crucial task in networking and data management. However, existing solutions assume a centralized model where one data owner has access to all data. In this paper, we put forward a study of secure quantile aggregation between private data streams, where data streams owned by different parties would like to obtain a quantile of the union of their data without revealing anything else about their inputs. To this end, we designed efficient...

Measuring people’s interactions that span multiple websites can provide unique insight that enables better products and improves people’s experiences, but directly observing people’s individual journeys creates privacy risks that conflict with the newly emerging privacy model for the web. We propose a protocol that uses the combination of multi-party computation and differential privacy that enables the processing of peoples’ data such that only aggregate measurements are revealed, strictly...

Homomorphic encryption for approximate arithmetic allows one to encrypt discretized real/complex numbers and evaluate arithmetic circuits over them. The first scheme, called CKKS, was introduced by Cheon et al. (Asiacrypt 2017) and gained tremendous attention. The enthusiasm for CKKS-type encryption stems from its potential to be used in inference or multiparty computation tasks that do not require an exact output. A desirable property for homomorphic encryption is circuit privacy,...

We study the space complexity of the two related fields of differential privacy and adaptive data analysis. Specifically, (1) Under standard cryptographic assumptions, we show that there exists a problem $P$ that requires exponentially more space to be solved efficiently with differential privacy, compared to the space needed without privacy. To the best of our knowledge, this is the first separation between the space complexity of private and non-private algorithms. (2) The line of...

Privacy-preserving transaction systems on blockchain networks like Monero or Zcash provide complete transaction anonymity through cryptographic commitments or encryption. While this secures privacy, it inhibits the collection of statistical data, which current financial markets heavily rely on for economic and sociological research conducted by central banks, statistics bureaus, and research companies. Differential privacy techniques have been proposed to preserve individuals' privacy while...

In recent years, there has been significant work in studying data structures that provide privacy for the operations that are executed. These primitives aim to guarantee that observable access patterns to physical memory do not reveal substantial information about the queries and updates executed on the data structure. Multiple recent works, including Larsen and Nielsen [Crypto'18], Persiano and Yeo [Eurocrypt'19], Hubáček et al. [TCC'19] and Komargodski and Lin [Crypto'21], have shown that...

To bound information leakage in outputs of protocols, it is important to construct secure multiparty computation protocols which output differentially private values perturbed by the addition of noise. However, previous noise generation protocols have round and communication complexity growing with differential privacy budgets, or require parties to locally generate non-uniform noise, which makes it difficult to guarantee differential privacy against active adversaries. We propose three...

Differential obliviousness (DO) access pattern privacy is a privacy notion which guarantees that the access patterns of a program satisfy differential privacy. Differential obliviousness was studied in a sequence of recent works as a relaxation of full obliviousness. Earlier works showed that DO not only allows us to circumvent the logarithmic-overhead barrier of fully oblivious algorithms, in many cases, it also allows us to achieve polynomial speedup over full obliviousness, since it...

Modern Singular Value Decomposition (SVD) computation dates back to the 1960s when the basis for the eigensystem package and linear algebra package routines was created. Since then, SVD has gained attraction and been widely applied in various scenarios, such as recommendation systems and principal component analyses. Federated SVD has recently emerged, where different parties could collaboratively compute SVD without exchanging raw data. Besides its inherited privacy protection, noise...

Differential privacy (DP) is a key tool in privacy-preserving data analysis. Yet it remains challenging for non-privacy-experts to prove the DP of their algorithms. We propose a methodology for domain experts with limited data privacy background to empirically estimate the privacy of an arbitrary mechanism. Our Eureka moment is a new link---which we prove---between the problems of DP parameter-estimation and Bayes optimal classifiers in ML, which we believe can be of independent interest....

We consider the computation of sparse, $(\varepsilon, \delta)$-differentially private~(DP) histograms in the two-server model of secure multi-party computation~(MPC), which has recently gained traction in the context of privacy-preserving measurements of aggregate user data. We introduce protocols that enable two semi-honest non-colluding servers to compute histograms over the data held by multiple users, while only learning a private view of the data. Our solution achieves the same...

In the current digital world, large organizations (sometimes referred to as tech giants) provide service to extremely large numbers of users. The service provider is often interested in computing various data analyses over the private data of its users, which in turn have their incentives to cooperate, but do not necessarily trust the service provider. In this work, we introduce the \emph{Gulliver multi-party computation model} (GMPC) to realistically capture the above scenario. The GMPC...

Oblivious RAM (ORAM) is a machinery that protects any RAM from leaking information about its secret input by observing only the access pattern. It is known that every ORAM must incur a logarithmic overhead compared to the non-oblivious RAM. In fact, even the seemingly weaker notion of differential obliviousness, which intuitively ``protects'' a single access by guaranteeing that the observed access pattern for every two ``neighboring'' logical access sequences satisfy...

Contact discovery allows users of mobile messengers to conveniently connect with people in their address book. In this work, we demonstrate that severe privacy issues exist in currently deployed contact discovery methods and propose suitable mitigations. Our study of three popular messengers (WhatsApp, Signal, and Telegram) shows that large-scale crawling attacks are (still) possible. Using an accurate database of mobile phone number prefixes and very few resources, we queried 10% of US...

Recent work of Li and Micciancio (Eurocrypt 2021) has shown that the traditional formulation of indistinguishability under chosen plaintext attack (INDCPA) is not adequate to capture the security of approximate homomorphic encryption against passive adversaries, and identified a stronger INDCPA^D security definition (INDCPA with decryption oracles) as the appropriate security target for approximate encryption schemes. We show how to any approximate homomorphic encryption scheme achieving...

Deep learning (DL) models for natural language processing (NLP) tasks often handle private data, demanding protection against breaches and disclosures. Data protection laws, such as the European Union's General Data Protection Regulation (GDPR), thereby enforce the need for privacy. Although many privacy-preserving NLP methods have been proposed in recent years, no categories to organize them have been introduced yet, making it hard to follow the progress of the literature. To close this...

We study fine-grained error bounds for differentially private algorithms for averaging and counting in the continual observation model. For this, we use the completely bounded spectral norm (cb norm) from operator algebra. For a matrix $W$, its cb norm is defined as \[ \|{W}\|_{\mathsf{cb}} = \max_{Q} \left\{ \frac{\|{Q \bullet W}\|}{\|{Q}\|} \right\}, \] where $Q \bullet W$ denotes the Schur product and $\|{\cdot}\|$ denotes the spectral norm. We bound the cb norm of two fundamental...

The shuffle model has been extensively investigated in the distributed differential privacy (DP) literature. For a class of useful computational tasks, the shuffle model allows us to achieve privacy-utility tradeoff similar to those in the central model, while shifting the trust from a central data curator to a ``trusted shuffle'' which can be implemented through either trusted hardware or cryptography. Very recently, several works explored cryptographic instantiations of a new type of...

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...

We extend a previous framework for designing differentially private (DP) mechanisms via randomized graph colorings that was restricted to binary functions, corresponding to colorings in a graph, to multi-valued functions. As before, datasets are nodes in the graph and any two neighboring datasets are connected by an edge. In our setting, we assume each dataset has a preferential ordering for the possible outputs of the mechanism, which we refer to as a rainbow. Different rainbows partition...

Payment channel networks (PCNs) provide a faster and cheaper alternative to transactions recorded on the blockchain. Clients can trustlessly establish payment channels with relays by locking coins and then send signed payments that shift coin balances over the network's channels. Although payments are never published, anyone can track a client's payment by monitoring changes in coin balances over the network's channels. We present Twilight, the first PCN that provides a rigorous...

The use of data combined with tailored statistical analysis have presented a unique opportunity to organizations in diverse fields to observe users' behaviors and needs, and accordingly adapt and fine-tune their services. However, in order to offer utilizable, plausible, and personalized alternatives to users, this process usually also entails a breach of their privacy. The use of statistical databases for releasing data analytics is growing exponentially, and while many cryptographic...

Searchable symmetric encryption (SSE) enables users to make confidential queries over always encrypted data while confining information disclosure to pre-defined leakage profiles. Despite the well-understood performance and potentially broad applications of SSE, recent leakage-abuse attacks (LAAs) are questioning its real-world security implications. They show that a passive adversary with certain prior information of a database can recover queries by exploiting the legitimately admitted...

In the shuffle model for differential privacy, $n$ users locally randomize their data and submit the results to a trusted “shuffler” who mixes the results before sending them to a server for analysis. This is a promising model for real-world applications of differential privacy, as several recent results have shown that the shuffle model sometimes offers a strictly better privacy/utility tradeoff than what is possible in a purely local model. A downside of the shuffle model is its...

A recent work by Shi and Wu (Eurocrypt'21) sugested a new, non-interactive abstraction for anonymous routing, coined Non-Interactive Anonymous Router (\NIAR). They show how to construct a \NIAR scheme with succinct communication from bilinear groups. Unfortunately, the router needs to perform quadratic computation (in the number of senders/receivers) to perform each routing. In this paper, we show that if one is willing to relax the security notion to $(\epsilon, \delta)$-differential...

We propose a new theoretical approach for building anonymous mixing mechanisms for cryptocurrencies. Rather than requiring a fully uniform permutation during mixing, we relax the requirement, insisting only that neighboring permutations are similarly likely. This is defined formally by borrowing from the definition of differential privacy. This relaxed privacy definition allows us to greatly reduce the amount of interaction and computation in the mixing protocol. Our construction achieves...

We initiate a study of the composition properties of interactive differentially private mechanisms. An interactive differentially private mechanism is an algorithm that allows an analyst to adaptively ask queries about a sensitive dataset, with the property that an adversarial analyst's view of the interaction is approximately the same regardless of whether or not any individual's data is in the dataset. Previous studies of composition of differential privacy have focused on non-interactive...

Fuzzy Message Detection (FMD) is a recent cryptographic primitive invented by Beck et al. (CCS'21) where an untrusted server performs coarse message filtering for its clients in a recipient-anonymous way. In FMD --- besides the true positive messages --- the clients download from the server their cover messages determined by their false-positive detection rates. What is more, within FMD, the server cannot distinguish between genuine and cover traffic. In this paper, we formally analyze the...

Differential privacy is a concept to quantify the disclosure of private information that is controlled by the privacy parameter~$\varepsilon$. However, an intuitive interpretation of $\varepsilon$ is needed to explain the privacy loss to data engineers and data subjects. In this paper, we conduct a worst-case study of differential privacy risks. We generalize an existing model and reduce complexity to provide more understandable statements on the privacy loss. To this end, we analyze the...

Constant function market makers (CFMMs) are the most popular mechanism for facilitating decentralized trading. While these mechanisms have facilitated hundreds of billions of dollars of trades, they provide users with little to no privacy. Recent work illustrates that privacy cannot be achieved in CFMMs without forcing worse pricing and/or latency on end users. This paper more precisely quantifies the trade-off between pricing and privacy in CFMMs. We analyze a simple privacy-enhancing...

Symmetric Searchable Encryption (SSE) schemes enable keyword searches over encrypted documents. To obtain efficiency, SSE schemes incur a certain amount of leakage. The vast majority of the literature on SSE considers only leakage from one component of the overall SSE system, the encrypted search index. This component is used to identify which documents to return in response to a keyword query. The actual fetching of the documents is left to another component, usually left unspecified in the...

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...

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...

Information-theoretical privacy relies on randomness. Representatively, Differential Privacy (DP) has emerged as the gold standard to quantify the individual privacy preservation provided by given randomness. However, almost all randomness in existing differentially private optimization and learning algorithms is restricted to noise perturbation. In this paper, we set out to provide a privacy analysis framework to understand the privacy guarantee produced by other randomness commonly used in...

We tackle the problems of private learning where an owner wishes to outsource a training task to an honest-but-curious server while keeping its data private, and private collaborative learning where two (or more) mutually distrusting owners outsource respective training data sets to an honest-but-curious server while keeping their data sets private from the server and each other. The privacy property we provide is information-theoretic in nature, Probably Approximately Correct (PAC)...

Numerous high-profile works have shown that access patterns to even encrypted databases can leak secret information and sometimes even lead to reconstruction of the entire database. To thwart access pattern leakage, the literature has focused on {\it oblivious} algorithms, where obliviousness requires that the access patterns leak nothing about the input data. In this paper, we consider the {\tt Join} operator, an important database primitive that has been extensively studied and optimized....

Differential privacy has been widely applied to provide privacy guarantees by adding random noise to the function output. However, it inevitably fails in many high-stakes voting scenarios, where voting rules are required to be deterministic. In this work, we present the first framework for answering the question: ``How private are commonly-used voting rules?" Our answers are two-fold. First, we show that deterministic voting rules provide sufficient privacy in the sense of distributional...

We revisit private optimization and learning from an information processing view. The main contribution of this paper is twofold. First, different from the classic cryptographic framework of operation-by-operation obfuscation, a novel private learning and inference framework via either random or data-dependent transformation on the sample domain is proposed. Second, we propose a novel security analysis framework, termed probably approximately correct (PAC) inference resistance, which bridges...

Federated learning (FL) allows to train a massive amount of data privately due to its decentralized structure. Stochastic gradient descent (SGD) is commonly used for FL due to its good empirical performance, but sensitive user information can still be inferred from weight updates shared during FL iterations. We consider Gaussian mechanisms to preserve local differential privacy (LDP) of user data in the FL model with SGD. The trade-offs between user privacy, global utility, and transmission...

Federated Learning (FL) is a collaborative machine learning approach allowing participants to jointly train a model without having to share their private, potentially sensitive local datasets with others. Despite its benefits, FL is vulnerable to so-called backdoor attacks, in which an adversary injects manipulated model updates into the federated model aggregation process so that the resulting model will provide targeted false predictions for specific adversary-chosen inputs. Proposed...

Private Information Retrieval (PIR) allows several clients to query a database held by one or more servers, such that the contents of their queries remain private. Prior PIR schemes have achieved sublinear communication and computation by leveraging computational assumptions, federating trust among many servers, relaxing security to permit differentially private leakage, refactoring effort into an offline stage to reduce online costs, or amortizing costs over a large batch of queries. In...

Secure computation is a promising privacy enhancing technology, but it is often not scalable enough for data intensive applications. On the other hand, the use of sketches has gained popularity in data mining, because sketches often give rise to highly efficient and scalable sub-linear algorithms. It is natural to ask: what if we put secure computation and sketches together? We investigated the question and the findings are interesting: we can get security, we can get scalability, and...

Functional Encryption (FE) allows users who hold a specific secret key (known as the functional key) to learn a specific function of encrypted data whilst learning nothing about the content of the underlying data. Considering this functionality and the fact that the field of FE is still in its infancy, we sought a route to apply this potent tool to design efficient applications. To this end, we first built a symmetric FE scheme for the $\ell_1$ norm of a vector space, which allows us to...

The shuffle model of differential privacy [Bittau et al. SOSP 2017; Erlingsson et al. SODA 2019; Cheu et al. EUROCRYPT 2019] was proposed as a viable model for performing distributed differentially private computations. Informally, the model consists of an untrusted analyzer that receives messages sent by participating parties via a shuffle functionality, the latter potentially disassociates messages from their senders. Prior work focused on one-round differentially private shuffle model...

Secure aggregation is a cryptographic primitive that enables a server to learn the sum of the vector inputs of many clients. Bonawitz et al. (CCS 2017) presented a construction that incurs computation and communication for each client linear in the number of parties. While this functionality enables a broad range of privacy preserving computational tasks, scaling concerns limit its scope of use. We present the first constructions for secure aggregation that achieve polylogarithmic...

Recent work has shown that cell phone mobility data has the unique potential to create accurate models for human mobility and consequently the spread of infected diseases. While prior studies have exclusively relied on a mobile network operator's subscribers' aggregated data in modelling disease dynamics, it may be preferable to contemplate aggregated mobility data of infected individuals only. Clearly, naively linking mobile phone data with health records would violate privacy by either...

New generation head-mounted displays, such as VR and AR glasses, are coming into the market with already integrated eye tracking and are expected to enable novel ways of human-computer interaction in numerous applications. However, since eye movement properties contain biometric information, privacy concerns have to be handled properly. Privacy-preservation techniques such as differential privacy mechanisms have recently been applied to eye movement data obtained from such displays. Standard...

This paper introduces a new cryptographic primitive called a private resource allocator (PRA) that can be used to allocate resources (e.g., network bandwidth, CPUs) to a set of clients without revealing to the clients whether any other clients received resources. We give several constructions of PRAs that provide guarantees ranging from information-theoretic to differential privacy. PRAs are useful in preventing a new class of attacks that we call allocation-based side-channel attacks. These...

We consider a scenario where multiple organizations holding large amounts of sensitive data from their users wish to compute aggregate statistics on this data while protecting the privacy of individual users. To support large-scale analytics we investigate how this privacy can be provided for the case of sketching algorithms running in time sub-linear of the input size. We begin with the well-known LogLog sketch for computing the number of unique elements in a data stream. We show that this...

An exciting new development in differential privacy is the shuffled model, in which an anonymous channel enables non-interactive, differentially private protocols with error much smaller than what is possible in the local model, while relying on weaker trust assumptions than in the central model. In this paper, we study basic counting problems in the shuffled model and establish separations between the error that can be achieved in the single-message shuffled model and in the shuffled model...

Machine Learning has been widely applied in practice, such as disease diagnosis, target detection. Commonly, a good model relies on massive training data collected from different sources. However, the collected data might expose sensitive information. To solve the problem, researchers have proposed many excellent methods that combine machine learning with privacy protection technologies, such as secure multiparty computation(MPC), homomorphic encryption(HE), and differential privacy. In the...

Volume leakage has recently been identified as a major threat to the security of cryptographic cloud-based data structures by Kellaris et al. [CCS’16] (see also the attacks in Grubbs et al. [CCS’18] and Lacharité et al. [S&P’18]). In this work, we focus on volume-hiding implementations of encrypted multi-maps as first considered by Kamara and Moataz [Eurocrypt’19]. Encrypted multi-maps consist of outsourcing the storage of a multi-map to an untrusted server, such as a cloud storage system,...

We design an efficient method for sampling a large batch of $d$ independent coins with a given bias $p \in [0,1]$. The folklore secure computation method for doing so requires $O(\lambda + \log d)$ communication and computation per coin to achieve total statistical difference $2^{-\lambda}$. We present an exponential improvement over the folklore method that uses just $O(\log(\lambda+\log d))$ gates per coin when sampling $d$ coins with total statistical difference $2^{-\lambda}$. We...

Data minimization has become a paradigm to address privacy concerns when collecting and storing personal data. In this paper we present two new approaches, RSTxFM and RRTxFM, to estimate the cardinality of a dataset while ensuring differential privacy. We argue that privacy-preserving cardinality estimators are able to realize strong privacy requirements. Both approaches are based on a probabilistic counting algorithm which has a logarithmic space complexity. We combine this with a...

Consider a PPT two-party protocol $\Pi=(A,B)$ in which the parties get no private inputs and obtain outputs $O^A,O^B\in \{0,1\}$, and let $V^A$ and $V^B$ denote the parties' individual views. Protocol $\Pi$ has $\alpha$-agreement if $Pr[O^A=O^B]=1/2+\alpha$. The leakage of $\epsilon$ is the amount of information a party obtains about the event $\{O^A=O^B\}$; that is, the leakage $\epsilon$ is the maximum, over $P\in \{A,B\}$, of the distance between $V^P|_{O^A=O^B}$ and $V^P|_{O^A\neq O^B}$....

Oblivious RAM (ORAM) and private information retrieval (PIR) are classic cryptographic primitives used to hide the access pattern to data whose storage has been outsourced to an untrusted server. Unfortunately, both primitives require considerable overhead compared to plaintext access. For large-scale storage infrastructure with highly frequent access requests, the degradation in response time and the exorbitant increase in resource costs incurred by either ORAM or PIR prevent their usage....

We consider the problem of designing scalable, robust protocols for computing statistics about sensitive data. Specifically, we look at how best to design differentially private protocols in a distributed setting, where each user holds a private datum. The literature has mostly considered two models: the "central" model, in which a trusted server collects users' data in the clear, which allows greater accuracy; and the "local" model, in which users individually randomize their data, and...

In this work we demonstrate that allowing differentially private leakage can significantly improve the concrete performance of secure 2-party computation (2PC) protocols. Specifically, we focus on the private set intersection (PSI) protocol of Rindal and Rosulek (CCS 2017), which is the fastest PSI protocol with security against malicious participants. We show that if differentially private leakage is allowed, the cost of the protocol can be reduced by up to 63%, depending on the desired...

Dodis and Yu (TCC 2013) studied how the security of cryptographic primitives that are secure in the "ideal" model in which the distribution of a randomness is the uniform distribution, is degraded when the ideal distribution of a randomness is switched to a "real-world" (possibly biased) distribution that has some lowerbound on its min-entropy or collision-entropy. However, in many constructions, their security is guaranteed only when a randomness is sampled from some non-uniform...

In this work, we study privacy-preserving storage primitives that are suitable for use in data analysis on outsourced databases within the differential privacy framework. The goal in differentially private data analysis is to disclose global properties of a group without compromising any individual’s privacy. Typically, differentially private adversaries only ever learn global properties. For the case of outsourced databases, the adversary also views the patterns of access to data. Oblivious...

Fairness in classification has become an increasingly relevant and controversial issue as computers replace humans in many of today’s classification tasks. In particular, a subject of much recent debate is that of finding, and subsequently achieving, suitable definitions of fairness in an algorithmic context. In this work, following the work of Hardt et al. (NIPS’16), we consider and formalize the task of sanitizing an unfair classifier C into a classifier C' satisfying an approximate notion...

The problem of privatizing statistical databases is a well-studied topic that has culminated with the notion of differential privacy. The complementary problem of securing these databases, however, has---as far as we know---not been considered in the past. While the security of private databases is in theory orthogonal to the problem of private statistical analysis (e.g., in the central model of differential privacy the curator is trusted) the recent real-world deployments of...

Quantifying the privacy loss of a privacy-preserving mechanism on potentially sensitive data is a complex and well-researched topic; the de-facto standard for privacy measures are $\varepsilon$-differential privacy (DP) and its versatile relaxation $(\varepsilon,\delta)$-approximate differential privacy (ADP). Recently, novel variants of (A)DP focused on giving tighter privacy bounds under continual observation. In this paper we unify many previous works via the \emph{privacy loss...

Secure multi-party computation (MPC) is a general cryptographic technique that allows distrusting parties to compute a function of their individual inputs, while only revealing the output of the function. It has found applications in areas such as auctioning, email filtering, and secure teleconference. Given their importance, it is crucial that the protocols are specified and implemented correctly. In the programming language community, it has become good practice to use computer proof...

Privacy-preserving collaborative data analysis enables richer models than what each party can learn with their own data. Secure Multi-Party Computation (MPC) offers a robust cryptographic approach to this problem, and in fact several protocols have been proposed for various data analysis and machine learning tasks. In this work, we focus on secure similarity computation between text documents, and the application to $k$-nearest neighbors (\knn) classification. Due to its non-parametric...

This technical report discusses three subtleties related to the widely used notion of differential privacy (DP). First, we discuss how the choice of a distinguisher influences the privacy notion and why we should always have a distinguisher if we consider approximate DP. Secondly, we draw a line between the very intuitive probabilistic differential privacy (with probability $1-\delta$ we have $\varepsilon$-DP) and the commonly used approximate differential privacy...

Bad choices of passwords were and are a pervasive problem. Users choosing weak passwords do not only compromise themselves, but the whole ecosystem. E.g, common and default passwords in IoT devices were exploited by hackers to create botnets and mount severe attacks on large Internet services, such as the Mirai botnet DDoS attack. We present a method to help protect the Internet from such large scale attacks. Our method enables a server to identify popular passwords (heavy hitters), and...

In this work we seek to construct collusion-resistant traitor tracing systems with small ciphertexts from standard assumptions that also move toward practical efficiency. In our approach we will hold steadfast to the principle of collusion resistance, but relax the requirement on catching a traitor from a successful decoding algorithm. We define a $f$-risky traitor tracing system as one where the probability of identifying a traitor is $f(\lambda,n)$ times the probability a successful box is...