Federated Learning (FL) is currently referred to as one of the privacy-enhancing technologies because of its service architecture. However, recent advancements in FL have high-lighted its potential not only as a new framework of privacy but also as a key enabler of sustainable computing, which is expected to minimize the impact of an individual party to further improve the capacity of the machine learning model, energy efficiency, and reliability. For the above requirement of sustainability, resource allocation and trust management in FL are very infrastructural tasks of energy efficiency and reliability. In this paper, we present a framework, called Veri-SFL, to indicate verification for resource allocation and trust measurement in FL. We use trust scores to represent the credibility of each dataset without leaking any privacy, and utilize collaborative zk-SNARKs to verify the trust scores of each local dataset. Then, after verifying the correctness of trust levels, we present a solution to verify whether workers (model owners) are training according to the required distribution ratio by using collaborative zk-SNARKs.

Abstract In a smart city, IoT devices are required to support monitoring of normal operations such as traffic, infrastructure, and the crowd of people. IoT‐enabled systems offered by many IoT devices are expected to achieve sustainable developments from the information collected by the smart city. Indeed, artificial intelligence (AI) and machine learning (ML) are well‐known methods for achieving this goal as long as the system framework and problem statement are well prepared. However, to better use AI/ML, the training data should be as global as possible, which can prevent the model from working only on local data. Such data can be obtained from different sources, but this induces the privacy issue where at least one party collects all data in the plain. The main focus of this article is on support vector machines (SVM). We aim to present a solution to the privacy issue and provide confidentiality to protect the data. We build a privacy‐preserving scheme for SVM (SecretSVM) based on the framework of federated learning and distributed consensus. In this scheme, data providers self‐organize and obtain training parameters of SVM without revealing their own models. Finally, experiments with real data analysis show the feasibility of potential applications in smart cities. This article is the extended version of that of Hsu et al. (Proceedings of the 15th ACM Asia Conference on Computer and Communications Security. ACM; 2020:904‐906).

With the prevalence of artificial intelligence, people collect data through numerous sensors and use machine learning to create models for intelligent services. However, data privacy and massive data issues are raised with the proliferation of devices. Although secret sharing can be the solution to providing privacy and performing efficient computations (than homomorphic encryption-based) in the privacy domain, sharing large amounts of data may impose a considerable storage burden on resource-restricted devices. Therefore, we aim to reduce the shares that participants need to hold and construct secure computation protocols that form privacy-preserving data utilization with multi-secret sharing. In this paper, we study the mechanism of persistent computation with multi-secret sharing for privacy protection, which avoids the requirements for participants to store extra information during secure computation. We present the mask techniques to convert secrets into protected data and upload them separately to non-colluding dual-cloud servers through multi-secret sharing. Then, we can locally evaluate the operation result shares by revealing a part of the protected data, where the true secret consistently remains secure as the other part stays in a shared state. Eventually, we outsource a dataset to the cloud, build a privacy-preserving multi-party kNN classification based on our scheme, and provide some experiments to demonstrate the feasibility and usability of storage size.

In the era of big data with an ever-increasing amount of data, the demand for data processing is increasing, and various computing service agreements have also been created. While data are widely used, data providers worry that their privacy will be infringed, so the security of the data has also received more attention. Private set intersection (PSI) allows two data-owning parties to jointly compute the intersection of private datasets. The party cannot obtain any other information except for the intersection results. However, Bloom filter (BF) is a naive solution that can be used to potentially obtain PSI, but the verifiability of using BF is usually a concern for effectiveness. Hence, garbled Bloom filter is indeed expected to overcome this concern in two-party setting. We aim for the outsourced scenario where there are many data-owning parties in distributed setting. These parties would like to jointly obtain the set intersection in a private manner supported by an outsourced cloud, and further to obtain data integration. The main challenges are that the previous solutions cannot be directly transformed into an outsourced scenario. In this paper, the proposed joint and effective privacy preserving outsourced set intersection and data integration protocols are called JEDI. They are generic constructions. We start by introducing a basic outsourced PSI protocol from the cryptographic building blocks and new BF and GBF techniques. Then, we modify the basic protocol and propose the full-fledged scheme to achieve an outsourced multi-party PSI protocol. Finally, we analyze the security and execution efficiency aspects to ensure usability.

With the rapid development of cloud computing, an increasing number of companies are adopting cloud storage technology to reduce overhead. However, to ensure the privacy of sensitive data, the uploaded data need to be encrypted before being outsourced to the cloud. The concept of public-key encryption with keyword search (PEKS) was introduced by Boneh et al. to provide flexible usage of the encrypted data. Unfortunately, most of the PEKS schemes are not secure against inside keyword guessing attacks (IKGA), so the keyword information of the trapdoor may be leaked to the adversary. To solve this issue, Huang and Li presented public key authenticated encryption with keyword search (PAEKS) in which the trapdoor generated by the receiver is only valid for authenticated ciphertexts. With their seminal work, many PAEKS schemes have been introduced for the enhanced security of PAEKS. Some of them further consider the upcoming quantum attacks. However, our cryptanalysis indicated that in fact, these schemes could not withstand IKGA. To fight against the attacks from quantum adversaries and support the privacy-preserving search functionality, we first introduce a novel generic PAEKS construction in this work. Then, we further present the first quantum-resistant PAEKS instantiation based on lattices. The security proofs show that our instantiation not only satisfies the basic requirements but also achieves enhanced security models, namely the multi-ciphertext indistinguishability and multi-trapdoor privacy. Furthermore, the comparative results indicate that with only some additional expenditure, the proposed instantiation provides more secure properties, making it suitable for more diverse application environments.

The k-nearest neighbor (kNN) algorithm is a very simple manner in the area of machine learning. It is a supervised method to classify according to the distance between different instances and is also widely used in solving some classification problems. It is expected to obtain better training with a larger dataset. However, how to perform kNN algorithm efficiently is an issue with privacy-preserving. In this paper, we proposed a privacy-preserving k -nearest neighboring scheme by secret sharing and improve the kNN classification by preprocessing with tree structures. Finally, the applicability of our method is shown by experiments with real datasets.

Abstract The industrial Internet of Things (IIoT) integrates sensors, instruments, equipment and industrial applications, enabling traditional industries to automate and intelligently process data. To reduce the cost and demand of required service equipment, IIoT relies on cloud computing to further process and store data. Public-key encryption with keyword search (PEKS) plays an important role, due to its search functionality, to ensure the privacy and confidentiality of the outsourced data and the maintenance of flexibility in the use of the data. Recently, Huang and Li proposed the ‘public-key authenticated encryption with keyword search’ (PAEKS) to avoid the insider keyword guessing attacks (IKGAs) in the previous PEKS schemes. However, all current PAEKS schemes are based on the discrete logarithm assumption and are therefore vulnerable to quantum attacks. In this study, we first introduce a generic PAEKS construction, with the assistance of a trusted authority, that enjoys the security against IKGA in the standard model, if all building blocks are secure under standard model. Based on the framework, we further propose a novel instantiation of quantum-resistant PAEKS that is based on NTRU assumption under random oracle. Compared with its state-of-the-art counterparts, the experiment result indicates that our instantiation is more efficient and secure.

Abstract Identity‐based encryption (IBE), introduced by Shamir, eliminates the need for public‐key infrastructure. The sender can simply encrypt a message by using the recipient's identity (such as email or IP address) without needing to look up the public key. In particular, when ciphertexts of an IBE do not reveal recipient's identity, this scheme is known as an anonymous IBE scheme. Recently, Blazy et al. (ARES '19) analysed the trade‐off between public safety and unconditional privacy in anonymous IBE and introduced a new notion that incorporates traceability into anonymous IBE, called anonymous IBE with traceable identities (AIBET). However, their construction is based on the discrete logarithm assumption, which is insecure in the quantum era. In this paper, we first formalize the consistency of tracing key of the AIBET scheme to ensure that a ciphertext cannot be traced with the use of wrong tracing keys. Subsequently, we present a generic formulation concept that can be used to transform structure‐specific lattice‐based anonymous IBE schemes into an AIBET. Finally, we apply this concept to Katsumata and Yamada's compact anonymous IBE scheme (Asiacrypt '16) to obtain the first quantum‐resistant AIBET scheme that is adaptively secure under the ring learning with errors assumption without random oracle.

Blockchain has become very popular and suitable to the Internet of Things (IoT) field due to its nontamperability and decentralization properties. The number of IoT devices and leaders (who own IoT devices) is increased exponentially, and thus, data privacy and security are undoubtedly significant concerns. In this paper, we summarize some issues for the BeeKeeper system, a blockchain‐based IoT system, proposed by Zhou et al., and then aim for presenting an improved solution for decentralized data aggregation (DDA) on IoT. Firstly, we formally state the security requirements of DDA. Secondly, we propose our basic DDA system by using secret sharing to improve its efficiency and smart contracts as the computing processors. Moreover, the proposed full‐fledged system achieves data sharing (e.g., a leader to access data of others’ devices), which is realized by using local differential privacy and cryptographic primitives such as token‐based encryption. Finally, to show the feasibility, we provide some implementations and experiments for the DDA systems.

Data stored on centralized cloud servers may have some risks. Moreover, it may leak the data access pattern when accessing data on cloud servers. Oblivious RAM (ORAM) is a candidate solution to hide the data access pattern, but it inherently induces some overheads of accessing data. In this paper, we address the issues above, and propose the oblivious access system on decentralized database over parallel smart contract model. The proposed system can slightly reduce the overhead of ORAM and overcome the problem from centralized servers. The main techniques are to use parallel smart contract model to improve the performance of smart contract execution and garbled circuit to reduce the cost of communication.

Support vector machine (SVM) is widely used because of its efficiency in data processing. Single source data may not train a nice SVM, since a single entity isn't having enough data with adequate attributes. Thus, multiple source data need to share data to combine a dataset with different attributes, and then jointly train a classifier. However, outsourcing data to a cloud for training induces two security concerns, data privacy and collusion with the cloud and providers. In this paper, we consider a distributed scenario without any centralized party in which IoT data providers will jointly serve as the leader to obtain some parameters. We propose a privacy-preserving and collusion-free SVM (so-called SecretSVM)built from secret sharing and distributed consensus. Participants train intermediate values to provide the necessary interaction andprevent against collusion attacks.

Reversible data hiding in encrypted images (RDHEI) has been introduced for preserving image privacy and data embedding. RDHEI usually involves three parties, namely, the image provider, data hider, and receiver. On the security with key setting, there are three categories: share independent secret keys (SIK), shared one key (SOK), and share no secret keys (SNK). In SIK, the image provider and data hider must respectively and independently share secret keys with the receiver, whereas in SNK, no secret key is shared. However, the literature works proposed SNK-type schemes by using homomorphic encryption (with exorbitant computation cost). In this paper, we address the SOK setting, where only the image provider shares a secret key with the receiver, and the data hider can embed a secret message without any knowledge of this key. To realize our SOK scheme in a simple manner, we propose a new technique by using multi-secret sharing as the underlying encryption, which indeed induces a blow-up issue of the key size. For preserving the efficiency of the key size, we apply a compression by using lightweight cryptographic algorithms. Then, we demonstrate our SOK scheme based on the proposed techniques, and show effectiveness, efficiency, and security by experiments and analysis.

Visual cryptography (VC) is a variant form of secret sharing. In general threshold setting, the k-out-of-n VC allows that, in a set of n participants, any k can recover and reconstruct the secret by stacking their shares. Recently, the notion of multiple-secret VC has been introduced to embed multiple secrets. Region incrementing visual cryptography (RIVC) is referred to as a new type of multi-secret VC. RIVC defines s layers and takes s secrets, and then embeds each secret into each layer. The layers are defined by the number of participants; for example, let two secrets and two layers be S2, S3and L2, L3in two-out-of-three RIVC, where any two participants in L2can recover S2and three in L3can recover S2, S3. However, there is another multi-secret VC, called fully incrementing visual cryptography (FIVC), which also has the layers, but only one secret Si will reveal in one layer Li. In this paper, our stating point is to propose a new notion of non-monotonic visual cryptography (NVC) for human vision system as a primitive to construct FIVC. We first present an ideal construction of simple NVC, which relies on a slightly unreasonable assumption. Based on the simple NVC, we show a few methods to extend the functionality for complicated cases of NVC. Then, the generic construction is presented as a systematic manner to eliminate the above-mentioned assumption. Finally, we formally introduce a transformation NVC-to-FIVC algorithm, which takes NVC as input and then produce a construction of FIVC. Also, show a demonstration the NVC-to-RIVC algorithm, and analyze some properties regarding NVC. We believe that the notion of NVC can potentially find other applications and is of independent interest.

Since many cryptographic schemes are about performing computation on data, it is important to consider a computation model which captures the prominent features of modern system architecture. Parallel random access machine (PRAM) is such an abstraction which not only models multiprocessor platforms, but also new frameworks supporting massive parallel computation such as MapReduce.

Cloud storage allows users to outsource their data to a storage server. For general security and privacy concerns, users prefer storing encrypted data to pure ones so that servers do not learn anything about privacy. However, there is a natural issue that servers have worked some analyses (i.e. statistics) or routines for encrypted data without losing privacy. In this paper, we address the basic functionality, equality test, over encrypted data, which at least can be applied to specific analyses like private information retrieval. We introduce a new system, called filtered equality test, which is an additional functionality for existing public key encryption schemes. It satisfies the following scenario: a ciphertext-receiver selects several messages as a set and produces its related warrant; then, on receiving this warrant, an user is able to perform equality test on the receiver's ciphertext without decryption when the hidden message belongs to that message set. Similar to the attribute based encryption, ABE. In ABE schemes, those ones who match the settled conditions could get the privilege of decryption. In FET schemes, those ‘messages inside selected set’ can be equality tested. Combining PKE schemes and filtered equality test, we propose a framework of public key encryption scheme with filtered equality test, abbreviated as PKE-FET. Then, taking ElGamal for example, we propose a concrete PKE-FET scheme based on secret sharing and bilinear map. Finally, we prove our proposition with semantic security in the standard model.

Public key encryption with keyword search (PEKS) is a system for realizing keyword search over encrypted data, but communication must rely on a secure channel. In PEKS, a sender would like to share data with a receiver via a storage server. For security and privacy purpose, he must upload the encrypted data to the server, and further the server can search encrypted data (sent by the sender) by using a keyword trapdoor (given by the receiver). In the literature, a new system, server-designation public key encryption with keyword search (dPEKS), is introduced to eliminate the assumption of the secure channel in PEKS. The security models are defined for trapdoor security. However, dPEKS suffers from the on-line keyword guessing attack. In this paper, we first analyze this weakness of dPEKS and enhance the security models with practical adversaries. Secondly, owing to the shown weakness, we consider and define ‘original ciphertext indistinguishability’ regarding the task after the server's search. Finally, we present a new framework of secure server-designation public key encryption with keyword search (SPEKS), and analyze its security.

Recently, Luo et al. proposed a new block-based reversible data hiding method that divides an image into blocks, sorts all pixels within a block and then chooses the middle sorted pixel as a basic-point. To maintain the same basic-point before and after embedding, Luo et al.'s method classifies blocks into four types and adopts four different strategies, respectively, to embed a secret message. In this paper, we propose a new method which relies on only one strategy to reach the same effectiveness. The proposed method includes a complexity controlled mechanism to filter blocks that are not suitable for hiding the secret message. High image quality is thereby retained. Based on just-noticeable difference mechanism, the proposed scheme also estimates allowable changes such that the stego-image is able to meet the human visual system. Experimental results showed that the proposed method allows more payload and higher image quality than the method of Luo et al. and other related methods.

In this paper, we propose a new notion of public key encryption scheme with authorized equality test (PKE-AET), which allows authorized users those who have warrants to test the equivalence between two messages, where the messages are encrypted using different public keys. Comparing with the existing researches, our PKE-AET provides two kinds of warrants that are referred to as receiver's warrants and cipher-warrants. The proposed PKE-AET is able to deal with the following complicated scenario: Assume that a receiver authorizes a receiver's warrant to a tester, which makes the tester be able to perform equality test on all of receivers’ ciphertext; on the other hand, if receiver authorizes a cipher-warrant corresponding to a specific ciphertext to the tester, then the tester only acquires the equality test on that particular ciphertext. The equality between two ciphertexts can be verified by the tester without decryption after he or she receives two warrants and varies their validations. Moreover, for security analysis, we define two types of adversaries and security notions for PKE-AET in the multi-user setting. Furthermore, we prove that our PKE-AET is one-way CCA secure against type-I adversaries and IND-CCA secure against type-II adversaries. Finally, the proposed scheme leads better efficiency than most of previous equality test schemes.

Certificateless public key cryptography CL-PKC is a cryptosystem solving the key escrow problem of identity-based cryptography. One of the applications of CL-PKC is certificateless aggregate signature CLAS that in practice can be used to efficiently verify concealed data aggregation in wireless sensor networks. CLAS is referred to as an extension of certificateless signature, which in particular performs verification for many signatures efficiently. Therefore, not only plenty of CLAS schemes have been proposed but also the security models of CLAS were introduced in the literature. Recently, some CLAS schemes are extended from specific certificateless signature CLS schemes. However, we found that two certificateless signature CLS and their corresponding CLAS schemes are not secure. In this paper, we simplify the relation of security definitions of CLS and CLAS. Then, a new CLAS scheme is proposed, which leads to the advantages of both certificateless cryptography and aggregate signature. Moreover, our scheme only depends on constant pairing operations to verify a large number of signatures per time, because pairing is a complicated operation with high cost in computations. Copyright © 2014 John Wiley & Sons, Ltd.

Camera networks have been deployed to facilitate human tracking across multi-cameras in the modern surveillance systems. However, human privacy is an important concern on security surveillance. More specifically, in the real world, surveillance cameras are commonly installed by different entities (such as departments or companies), and any recorded video by one entity should not be shared with others to protect the privacy of tracked humans, while maintaining the knowledge of those moving trajectories of tracked humans in a centralized cloud server. This tracking across multi-cameras information can serve as a very powerful analysis tool for locating crime suspects or collecting business statistics. This paper is the first to aim at the importance of privacy-preserving in a multiple-camera tracking system. We address the problems of privacy-preserving human tracking based on Paillier encryption without revealing any recorded video or data, and introduce the secure multiple-camera system which consists of two stages: training and testing stages. Finally, the security analyses and simulations show the effectiveness of the proposed scheme.

In this study, the authors consider certificateless signature (CLS) schemes for strong non‐repudiation. They show that previous security models which ensure that any user can have a unique key pair cannot guarantee a CLS scheme to achieve strong non‐repudiation. The authors then fix the security model and propose a new CLS scheme which provides strong non‐repudiation under the new model, assuming the computational Diffie–Hellman problem is intractable.

Visual cryptography (VC), proposed by Naor and Shamir, has numerous applications, including visual authentication and identification, steganography, and image encryption. In 2006, Horng showed that cheating is possible in VC, where some participants can deceive the remaining participants by forged transparencies. Since then, designing cheating-prevention visual secret-sharing (CPVSS) schemes has been studied by many researchers. In this paper, we cryptanalyze the Hu-Tzeng CPVSS scheme and show that it is not cheating immune. We also outline an improvement that helps to overcome the problem.

Blind decoding schemes are proposed as tools for protecting customerspsila privacy in on-line shopping for electronic documents such that the company has no way of knowing which documents the customers have purchased. Most of the blind decoding schemes suffer from the oracle problem. Schemes utilizing the transformability of digital signatures were proposed to ensure the correctness of the requests from the customers. In this paper, a secure blind decoding scheme based on RSA scheme is proposed. It does not utilize the transformability of RSA digital signature.