Biblio

Companies store increasing amounts of data, requiring the implementation of mechanisms to protect them from malicious people. There are techniques and procedures that aim to increase the security of computer systems, such as network protection services, firewalls. They are intended to filter packets that enter and leave a network. Its settings depend on security policies, which consist of documents that describe what is allowed to travel on the network and what is prohibited. The transcription of security policies into rules, written in native firewall language, that represent them, is the main source of errors in firewall configurations. In this work, concepts related to security between networks and firewalls are presented. Related works on security policies and their translations into firewall rules are also referenced. Furthermore, the developed tool, named Fireasy, is presented, which allows the modeling of security policies through graphic elements, and the maintenance of rules written in native firewall language, also representing them in graphic elements. Finally, a controlled experiment was conducted to validate the approach, which indicated, in addition to the correct functioning of the tool, an improvement in the translation of security policies into firewall rules using the tool. In the task of understanding firewall rules, there was a homogenization of the participants' performance when they used the tool.

In this paper, we design a new framework that can utilize the current global optimization heuristics for solving the straight-line program (SLP) problem. We combine Paar1, Paar2, BP (Boyar-Peralta), BFI, RNBP (Random-Boyar Peralta), A1, A2, XZLBZ, and LWFWSW (backward search) state-of-the-art heuristics by taking the XOR (exclusive OR) count metrics into consideration. Thus, by using the proposed framework, optimal circuit implementations of a given diffusion (or linear) layer can be found with fewer XOR gate counts.

In recent years, the security of AI systems has drawn increasing research attention, especially in the medical imaging realm. To develop a secure medical image analysis (MIA) system, it is a must to study possible backdoor attacks (BAs), which can embed hidden malicious behaviors into the system. However, designing a unified BA method that can be applied to various MIA systems is challenging due to the diversity of imaging modalities (e.g., X-Ray, CT, and MRI) and analysis tasks (e.g., classification, detection, and segmentation). Most existing BA methods are designed to attack natural image classification models, which apply spatial triggers to training images and inevitably corrupt the semantics of poisoned pixels, leading to the failures of attacking dense prediction models. To address this issue, we propose a novel Frequency-Injection based Backdoor Attack method (FIBA) that is capable of delivering attacks in various MIA tasks. Specifically, FIBA leverages a trigger function in the frequency domain that can inject the low-frequency information of a trigger image into the poisoned image by linearly combining the spectral amplitude of both images. Since it preserves the semantics of the poisoned image pixels, FIBA can perform attacks on both classification and dense prediction models. Experiments on three benchmarks in MIA (i.e., ISIC-2019 [4] for skin lesion classification, KiTS-19 [17] for kidney tumor segmentation, and EAD-2019 [1] for endoscopic artifact detection), validate the effectiveness of FIBA and its superiority over stateof-the-art methods in attacking MIA models and bypassing backdoor defense. Source code will be available at code.

Research done in Facial Privacy so far has entrenched the scope of gleaning race, age, and gender from a human’s facial image that are classifiable and compliant biometric attributes. Noticeable distortions, morphing, and face-swapping are some of the techniques that have been researched to restore consumers’ privacy. By fooling face recognition models, these techniques cater superficially to the needs of user privacy, however, the presence of visible manipulations negatively affects the aesthetic of the image. The objective of this work is to highlight common adversarial techniques that can be used to introduce granular pixel distortions using white-box and black-box perturbation algorithms that ensure the privacy of users’ sensitive or personal data in face images, fooling AI facial recognition models while maintaining the aesthetics of and visual integrity of the image.

Many organizations process and store classified data within their computer networks. Owing to the value of data that they hold; such organizations are more vulnerable to targets from adversaries. Accordingly, the sensitive organizations resort to an ‘air-gap’ approach on their networks, to ensure better protection. However, despite the physical and logical isolation, the attackers have successfully manifested their capabilities by compromising such networks; examples of Stuxnet and Agent.btz in view. Such attacks were possible due to the successful manipulation of human beings. It has been observed that to build up such attacks, persistent reconnaissance of the employees, and their data collection often forms the first step. With the rapid integration of social media into our daily lives, the prospects for data-seekers through that platform are higher. The inherent risks and vulnerabilities of social networking sites/apps have cultivated a rich environment for foreign adversaries to cherry-pick personal information and carry out successful profiling of employees assigned with sensitive appointments. With further targeted social engineering techniques against the identified employees and their families, attackers extract more and more relevant data to make an intelligent picture. Finally, all the information is fused to design their further sophisticated attacks against the air-gapped facility for data pilferage. In this regard, the success of the adversaries in harvesting the personal information of the victims largely depends upon the common errors committed by legitimate users while on duty, in transit, and after their retreat. Such errors would keep on repeating unless these are aligned with their underlying human behaviors and weaknesses, and the requisite mitigation framework is worked out.

Sybil attack is one of the most dangerous internal attacks in Vehicular Ad Hoc Network (VANET). It affects the function of the VANET network by maliciously claiming or stealing multiple identity propagation error messages. In order to prevent VANET from Sybil attacks, many solutions have been proposed. However, the existing solutions are specific to the physical or application layer's single-level data and lack research on cross-layer information fusion detection. Moreover, these schemes involve a large number of sensitive data access and transmission, do not consider users' privacy, and can also bring a severe communication burden, which will make these schemes unable to be actually implemented. In this context, this paper introduces FedMix, the first federated Sybil attack detection system that considers cross-layer information fusion and provides privacy protection. The system can integrate VANET physical layer data and application layer data for joint analyses simultaneously. The data resides locally in the vehicle for local training. Then, the central agency only aggregates the generated model and finally distributes it to the vehicles for attack detection. This process does not involve transmitting and accessing any vehicle's original data. Meanwhile, we also designed a new model aggregation algorithm called SFedAvg to solve the problems of unbalanced vehicle data quality and low aggregation efficiency. Experiments show that FedMix can provide an intelligent model with equivalent performance under the premise of privacy protection and significantly reduce communication overhead, compared with the traditional centralized training attack detection model. In addition, the SFedAvg algorithm and cross-layer information fusion bring better aggregation efficiency and detection performance, respectively.

The rapid growth of number of devices that are connected to internet of things (IoT) networks, increases the severity of security problems that need to be solved in order to provide safe environment for network data exchange. The discovery of new vulnerabilities is everyday challenge for security experts and many novel methods for detection and prevention of intrusions are being developed for dealing with this issue. To overcome these shortcomings, artificial intelligence (AI) can be used in development of advanced intrusion detection systems (IDS). This allows such system to adapt to emerging threats, react in real-time and adjust its behavior based on previous experiences. On the other hand, the traffic classification task becomes more difficult because of the large amount of data generated by network systems and high processing demands. For this reason, feature selection (FS) process is applied to reduce data complexity by removing less relevant data for the active classification task and therefore improving algorithm's accuracy. In this work, hybrid version of recently proposed sand cat swarm optimizer algorithm is proposed for feature selection with the goal of increasing performance of extreme learning machine classifier. The performance improvements are demonstrated by validating the proposed method on two well-known datasets - UNSW-NB15 and CICIDS-2017, and comparing the results with those reported for other cutting-edge algorithms that are dealing with the same problems and work in a similar configuration.

Today billions of people are accessing the internet around the world. There is a need for new technology to provide security against malicious activities that can take preventive/ defensive actions against constantly evolving attacks. A new generation of technology that keeps an eye on such activities and responds intelligently to them is the intrusion detection system employing machine learning. It is difficult for traditional techniques to analyze network generated data due to nature, amount, and speed with which the data is generated. The evolution of advanced cyber threats makes it difficult for existing IDS to perform up to the mark. In addition, managing large volumes of data is beyond the capabilities of computer hardware and software. This data is not only vast in scope, but it is also moving quickly. The system architecture suggested in this study uses SVM to train the model and feature selection based on the information gain ratio measure ranking approach to boost the overall system's efficiency and increase the attack detection rate. This work also addresses the issue of false alarms and trying to reduce them. In the proposed framework, the UNSW-NB15 dataset is used. For analysis, the UNSW-NB15 and NSL-KDD datasets are used. Along with SVM, we have also trained various models using Naive Bayes, ANN, RF, etc. We have compared the result of various models. Also, we can extend these trained models to create an ensemble approach to improve the performance of IDS.

This paper introduces an application of machine learning algorithms. In fact, support vector machine and decision tree approaches are studied and applied to compare their performances in detecting, classifying, and locating faults in the transmission network. The IEEE 14-bus transmission network is considered in this work. Besides, 13 types of faults are tested. Particularly, the one fault and the multiple fault cases are investigated and tested separately. Fault simulations are performed using the SimPowerSystems toolbox in Matlab. Basing on the accuracy score, a comparison is made between the proposed approaches while testing simple faults, on the one hand, and when complicated faults are integrated, on the other hand. Simulation results prove that the support vector machine technique can achieve an accuracy of 87% compared to the decision tree which had an accuracy of 53% in complicated cases.

Payment is an essential part of e-commerce. Merchants usually rely on third-parties, so-called payment processors, who take care of transferring the payment from the customer to the merchant. How a payment processor interacts with the customer and the merchant varies a lot. Each payment processor typically invents its own protocol that has to be integrated into the merchant’s application and provides the user with a new, potentially unknown and confusing user experience.Pushed by major companies, including Apple, Google, Master-card, and Visa, the W3C is currently developing a new set of standards to unify the online checkout process and “streamline the user’s payment experience”. The main idea is to integrate payment as a native functionality into web browsers, referred to as the Web Payment APIs. While this new checkout process will indeed be simple and convenient from an end-user perspective, the technical realization requires rather significant changes to browsers.Many major browsers, such as Chrome, Firefox, Edge, Safari, and Opera, already implement these new standards, and many payment processors, such as Google Pay, Apple Pay, or Stripe, support the use of Web Payment APIs for payments. The ecosystem is constantly growing, meaning that the Web Payment APIs will likely be used by millions of people worldwide.So far, there has been no in-depth security analysis of these new standards. In this paper, we present the first such analysis of the Web Payment APIs standards, a rigorous formal analysis. It is based on the Web Infrastructure Model (WIM), the most comprehensive model of the web infrastructure to date, which, among others, we extend to integrate the new payment functionality into the generic browser model.Our analysis reveals two new critical vulnerabilities that allow a malicious merchant to over-charge an unsuspecting customer. We have verified our attacks using the Chrome implementation and reported these problems to the W3C as well as the Chrome developers, who have acknowledged these problems. Moreover, we propose fixes to the standard, which by now have been adopted by the W3C and Chrome, and prove that the fixed Web Payment APIs indeed satisfy strong security properties.

The popularity of portable web browsers is increasing due to its convenient and compact nature along with the benefit of the data being stored and transferred easily using a USB drive. As technology gets updated frequently, developers are working on web browsers that can be portable in nature with additional security features like private mode browsing, built in ad blockers etc. The increased probability of using portable web browsers for carrying out nefarious activities is a result of cybercriminals with the thought that if they use portable web browsers in private mode it won't leave a digital footprint. Hence, the research paper aims at performing a comparative study of four portable web browsers namely Brave, TOR, Vivaldi, and Maxthon along with various memory acquisition tools to understand the quantity and quality of the data that can be recovered from the memory dump in two different conditions that is when the browser tabs were open and when the browser tabs were closed in a system to aid the forensic investigators.

The wide application of deep learning techniques is boosting the regulation of deep learning models, especially deep neural networks (DNN), as commercial products. A necessary prerequisite for such regulations is identifying the owner of deep neural networks, which is usually done through the watermark. Current DNN watermarking schemes, particularly white-box ones, are uniformly fragile against a family of functionality equivalence attacks, especially the neuron permutation. This operation can effortlessly invalidate the ownership proof and escape copyright regulations. To enhance the robustness of white-box DNN watermarking schemes, this paper presents a procedure that aligns neurons into the same order as when the watermark is embedded, so the watermark can be correctly recognized. This neuron alignment process significantly facilitates the functionality of established deep neural network watermarking schemes.

Vehicular Ad hoc Network (VANET) is an emerging technology that is used to provide communication between vehicle users. VANET provides communication between one vehicle node to another vehicle node, vehicle to the roadside unit, vehicle to pedestrian, and even vehicle to rail users. Communication between nodes should be very secure and confidential, Since VANET communicates through wireless mode, a malicious node may enter inside the communication zone to hack, inject false messages, and interrupt the communication. A strong protocol is necessary to detect malicious nodes and authenticate the VANET user to protect them from malicious attacks. In this paper, a fuzzy-based trust authentication scheme is used to detect malicious nodes with the Mamdani fuzzy Inference system. The parameter estimation, rules have been framed using MATLAB Mamdani Fuzzy Inference system to select a genuine node for data transmission.

Consensus is a basic building block in distributed systems for a myriad of related problems that involve agreement. For asynchronous networks, consensus has been proven impossible, and is well known as Augean task. Failure Detectors (FDs) have since emerged as a possible remedy, able to solve consensus in asynchronous systems under certain assumptions. With the increasing use of asynchronous, wireless Internet of Things (IoT) technologies, such as IEEE 802.15.4/6LoWPAN, the demand of applications that require some form of reliability and agreement is on the rise. What was missing so far is an FD that can operate under the tight constraints offered by Low Power and Lossy Networks (LLNs) without compromising the efficiency of the network. We present 6LoFD, an FD specifically aimed at energy and memory efficient operation in small scale, unreliable networks, and evaluate its working principles by using an ns-3 implementation of 6LoFD.

To improve efficiency and resource usage in data retrieval, an Internet of Things (IoT) search engine organizes a vast amount of scattered data and responds to client queries with processed results. Machine learning provides a deep understanding of complex patterns and enables enhanced feedback to users through well-trained models. Nonetheless, machine learning models are prone to adversarial attacks via the injection of elaborate perturbations, resulting in subverted outputs. Particularly, adversarial attacks on time-series data demand urgent attention, as sensors in IoT systems are collecting an increasing volume of sequential data. This paper investigates adversarial attacks on time-series analysis in an IoT search engine (IoTSE) system. Specifically, we consider the Long Short-Term Memory (LSTM) Recurrent Neural Network (RNN) as our base model, implemented in a simulated federated learning scheme. We propose the Federated Adversarial Learning for IoT Search Engine (FALIoTSE) that exploits the shared parameters of the federated model as the target for adversarial example generation and resiliency. Using a real-world smart parking garage dataset, the impact of an attack on FALIoTSE is demonstrated under various levels of perturbation. The experiments show that the training error increases significantly with noises from the gradient.

An ultra-low-power gesture and gait classification SoC is presented for rehabilitation application featuring (1) mixed-signal feature extraction and integrated low-noise amplifier eliminating expensive ADC and digital feature extraction, (2) an integrated distributed deep neural network (DNN) ASIC supporting a scalable multi-chip neural network for sensor fusion with distortion resiliency for low-cost front end modules, (3) onchip learning of DNN engine allowing in-situ training of user specific operations. A 12-channel 65nm CMOS test chip was fabricated with 1μW power per channel, less than 3ms computation latency, on-chip training for user-specific DNN model and multi-chip networking capability.

In this current age, stakeholders exchange legal documents, as well as documents that are official, sensitive and confidential via digital channels[1]. To securely communicate information between stakeholders is not an easy task considering the intentional or unintentional changes and possible attacks that can occur during communication. This paper focuses on protecting and securing data by hiding the data using steganography techniques, after encrypting the data to avoid unauthorized changes or modification made by adversaries to the data through using the Simplified Data Encryption Technique. By leveraging on these two approaches, secret data security intensifies to two levels and a steganography image of high quality is attained. Cryptography converts plaintext into cipher text (unreadable text); whereas steganography is the technique of hiding secret messages in other messages. First encryption of data is done using the Simplified Data Encryption Standard (S-DES) algorithm after which the message encrypted is embedded in the cover image by means of the Least Significant Bit (LSB) approach.

Recently, DDoS attack has developed rapidly and become one of the most important threats to the Internet. Traditional machine learning and deep learning methods can-not train a satisfactory model based on the data of a single client. Moreover, in the real scenes, there are a large number of devices used for traffic collection, these devices often do not want to share data between each other depending on the research and analysis value of the attack traffic, which limits the accuracy of the model. Therefore, to solve these problems, we design a DDoS attack detection model based on federated learning named FLDDoS, so that the local model can learn the data of each client without sharing the data. In addition, considering that the distribution of attack detection datasets is extremely imbalanced and the proportion of attack samples is very small, we propose a hierarchical aggregation algorithm based on K-Means and a data resampling method based on SMOTEENN. The result shows that our model improves the accuracy by 4% compared with the traditional method, and reduces the number of communication rounds by 40%.

Nowadays, source location privacy in underwater acoustic sensor networks (UASNs) has gained a lot of attention. The aim of source location privacy is to use specific technologies to protect the location of the source from being compromised. Among the many technologies available are fake packet technology, multi-path routing technology and so on. The fake packet technology uses a certain amount of fake packets to mask the transmission of the source packet, affecting the adversary's efficiency of hop-by-hop backtracking to the source. However, during the operation of the fake packet technology, the fake packet, and the source packet may interfere with each other. Focus on this, a fake packet time slot assignment-based source location privacy protection (FPTSA-SLP) scheme. The time slot assignment is adopted to avoid interference with the source packet. Also, a relay node selection method based on the handshake is further proposed to increase the diversity of the routing path to confuse the adversary. Compared with the comparison algorithm, the simulation results demonstrate that the proposed scheme has a better performance in safety time.

Fitbit Versa is the most popular of its predecessors and successors in the Fitbit faction. Increasingly data stored on these smart fitness devices, their linked applications and cloud datacenters are being used for criminal convictions. There is limited research for investigators on wearable devices and specifically exploring evidence identification and methods of extraction. In this paper we present our analysis of Fitbit Versa using Cellebrite UFED and MSAB XRY. We present a clear scope for investigation and data significance based on the findings from our experiments. The data recovery will include logical and physical extractions using devices running Android 9 and iOS 12, comparing between Cellebrite and XRY capabilities. This paper discusses databases and datatypes that can be recovered using different extraction and analysis techniques, providing a robust outlook of data availability. We also discuss the accuracy of recorded data compared to planned test instances, verifying the accuracy of individual data types. The verifiable accuracy of some datatypes could prove useful if such data was required during the evidentiary processes of a forensic investigation.

This paper presents a fully blind false data injection (FDI) attack against an industrial field-bus i.e. PROFINET that is widely used in Siemens distributed Input/Output (I/O) systems. In contrast to the existing academic efforts in the research community which assume that an attacker is already familiar with the target system, and has a full knowledge of what is being transferred from the sensors or to the actuators in the remote I/O module, our attack overcomes these strong assumptions successfully. For a real scenario, we first sniff and capture real time data packets (PNIO-RT) that are exchanged between the IO-Controller and the IO-Device. Based on the collected data, we create an I/O database that is utilized to replace the correct data with false one automatically and online. Our full attack-chain is implemented on a real industrial setting based on Siemens devices, and tested for two scenarios. In the first one, we manipulate the data that represents the actual sensor readings sent from the IO-Device to the IO-Controller, whereas in the second scenario we aim at manipulating the data that represents the actuator values sent from the IO-Controller to the IO-Device. Our results show that compromising PROFINET I/O systems in the both tested scenarios is feasible, and the physical process to be controlled is affected. Eventually we suggest some possible mitigation solutions to secure our systems from such threats.

Physical unclonable functions (PUFs) are used to create unique device identifiers from their inherent fabrication variability. Unstable readings and variation of the PUF response over time are key issues that limit the applicability of PUFs in real-world systems. In this project, we developed a fuzzy extractor (FE) to generate robust cryptographic keys from ReRAM-based PUFs. We tested the efficiency of the proposed FE using BCH and Polar error correction codes. We use ReRAM-based PUFs operating in pre-forming range to generate binary cryptographic keys at ultra-low power with an objective of tamper sensitivity. We investigate the performance of the proposed FE with real data using the reading of the resistance of pre-formed ReRAM cells under various noise conditions. The results show a bit error rate (BER) in the range of 10−5 for the Polar-codes based method when 10% of the ReRAM cell array is erroneous at Signal to Noise Ratio (SNR) of 20dB.This error rate is achieved by using helper data length of 512 bits for a 256 bit cryptographic key. Our method uses a 2:1 ratio for helper data and key, much lower than the majority of previously reported methods. This property makes our method more robust against helper data attacks.

The progression of cryptographic attacks in the ICT era doubtless leads to the development of new cryptographic algorithms and assessment, and evaluation of the existing ones. In this paper, the artificial intelligence application, through the fuzzy analytic hierarchy process (FAHP) implementation, is used to rank criteria and sub-criteria on which the algorithms are based to determine the most promising criteria and optimize their use. Out of fifteen criteria, security soundness, robustness and hardware failure distinguished as significant ones.

With the rapid advancement of the electrical grid, substation automation systems (SASs) have been developing continuously. However, with the introduction of advanced features, such as remote control, potential cyber security threats in SASs are also increased. Additionally, crucial components in SASs, such as protection relays, usually come from third-party vendors and may not be fully trusted. Untrusted devices may stealthily perform harmful or unauthorised behaviours which could compromise or damage SASs, and therefore, bring adverse impacts to the primary plant. Thus, it is necessary to detect abnormal behaviours from an untrusted device before it brings about catastrophic impacts. Anomaly detection techniques are suitable to detect anomalies in SASs as they only bring minimal side-effects to normal system operations. Many researchers have developed various machine learning algorithms and mathematical models to improve the accuracy of anomaly detection. However, without prudent feature selection, it is difficult to achieve high accuracy when detecting attacks launched from internal trusted networks, especially for stealthy message modification attacks which only modify message payloads slightly and imitate patterns of benign behaviours. Therefore, this paper presents choices of features which improve the accuracy of anomaly detection within SASs, especially for detecting “stealthy” attacks. By including two additional features, Boolean control data from message payloads and physical values from sensors, our method improved the accuracy of anomaly detection by decreasing the false-negative rate from 25% to 5% approximately.

This article discusses the problem of information security management in computer systems and describes the process of developing an algorithm that allows to determine measures to protect personal data. The organizational and technical measures formulated by the FSTEC are used as measures.