With the explosive growth of user data, recommenders have become increasingly complicated. State-of-the-art algorithms often have high computational complexity and heavily use non-linear transformations. This fact makes the privacy-preserving problem more challenging, despite the significant advances in cryptography. To alleviate this problem, we propose a privacy-friendly recommender, CryptoRec. It only relies on additions and multiplications, which are efficiently supported by most cryptographic primitives. Different from others, in CryptoRec, the parameter space only contains item features (user features can be directly computed from the item features). This property allows CryptoRec to, (1) naturally achieve transferability if two datasets share the same item entries, which can benefit differential privacy protection; (2) directly estimate the preference of new users whose data is not included in the training set, drastically improving recommendation efficiency. We first evaluate CryptoRec on three real-world datasets. The evaluation results show that the accuracy is competitive with state-of-the-art. Then, we build differential privacy into CryptoRec and leverage its transferability property to reduce the overall privacy loss. Lastly, we demonstrate the simplicity and efficiency of using CryptoRec to construct secure recommendation protocols based on homomorphic encryption schemes. Our results show that CryptoRec outperforms existing solutions in terms of both accuracy and efficiency.

Nowadays, recommender system is an indispensable tool in many information services, and a large number of algorithms have been designed and implemented. However, fed with very large datasets, state-of-the-art recommendation algorithms often face an efficiency bottleneck, i.e., it takes huge amount of computing resources to train a recommendation model. In order to satisfy the needs of privacy-savvy users who do not want to disclose their information to the service provider, the complexity of most existing solutions becomes prohibitive. As such, it is an interesting research question to design simple and efficient recommendation algorithms that achieve reasonable accuracy and facilitate privacy protection at the same time. In this paper, we propose an efficient recommendation algorithm, named CryptoRec, which has two nice properties: (1) can estimate a new user's preferences by directly using a model pre-learned from an expert dataset, and the new user's data is not required to train the model; (2) can compute recommendations with only addition and multiplication operations. As to the evaluation, we first test the recommendation accuracy on three real-world datasets and show that CryptoRec is competitive with state-of-the-art recommenders. Then, we evaluate the performance of the privacy-preserving variants of CryptoRec and show that predictions can be computed in seconds on a PC. In contrast, existing solutions will need tens or hundreds of hours on more powerful computers.

Abstract Machine learning algorithms have reached mainstream status and are widely deployed in many applications. The accuracy of such algorithms depends significantly on the size of the underlying training dataset; in reality a small or medium sized organization often does not have the necessary data to train a reasonably accurate model. For such organizations, a realistic solution is to train their machine learning models based on their joint dataset (which is a union of the individual ones). Unfortunately, privacy concerns prevent them from straightforwardly doing so. While a number of privacy-preserving solutions exist for collaborating organizations to securely aggregate the parameters in the process of training the models, we are not aware of any work that provides a rational framework for the participants to precisely balance the privacy loss and accuracy gain in their collaboration. In this paper, by focusing on a two-player setting, we model the collaborative training process as a two-player game where each player aims to achieve higher accuracy while preserving the privacy of its own dataset. We introduce the notion of Price of Privacy, a novel approach for measuring the impact of privacy protection on the accuracy in the proposed framework. Furthermore, we develop a game-theoretical model for different player types, and then either find or prove the existence of a Nash Equilibrium with regard to the strength of privacy protection for each player. Using recommendation systems as our main use case, we demonstrate how two players can make practical use of the proposed theoretical framework, including setting up the parameters and approximating the non-trivial Nash Equilibrium.

In this work, we consider a scenario that includes an individual sharing his genomic data (or results obtained from his genomic data) with a service provider. In this scenario, (i) the service provider wants to make sure that received genomic data (or results) in fact belongs to the corresponding individual (and computed correctly), (ii) the individual wants to provide a digital consent along with his data specifying whether the service provider is allowed to further share his data, and (iii) if his data is shared without his consent, the individual wants to determine the service provider that is responsible for this leakage. We propose two schemes based on homomorphic signature and aggregate signature that links the information about the legitimacy of the data to the consent and the phenotype of the individual. Thus, to verify the data, each party also needs to use the correct consent and phenotype of the individual who owns the data.

Machine-Learning-as-a-Service has become increasingly popular, with Recommendation-as-a-Service as one of the representative examples. In such services, providing privacy protection for the users is an important topic. Reviewing privacy-preserving solutions which were proposed in the past decade, privacy and machine learning are often seen as two competing goals at stake. Though improving cryptographic primitives (e.g., secure multi-party computation (SMC) or homomorphic encryption (HE)) or devising sophisticated secure protocols has made a remarkable achievement, but in conjunction with state-of-the-art recommender systems often yields far-from-practical solutions. We tackle this problem from the direction of machine learning. We aim to design crypto-friendly recommendation algorithms, thus to obtain efficient solutions by directly using existing cryptographic tools. In particular, we propose an HE-friendly recommender system, refer to as CryptoRec, which (1) decouples user features from latent feature space, avoiding training the recommendation model on encrypted data; (2) only relies on addition and multiplication operations, making the model straightforwardly compatible with HE schemes. The properties turn recommendation-computations into a simple matrix-multiplication operation. To further improve efficiency, we introduce a sparse-quantization-reuse method which reduces the recommendation-computation time by $9\times$ (compared to using CryptoRec directly), without compromising the accuracy. We demonstrate the efficiency and accuracy of CryptoRec on three real-world datasets. CryptoRec allows a server to estimate a user's preferences on thousands of items within a few seconds on a single PC, with the user's data homomorphically encrypted, while its prediction accuracy is still competitive with state-of-the-art recommender systems computing over clear data. Our solution enables Recommendation-as-a-Service on large datasets in a nearly real-time (seconds) level.

Machine learning algorithms have reached mainstream status and are widely deployed in many applications. The accuracy of such algorithms depends significantly on the size of the underlying training dataset; in reality a small or medium sized organization often does not have enough data to train a reasonably accurate model. For such organizations, a realistic solution is to train machine learning models based on a joint dataset (which is a union of the individual ones). Unfortunately, privacy concerns prevent them from straightforwardly doing so. While a number of privacy-preserving solutions exist for collaborating organizations to securely aggregate the parameters in the process of training the models, we are not aware of any work that provides a rational framework for the participants to precisely balance the privacy loss and accuracy gain in their collaboration. In this paper, we model the collaborative training process as a two-player game where each player aims to achieve higher accuracy while preserving the privacy of its own dataset. We introduce the notion of Price of Privacy, a novel approach for measuring the impact of privacy protection on the accuracy in the proposed framework. Furthermore, we develop a game-theoretical model for different player types, and then either find or prove the existence of a Nash Equilibrium with regard to the strength of privacy protection for each player.

In this paper, we introduce new methods to evaluate integer polynomials with the Gentry-Sahai-Waters fully homomorphic encryption (GSW FHE) scheme. Our solution has much slower noise growth and per homomorphic integer multiplication cost, with a ratio O((logk)w+1kw⋅n) in comparison to the original GSW scheme where k is the input plaintext width, n is the LWE dimension and ω=2.373⁠. Technically, we reduce the integer multiplication noise by restricting the evaluation to be between two kinds of ciphertexts: one is for the message space Zq and the other is for message space F2⌈logq⌉⁠. To achieve generality, we propose an integer bootstrapping scheme which converts these two kinds of ciphertexts into each other. To solve the ciphertext expansion problem due to ciphertexts in F2⌈logq⌉⁠, we propose a solution based on symmetric-key encryption with stream ciphers.

In the public‐key setting, known constructions of function‐private functional encryption (FPFE) were limited to very restricted classes of functionalities like inner‐product. Moreover, its power has not been well investigated. The authors construct FPFE for general functions and explore its powerful applications, both for general and specific functionalities. One key observation entailed by their results is that attribute‐based encryption with function privacy implies FE, a notable fact that sheds light on the importance of the function privacy property for FE.

Privacy-preserving recommender systems have been an active research topic for many years. However, until today, it is still a challenge to design an efficient solution without involving a fully trusted third party or multiple semi-trusted third parties. The key obstacle is the large underlying user populations (i.e., huge input size) in the systems. In this paper, we revisit the concept of friendship-based recommender systems, proposed by Jeckmans et al. and Tang and Wang. These solutions are very promising because recommendations are computed based on inputs from a very small subset of the overall user population (precisely, a user's friends and some randomly chosen strangers). We first clarify the single prediction protocol and Top-n protocol by Tang and Wang, by correcting some flaws and improving the efficiency of the single prediction protocol. We then design a decentralized single protocol by getting rid of the semi-honest service provider. In order to validate the designed protocols, we crawl Twitter and construct two datasets (FMT and 10-FMT) which are equipped with auxiliary friendship information. Based on 10-FMT and MovieLens 100k dataset with simulated friendships, we show that even if our protocols use a very small subset of the datasets, their accuracy can still be equal to or better than some baseline algorithm. Based on these datasets, we further demonstrate that the outputs of our protocols leak very small amount of information of the inputs, and the leakage decreases when the input size increases. We finally show that he single prediction protocol is quite efficient but the Top-n is not. However, we observe that the efficiency of the Top-n protocol can be dramatically improved if we slightly relax the desired security guarantee.

In the cloud computing era, in order to avoid the computational burdens, many recommendation service providers tend to outsource their collaborative filtering computations to third-party cloud servers. In order to protect service quality, the integrity of computation results needs to be guaranteed. In this paper, we analyze two integrity verification approaches by Vaidya et al. and demonstrate their performances. In particular, we analyze the verification via auxiliary data approach which is only briefly mentioned in the original paper, and demonstrate the experimental results. We then propose a new solution to outsource all computations of the weighted Slope One algorithm in two-server setting and provide experimental results.

peer reviewed

We revisit the concept of public key encryption with delegated search (PKEDS), a concept proposed by Ibraimi et al. A PKEDS scheme allows a receiver to authorize a third-party server to search in two ways: either according to a message chosen by the server itself or according to a trapdoor sent by the receiver. We first show some observations on the primitive formulation and the proposed PKEDS scheme by Ibraimi et al. and point out that the single-server setting may limit the application of this primitive. We then extend the concept to the multiple-server setting and present a new security model. We finally propose a new PKEDS scheme, which is proven secure, and compare it with that by Ibraimi et al.

Searchable Symmetric Encryption (SSE) allows a client to store encrypted data on a storage provider in such a way, that the client is able to search and retrieve the data selectively without the storage provider learning the contents of the data or the words being searched for. Practical SSE schemes usually leak (sensitive) information during or after a query (e.g., the search pattern). Secure schemes on the other hand are not practical, namely they are neither efficient in the computational search complexity, nor scalable with large data sets. To achieve efficiency and security at the same time, we introduce the concept of distributed SSE (DSSE), which uses a query proxy in addition to the storage provider. We give a construction that combines an inverted index approach (for efficiency) with scrambling functions used in private information retrieval (PIR) (for security). The proposed scheme, which is entirely based on XOR operations and pseudo-random functions, is efficient and does not leak the search pattern. For instance, a secure search in an index over one million documents and 500 keywords is executed in less than 1 second.

Most existing symmetric searchable encryption schemes aim at allowing a user to outsource her encrypted data to a cloud server and delegate the latter to search on her behalf. These schemes do not qualify as a secure and scalable solution for the multiparty setting, where users outsource their encrypted data to a cloud server and selectively authorize each other to search. Due to the possibility that the cloud server may collude with some malicious users, it is a challenge to have a secure and scalable multiparty searchable encryption (MPSE) scheme. This is shown by our analysis on the Popa-Zeldovich scheme, which says that an honest user may leak all her search patterns even if she shares only one of her documents with another malicious user. Based on our analysis, we present a new security model for MPSE by considering the worst case and average-case scenarios, which capture different server-user collusion possibilities. We then propose a MPSE scheme by employing the bilinear property of Type-3 pairings and prove its security based on the bilinear Diffie-Hellman variant and symmetric external Diffie-Hellman assumptions in the random oracle model.

With the rapid development of cloud services, the techniques for securely outsourcing the prohibitively expensive computations to untrusted servers are getting more and more attention in the scientific community. Exponentiations modulo a large prime have been considered the most expensive operations in discrete-logarithm-based cryptographic protocols, and they may be burdensome for the resource-limited devices such as RFID tags or smartcards. Therefore, it is important to present an efficient method to securely outsource such operations to (untrusted) cloud servers. In this paper, we propose a new secure outsourcing algorithm for (variable-exponent, variable-base) exponentiation modulo a prime in the two untrusted program model. Compared with the state-of-the-art algorithm, the proposed algorithm is superior in both efficiency and checkability. Based on this algorithm, we show how to achieve outsource-secure Cramer-Shoup encryptions and Schnorr signatures. We then propose the first efficient outsource-secure algorithm for simultaneous modular exponentiations. Finally, we provide the experimental evaluation that demonstrates the efficiency and effectiveness of the proposed outsourcing algorithms and schemes.

When outsourcing data to third-party servers, searchable encryption is an important enabling technique which simultaneously allows the data owner to keep his data in encrypted form and the third-party servers to search in the ciphertexts. Motivated by an encrypted email retrieval and archive scenario, we investigate asymmetric searchable encryption (ASE) schemes which support two special features, namely message recovery and flexible search authorization. With this new primitive, a data owner can keep his data encrypted under his public key and assign different search privileges to third-party servers. In the security model, we define the standard IND-CCA security against any outside attacker and define adapted ciphertext indistinguishability properties against inside attackers according to their functionalities. Moreover, we take into account the potential information leakage from trapdoors, and define two trapdoor security properties. Employing the bilinear property of pairings and a deliberately-designed double encryption technique, we present a provably secure instantiation of the primitive based on the DLIN and BDH assumptions in the random oracle model.

ABSTRACT In this paper, we investigate a category of public key encryption schemes that supports plaintext equality test and user‐specified authorization. With this new primitive, two users, who possess their own public/private key pairs, can issue token(s) to a proxy to authorize it to perform plaintext equality test from their ciphertexts. We provide a formal formulation for this primitive and present a construction with provable security in our security model. To mitigate the risks against the semi‐trusted proxies, we enhance the proposed cryptosystem by integrating the concept of computational client puzzles. As a showcase, we construct a secure personal health record application on the basis of this primitive. Copyright © 2012 John Wiley & Sons, Ltd.

In this paper, we outline a privacy-preserving matching protocol for OSN (online social network) users to find their potential friends. With the proposed protocol, a logged-in user can match her profile with that of an off-line stranger, while both profiles are maximally protected. Our solution successfully eliminates the requirement of "out-of-band" communication channels, which is one of the biggest obstacles facing cryptographic solutions for OSNs.

In this paper the author shows that the group key agreement protocol proposed by Tseng suffers from a number of serious security vulnerabilities.