Machine Learning (ML) has become one of the most impactful fields of data science in recent years. However, a significant concern with ML is its privacy risks due to rising attacks against ML models. Privacy-Preserving Machine Learning (PPML) methods have been proposed to mitigate the privacy and security risks of ML models. A popular approach to achieving PPML uses Homomorphic Encryption (HE). However, the highly publicized inefficiencies of HE make it unsuitable for highly scalable scenarios with resource-constrained devices. Hence, Hybrid Homomorphic Encryption (HHE) – a modern encryption scheme that combines symmetric cryptography with HE – has recently been introduced to overcome these challenges. HHE potentially provides a foundation to build new efficient and privacy-preserving services that transfer expensive HE operations to the cloud. This work introduces HHE to the ML field by proposing resource-friendly PPML protocols for edge devices. More precisely, we utilize HHE as the primary building block of our PPML protocols. We assess the performance of our protocols by first extensively evaluating each party’s communication and computational cost on a dummy dataset and show the efficiency of our protocols by comparing them with similar protocols implemented using plain BFV. Subsequently, we demonstrate the real-world applicability of our construction by building an actual PPML application that uses HHE as its foundation to classify heart disease based on sensitive ECG data.

The popularity of Machine Learning (ML) makes the privacy of sensitive data more imperative than ever. Collaborative learning techniques like Split Learning (SL) aim to protect client data while enhancing ML processes. Though promising, SL has been proved to be vulnerable to a plethora of attacks, thus raising concerns about its effectiveness on data privacy. In this work, we introduce a hybrid approach combining SL and Function Secret Sharing (FSS) to ensure client data privacy. The client adds a random mask to the activation map before sending it to the servers. The servers cannot access the original function but instead work with shares generated using FSS. Consequently, during both forward and backward propagation, the servers cannot reconstruct the client's raw data from the activation map. Furthermore, through visual invertibility, we demonstrate that the server is incapable of reconstructing the raw image data from the activation map when using FSS. It enhances privacy by reducing privacy leakage compared to other SL-based approaches where the server can access client input information. Our approach also ensures security against feature space hijacking attack, protecting sensitive information from potential manipulation. Our protocols yield promising results, reducing communication overhead by over 2× and training time by over 7× compared to the same model with FSS, without any SL. Also, we show that our approach achieves > 96% accuracy and remains equivalent to the plaintext models.

Machine Learning (ML), addresses a multitude of complex issues in multiple disciplines, including social sciences, finance, and medical research. ML models require substantial computing power and are only as powerful as the data utilized. Due to the high computational cost of ML methods, data scientists frequently use Machine Learning-as-a-Service (MLaaS) to outsource computation to external servers. However, when working with private information, like financial data or health records, outsourcing the computation might result in privacy issues. Recent advances in Privacy-Preserving Techniques (PPTs) have enabled ML training and inference over protected data through the use of Privacy-Preserving Machine Learning (PPML). However, these techniques are still at a preliminary stage and their application in real-world situations is demanding. In order to comprehend the discrepancy between theoretical research suggestions and actual applications, this work examines the past and present of PPML, focusing on Homomorphic Encryption (HE) and Secure Multi-party Computation (SMPC) applied to ML. This work primarily focuses on the ML model's training phase, where maintaining user data privacy is of utmost importance. We provide a solid theoretical background that eases the understanding of current approaches and their limitations. We also provide some preliminaries of SMPC, HE, and ML. In addition, we present a systemization of knowledge of the most recent PPML frameworks for model training and provide a comprehensive comparison in terms of the unique properties and performances on standard benchmarks. Also, we reproduce the results for some of the surveyed papers and examine at what level existing works in the field provide support for open science. We believe our work serves as a valuable contribution by raising awareness about the current gap between theoretical advancements and real-world applications in PPML, specifically regarding open-source availability, reproducibility, and usability.

Split learning (SL) is a new collaborative learning technique that allows participants, e.g. a client and a server, to train machine learning models without the client sharing raw data. In this setting, the client initially applies its part of the machine learning model on the raw data to generate activation maps and then sends them to the server to continue the training process. Previous works in the field demonstrated that reconstructing activation maps could result in privacy leakage of client data. In addition to that, existing mitigation techniques that overcome the privacy leakage of SL prove to be significantly worse in terms of accuracy. In this paper, we improve upon previous works by constructing a protocol based on U-shaped SL that can operate on homomorphically encrypted data. More precisely, in our approach, the client applies homomorphic encryption on the activation maps before sending them to the server, thus protecting user privacy. This is an important improvement that reduces privacy leakage in comparison to other SL-based works. Finally, our results show that, with the optimum set of parameters, training with HE data in the U-shaped SL setting only reduces accuracy by 2.65% compared to training on plaintext. In addition, raw training data privacy is preserved.

Over the past few years, a tremendous growth of machine learning was brought about by a significant increase in adoption and implementation of cloud-based services. As a result, various solutions have been proposed in which the machine learning models run on a remote cloud provider and not locally on a user’s machine. However, when such a model is deployed on an untrusted cloud provider, it is of vital importance that the users’ privacy is preserved. To this end, we propose Learning in the Dark – a hybrid machine learning model in which the training phase occurs in plaintext data, but the classification of the users’ inputs is performed directly on homomorphically encrypted ciphertexts. To make our construction compatible with homomorphic encryption, we approximate the ReLU and Sigmoid activation functions using low-degree Chebyshev polynomials. This allowed us to build Learning in the Dark – a privacy-preserving machine learning model that can classify encrypted images with high accuracy. Learning in the Dark preserves users’ privacy since it is capable of performing high accuracy predictions by performing computations directly on encrypted data. In addition to that, the output of Learning in the Dark is generated in a blind and therefore privacy-preserving way by utilizing the properties of homomorphic encryption.

Over the past few years, a tremendous growth of machine learning was brought about by a significant increase in adoption of cloud-based services. As a result, various solutions have been proposed in which the machine learning models run on a remote cloud provider. However, when such a model is deployed on an untrusted cloud, it is of vital importance that the users' privacy is preserved. To this end, we propose Blind Faith - a machine learning model in which the training phase occurs in plaintext data, but the classification of the users' inputs is performed on homomorphically encrypted ciphertexts. To make our construction compatible with homomorphic encryption, we approximate the activation functions using Chebyshev polynomials. This allowed us to build a privacy-preserving machine learning model that can classify encrypted images. Blind Faith preserves users' privacy since it can perform high accuracy predictions by performing computations directly on encrypted data.

Social Networks' omnipresence and ease of use has revolutionized the generation and distribution of information in today's world. However, easy access to information does not equal an increased level of public knowledge. Unlike traditional media channels, social networks also facilitate faster and wider spread of disinformation and misinformation. Viral spread of false information has serious implications on the behaviours, attitudes and beliefs of the public, and ultimately can seriously endanger the democratic processes. Limiting false information's negative impact through early detection and control of extensive spread presents the main challenge facing researchers today. In this survey paper, we extensively analyze a wide range of different solutions for the early detection of fake news in the existing literature. More precisely, we examine Machine Learning (ML) models for the identification and classification of fake news, online fake news detection competitions, statistical outputs as well as the advantages and disadvantages of some of the available data sets. Finally, we evaluate the online web browsing tools available for detecting and mitigating fake news and present some open research challenges.

Social networking and micro-blogging services, such as Twitter, play an important role in sharing digital information. Despite the popularity and usefulness of social media, they are regularly abused by corrupt users. One of these nefarious activities is so-called fake news - a virus that has been spreading rapidly thanks to the hospitable environment provided by social media platforms. The extensive spread of fake news is now becoming a major problem with far-reaching negative repercussions on both individuals and society. Hence, the identification of fake news on social media is a problem of utmost importance that has attracted the interest not only of the research community but most of the big players on both sides - such as Facebook, on the industry side, and political parties on the societal one. In this work, we create a model through which we hope to be able to offer a solution that will instill trust in social network communities. Our model analyses the behaviour of 50,000 politicians on Twitter and assigns an influence score for each evaluated user based on several collected and analysed features and attributes. Next, we classify political Twitter users as either trustworthy or untrustworthy using random forest and support vector machine classifiers. An active learning model has been used to classify any unlabeled ambiguous records from our dataset. Finally, to measure the performance of the proposed model, we used accuracy as the main evaluation metric.

NoSQL-based databases are attractive to store and manage big data mainly due to high scalability and data modeling flexibility. However, security in NoSQL-based databases is weak which raises concerns for users. Specifically, security of data at rest is a high concern for the users deployed their NoSQL-based solutions on the cloud because unauthorized access to the servers will expose the data easily. There have been some efforts to enable encryption for data at rest for NoSQL databases. However, existing solutions do not support secure query processing, and data communication over the Internet and performance of the proposed solutions are also not good. In this article, the authors address NoSQL data at rest security concern by introducing a system which is capable to dynamically encrypt/decrypt data, support secure query processing, and seamlessly integrate with any NoSQL- based database. The proposed solution is based on a combination of chaotic encryption and Order Preserving Encryption (OPE). The experimental evaluation showed excellent results when integrated the solution with MongoDB and compared with the state-of-the-art existing work.

Online voting systems are gaining acceptance with the widespread use of secure web services and cloud computing such as electronic currency and online banking. However, they still face privacy, security and accountability issues. Designing a system that covers all the general requirements of secure voting is a research challenge. In this paper, we propose a secure online voting protocol based on a partially homomorphic encryption scheme. Our protocol ensures the anonymity of voters while preserving the integrity of the results. Experiments show the viability of our protocol in terms of both security and scalability.

Large propagation delay, high error rate, low band-width and limited energy in Underwater Sensor Networks (UWSNs) attract the attention of most researchers. In UWSNs, efficient utilization of energy is one of the major issue, as the replacement of energy sources in such environment is very expensive. In this paper, we have proposed a Cluster Depth Based Routing (cDBR) that is based on existing Depth Based Routing (DBR) protocol. In DBR, routing is based on the depth of the sensor nodes: the nodes having less depth are used as a forward nodes and consumes more energy as compared to the rest of nodes. As a result, nodes nearer to sink dies first because of more load. In cDBR, cluster based approach is used. In order to minimize the energy consumption, load among all the nodes are distributed equally. The energy consumption of each node is equally utilized as each node has equal probability to be selected as a Cluster Head (CH). This improves the stability period of network from DBR. In cDBR Cluster Heads (CHs) are used for forwarding packets that maximizes throughput of the network. We have compared our results with DBR and Energy Efficient DBR (EEDBR). The simulation result validates that cDBR achieves better stability period and high throughput comparatively to DBR and EEDBR.

Under water sensor networks(UWSNs) have attracted significantly to explore natural and undersea resources and gathering scientific data in aqueous conditions. The adverse characteristics in UWSNs communication and high cost limit the sensor nodes to spare deployment, causing delay, low propagation, power efficiency and floating node mobility. This proposed protocol is developed to handle these problems in under water sensor networks, two static sinks and four courier nodes are used to perform routing. Sensor nodes select their appropriate nearby static courier node to forward their data towards destination. Courier nodes have maximum energy, as compare to sensor nodes causing to enhance the network life time and provide equal distribution of energy consumption resulting to provide maximum throughput and stability of the network. Network field is hundred by hundred and providing maximum rounds through which we can closely over view the network life time, energy efficiency and throughput. Simulation results show maximum packet delivery per round. Courier nodes have maximum energy so the maximum routing will be performed by the courier nodes and sensor nodes will only sense their data and forward it to courier nodes causing to minimize the destabilization period of the network. Simulation results provide maximum throughput, minimum dead versus alive nodes and equal energy consumption per round.

We propose a new routing protocol MobiSink (mobile sink) for underwater sensor networks (UWSNs). We deploy the sink mobility in four horizontal regions of the network. The mobile sink moves in its own region to collect data from the transmission range sensor nodes. The transmission range of a node is calculated after fixed interval of time for mobile sink. In MobiSink nodes also take help of transmission range neighbors to communicate with sink cooperatively, if sink is out of range. The mobility pattern of sink and cooperative routing achieved better results as compared with other depth based routing protocols. The MobiSink scheme is validated via simulation, which shows better performance compared with depth based routing (DBR) and energy efficient depth based routing (EEDBR) protocols in terms of network life time, throughput and energy consumption.

In this paper, we overcome the problem of energy holes in UWSNs while considering the unique characteristics of underwater communication. In proposed scheme we consider UWSNs where nodes are manually deployed according to the defined deployment pattern to satisfy our application requirements in terms of energy saving. We used mixed routing technique i.e. direct transmission and hop-by-hop transmission for energy balancing in continuous monitoring applications for UWSNs. Sensor nodes forward the total data traffic (generated plus received) periodically to the sink with calculated load weights using variable communication ranges. The transmission ranges for 1-hop, 2-hop and direct transmission to the sink are used for data transmission to achieve load balancing for balanced energy consumption of all sensor nodes in the network. We prove that our scheme outperforms the existing selected schemes in terms of network lifetime and energy conservation. We select an optimal result from the simulation results for different possible combinations of transmissions.

In this paper, we propose a reliable and interference-aware routing protocol for underwater wireless sensor networks (UWSNs). Proposed protocol follows end-to-end path from source node to sink and selects next forwarder node of a data packet on the basis, having already established a path to sink. In this way, the problem of encounters void hole in depth based routing protocol is eliminated. Furthermore, during the selection of forwarding node, channel interference is also considered as routing metric to provide reliable communication. Therefore, proposed scheme reduces the probability of collision at the network layer, by selecting a neighbor node as the next forwarder of the data packet from the source node to the destination where the chance of channel interference is minimum. Simulation results verify the effectiveness of the proposed scheme in term of energy consumption, end-to-end delay and packet delivery ratio especially in a sparse network.