Intentional Attacks Research Articles

Vulnerability Analysis to Maximize the Resilience of Power Systems Considering Demand Response and Distributed Generation

Electric power systems are subject to failures, due to both deliberate and fortuitous events. This paper addresses the first case in which a disruptive agent aims at maximizing the damage to the network (expressed through the total cost of operation), while the system operator takes the necessary measures to mitigate the effects of this attack. The interaction between these two agents is modeled by means of a bi-level optimization problem. On one hand, the disruptive agent is positioned in the upper-level optimization problem and must decide which elements to render out of service (lines and generators), given a limited destructive budget. On the other hand, the system operator, located in a lower-level optimization problem, reacts to the attack by deploying mitigation measures in order to minimize cost overruns in system operation. Based on the aforementioned dynamic, this paper proposes a novel approach to maximize the resiliency of the power system under intentional attacks through the implementation of distributed energy resources (DERs), namely, distributed generation (DG) and demand response (DR). Three metrics are proposed to assess resilience by assigning DERs in islands generated by the destruction of lines and generators. The results obtained in a didactic 5-bus test system and the IEEE RTS-24 bus test system demonstrate the applicability and effectiveness of the proposed methodology.

The Integration of Protection, Restoration, and Adaptive Flow Redistribution in Building Resilient Networked Critical Infrastructures Against Intentional Attacks

It is paramount to ensure the resilience of networked critical infrastructures, which are vital for the economic development and our daily requirements but threatened by intentional attacks. For this purpose, their protection before and restoration after intentional attacks are of great significance. Moreover, in order to keep the prescribed demand continuously satisfied, postdisruption flow redistribution is also necessary. However, due to the severity of intentional attacks, certain components may need to be overloaded as a result of the redistribution and, thus, the network reliability is influenced. To minimize such influence, we first present a new metric of resilience to reflect the incurred reliability changes. Then, a trilevel optimization model integrating protection, restoration, and adaptive flow redistribution to maximize the network resilience is proposed. To deal with the complexity of the proposed problem, an evolutionary algorithm is employed with necessary adaption.

Exacerbating Algorithmic Bias through Fairness Attacks

Algorithmic fairness has attracted significant attention in recent years, with many quantitative measures suggested for characterizing the fairness of different machine learning algorithms. Despite this interest, the robustness of those fairness measures with respect to an intentional adversarial attack has not been properly addressed. Indeed, most adversarial machine learning has focused on the impact of malicious attacks on the accuracy of the system, without any regard to the system's fairness. We propose new types of data poisoning attacks where an adversary intentionally targets the fairness of a system. Specifically, we propose two families of attacks that target fairness measures. In the anchoring attack, we skew the decision boundary by placing poisoned points near specific target points to bias the outcome. In the influence attack on fairness, we aim to maximize the covariance between the sensitive attributes and the decision outcome and affect the fairness of the model. We conduct extensive experiments that indicate the effectiveness of our proposed attacks.

RETRACTED: Vulnerability analysis of road network for dangerous goods transportation considering intentional attack: Based on Cellular Automata

• The vulnerability of RNDGT under cascading failure considering intentional attack is analyzed. • The time characteristics of load distribution and node recovery ability are considered. • AD, BD, CD, DD, TD and SLD are selected to study the load re-distribution. • DA, BA and TA are selected to study the vulnerability. • CA based cascading failure model is established. The vulnerability of road network for dangerous goods transportation (RNDGT) under cascading failure considering intentional attack is analyzed. We introduce the time characteristics of load distribution and node recovery ability into previous cascading failure model, subdivide the state of failed node into normal state, partial failure state and complete failure state. Six traffic load distribution strategies including Average Distribution (AD), Betweenness Distribution (BD), Capacity Distribution (CD), Degree Distribution (DD), Tightness Distribution (TD) and Surplus Load Distribution (SLD) are selected to study the load re-distribution of failed nodes. In addition, three kinds of intentional node attack strategies including Degree Attack (DA), Betweenness Attack (BA) and Tightness Attack (TA) are selected to study the impact on the vulnerability. By referring the application of cellular automata applied in epidemic spreading field, we establish a new cascading failure model of RNDGT. The improved maximum connectivity and node failure rate based on node degree are applied to analyze the vulnerability. A case study is conducted by using the RNDGT of Dalian as the background. The previous Motter-Lai model (M-L) is applied as the comparison approach. TA strategy has the least impact on increasing network vulnerability, SLD strategy is the best to reduce network vulnerability.

Defender–attacker–operator: Tri-level game-theoretic interdiction analysis of urban water distribution networks

The aim of this paper is to advance the field of network interdiction analysis by introducing an application to the urban water distribution networks (WDNs), deploying protective resources against intentional attacks. The resource allocation problem for urban water supply systems is considered as a three-player (i.e., defender–attacker–operator) game, in which the attacker aims to maximize disruption impacts via interdicting water plants in the network, the defender aims to minimize the worst-case disruption impacts achieved by the attacker while the system operators fulfill the water demand in the residual urban water supply network. Considering the operating characteristics of the water supply network, we adopted the method of hydraulic analysis in the third level to obtain its reliability, and use this as the game equilibrium index of the first two levels. An effective modified variable neighborhood search method is devised to obtain the solution to the game. Finally, a case study was conducted based on the data of water supply network of a certain city in China to evaluate the effectiveness of protection resources against intentional attacks.

Theoretical framework about optimal model of operation after intentional attacks considering transmission system switching

This research focuses on optimal post-intentional attack operation considering switching transmission systems. The applicable models for this process focus on the application of bi-level optimization methods that are capable of analyzing two possible scenarios in order to reduce the time of loss or departure from the demand of the electricity system. The main objective of this work is related to maintaining the minimum requirements that allow the operation of the Electric Power System, for this the approach of equations will be carried out that allows establishing the mathematical models in the event of intentional attacks, it must maintain the operation of the Electric System and react in the event of contingencies through the Optimal Switching of Transmission Lines.

Stackelberg production-protection games: Defending crop production against intentional attacks

Inspired by recent terrorist attacks on cereal production fields in Iraq, we introduce and study two types of Stackelberg games. In these games, the leader wants to maximize its production (e.g., cereal), while the follower tries to destroy this production as much as possible. In the first model, the leader can protect its production by spreading his production resources over multiple regions. In the second model, the leader can also decide to allocate some extra protection resources to the regions. For both games, we are interested in a follower’s and leader’s optimal strategy. We characterise optimal strategies for the follower and present two linear time algorithms (one for each game) that find an optimal strategy for the leader.

Cascading failures of overload behaviors on interdependent networks

The increasing dependence between infrastructure networks and the threat of cascading failures on network security motivated the analysis of the overload behaviors on interdependent networks. Here, considering the node load redistribution mechanism, we study the robustness of interdependent networks with different attack strategies, connection strategies and load distribution mechanisms considering attack strength (r), coupling strength (q) and different network topologies (Erdős–Rényi network model and Albert–László Barabási network model). Compared to the different connection strategies, with q and qc, we conclude that the response of the cascade failure is more evident in the assorted connections which keep more concentrated load nodes. Meanwhile, by the proportional allocation of load in interdependent networks, we observe that the BA–BA networks with DC strategy (connect nodes with high load in network A and nodes with low load in network B) have the worst robust level against cascading failures in the case of the intentional attack. In addition, by considering the charge distribution mechanisms, we demonstrate that interdependent networks have the greatest capacity to withstand risk when the charge distribution preference (β) is closed at the initial charge of the nodes (τ). Theoretical solutions are useful for revealing cascading failures of overload behaviors in large-scale networks composed of multiple networks in the real world.

Protection of the patient data against intentional attacks using a hybrid robust watermarking code.

The security of patient information is important during the transfer of medical data. A hybrid spatial domain watermarking algorithm that includes encryption, integrity protection, and steganography is proposed to strengthen the information originality based on the authentication. The proposed algorithm checks whether the patient’s information has been deliberately changed or not. The created code is distributed at every pixel of the medical image and not only in the regions of non-interest pixels, while the image details are still preserved. To enhance the security of the watermarking code, SHA-1 is used to get the initial key for the Symmetric Encryption Algorithm. The target of this approach is to preserve the content of the image and the watermark simultaneously, this is achieved by synthesizing an encrypted watermark from one of the components of the original image and not by embedding a watermark in the image. To evaluate the proposed code the Least Significant Bit (LSB), Bit2SB, and Bit3SB were used. The evaluation of the proposed code showed that the LSB is of better quality but overall the Bit2SB is better in its ability against the active attacks up to a size of 2*2 pixels, and it preserves the high image quality.

Simulation-based vulnerability assessment in transit systems with cascade failures

The urban transit system is vulnerable to various types of hazards and failures, where the failure of a system component may lead to a redistribution of the pressure and result in cascade failures. It is of great significance to develop an efficient approach to model and assess the system vulnerability for achieving a sustainable transit system. This research develops a novel holistic approach to assess the vulnerability of the urban transit system toward sustainable development. The cascading failure mechanism is incorporated in the established complex network to simulate the process in a dependent system, where the failure of one or a few system components can trigger the failure of other components. Several attack strategies (i.e., static and dynamic intentional attack and random attack strategies) are used to test the system vulnerability, and the impacts of the tolerance parameter and passenger transfer rate on the loss of transport capacity are explored. A complex network consisting of 198 nodes and 415 edges is constructed to model the urban transit system with cascade failures. A realistic urban transit system in China is used as a case to testify the applicability and effectiveness of the developed approach. Results indicate that: (1) The interchange stations display high vulnerability in the network and Xianggang Rd station, Hongtu Boulevard station, Taipingyang station, and Zhongnan Rd are the most vulnerable stations, which should be paid more attention in the operation management; (2) The damage of the attacks in the intentional attack strategies are much more significant than that in the random attack, indicating that the proposed approach presents the ability to identify the important nodes; (3) The metric of node betweenness would be more efficient to identify the vulnerable nodes in the intentional attack; (4) The optimal value of the tolerance parameter is the range [0.8, 0.9], while keeping the passenger transfer rate in the range [0.3, 0.4] can effectively reduce the effect of cascade failures with minimum resources. The developed approach can be used as a decision tool to not only assess the vulnerability in the existing transit system but also optimize the planning of the new transit system.

A Power Dissipation Monitoring Circuit for Intrusion Detection and Botnet Prevention on IoT Devices

Recently, there has been a sharp increase in the production of smart devices and related networks, and consequently the Internet of Things. One concern for these devices, which is constantly becoming more critical, is their protection against attacks due to their heterogeneity and the absence of international standards to achieve this goal. Thus, these devices are becoming vulnerable, with many of them not even showing any signs of malfunction or suspicious behavior. The aim of the present work is to introduce a circuit that is connected in series with the power supply of a smart device, specifically an IP camera, which allows analysis of its behavior. The detection circuit operates in real time (real-time detection), sampling the supply current of the device, processing the sampled values and finally indicating any detection of abnormal activities, based on a comparison to normal operation conditions. By utilizing techniques borrowed by simple power analysis side channel attack, it was possible to detect deviations from the expected operation of the IP camera, as they occurred due to intentional attacks, quarantining the monitored device from the rest of the network. The circuit is analyzed and a low-cost implementation (under 5US$) is illustrated. It achieved 100% success in the test results, showing excellent performance in intrusion detection.

Community vulnerability perspective on robust protection planning in interdependent infrastructure networks

Critical infrastructure networks, including water, power, communication, and transportation, among others, are necessary to society’s functionality. In recent years, the threat of different types of disruptions to such infrastructure networks has become an increasingly important problem to address. Due to existing interdependencies, damage to a small area of one of the networks could have far-reaching effects on the ability to meet demand across the entire system. Common disruption scenarios include, among others, intentional malevolent attacks, natural disasters, and random failures. Similar works have focused on only one type of scenario, but combining a variety of disruptions may lead to more realistic results. Additionally, the concept of social vulnerability, which describes an area’s ability to prepare for and respond to a disruption, must be included. This should promote not only the protection of the most at-risk components but also ensure that socially vulnerable communities are given adequate resources. This work provides a decision making framework to determine the allocation of defensive resources that accounts for all these factors. Accordingly, we propose a multi-objective mathematical model with the objectives of: (i) minimizing the vulnerability of a system of interdependent infrastructure networks, and (ii) minimizing the total cost of the resource allocation strategy. Moreover, to account for uncertainty in the proposed model, this paper incorporates a means to address robustness in finding the most adaptable network protection plan to reduce the vulnerability of the system of interdependent networks to a variety of disruption scenarios. The proposed work is illustrated with an application to social vulnerability and interdependent power, gas, and water networks in Shelby County, Tennessee.

Compressed domain based robust digital video watermarking scheme to protect the copyright

<p><span>Digital video watermarking is an effective way to protect the ownership of the multimedia contents. A novel compressed domain based digital video watermarking algorithm scheme is proposed by exploiting MPEG-2 structure. Watermark bits are embedded in DC and AC coefficients both of only smooth discrete cosine transform (DCT) blocks from selected I-frames in the original digital video. The algorithms never exploited entire frames but explore only three location from the subset of DCT blocks from the subgroup of I-frames only. This process maintains the perceptibility of the watermarked video. Two parameters, normalized correlation (NC) and bit error rate (BER) are used to evaluate the degree of similarity and dissimilarity respectively to check the robustness against image processing and video specific intentional and non-intentional attacks. The security of embedded watermark is enhanced by applying three cryptographic keys. The experimental results demonstrated that the better robustness and perceptibility achieved by comparing the results with the state of art. </span></p>

Increasing robustness of deep neural network models against adversarial attacks

In Autonomous driving detecting correct object is important, further studies proved that by adding small pattern above object can also lead intentional fooling of network. Small intentional changes in the input can significantly distort output of a deep neural network model. This makes the machine learned model vulnerable to these small changes in images. Hence, these models have wide scope of failure. If we are able to tackle these intentional attacks it will help to make system more robust.In this project, we have combined multiple techniques used for defending against adversarial attacks. First technique is Adversarial training which include modifying training dataset, second technique is pre-processing input data before applying it to deep learned model and Third technique randomly selects image pre-processing technique. Third method is aimed to distract attacker who know the method used in pre-processing by randomly selecting from multiple methods in image pre-processing.We will measure robustness of deep learned model in terms of Accuracy on Designed system and previous deep learned model. Test samples and adversarial images generated from dataset will be used for testing on deep learned model.Among all methods which we have combined, Adversarial training proved best method to defend against white box attack. If we would have used strong defences in Random selection of image transformation then the system could have performed much better. However, Random selection have done its work of confusing the attacker by selecting random transformations.

A secure and robust color image watermarking using nature-inspired intelligence

The security of multimedia information is a prime concern in the present digital world. As a remedy, a robust color image watermarking in the transform domain using artificial intelligence is reported in this article. A color host image is secured by embedding a color watermark by utilizing the singular value decomposition and discrete wavelet transform. The color watermark information is scrambled with a chaotic map to give an additional stage of protection to overcome the problem of illegal copying and alteration of digital data by unauthorized users. All three RGB color channels of host singular values are modified with principal components of respective RGB information of scrambled watermark image. Ownership of data is provided by extracting the watermark in the presence of a secret key used while embedding process. To overcome the trade-off between the imperceptibility and robustness of the proposed algorithm artificial bee colony is used to optimize the scaling factor. The implementation of the proposed algorithm is performed on the MATLAB tool to compute the performance against the intentional and non-intentional image processing attacks. Comparative analysis with other related watermarking algorithms proves the robust, secure, and invisible nature of the proposed watermarking scheme.

Vulnerability modeling and assessment in urban transit systems considering disaster chains: A weighted complex network approach

This research develops a hybrid approach that integrates disaster chains and complex networks to model and assess vulnerability in urban transit systems. A weighted complex network is established by combining various disaster chains and their frequencies in the operation of the urban transit system, where disaster events and their transmission paths in chains are considered as nodes and edges in the disaster network, respectively. Several measures are proposed to quantify the vulnerability magnitude of the disaster nodes, edges, and network from both micro-level and macro-level perspectives. A weighted complex network consisting of 31 disaster nodes and 73 edges is constructed to model the disaster transmission in the operation of the metro system. Results show that (1) The steward misoperation (Node 15), the unqualified maintenance (Node 17), and the edge E38 (from Node 17 to Node 23) are identified as the most vulnerable nodes and edge, which should be paid more attention for disaster risk reduction; (2) The network efficiency is calculated to be 0.1244, acting as an anchor for improving system resilience against disasters; (3) The proposed edge vulnerability measure proves to be more effective in discovering more vulnerable edges than the traditional edge measure. Both node and edge attack analysis are performed to discuss the robustness of the developed complex network. It is found that all the intentional attack strategies display better performance for reducing network efficiency than random attack strategies, and global attack strategies are more efficient in both node and edge attack analysis than local attack strategies. This indicates that the developed approach is necessarily effective to identify the vulnerability of disaster events and links in the urban transit system, enabling to discover the best strategy for improving the resilience of the transit system against disasters.

A scenario-based simultaneous game approach for risk-averse facility protection problem

In this research, protecting facilities from intentional attacks under operational and disruption risks is modeled as a discrete scenario-based simultaneous game between a defender and an attacker, in which the information of their resource allocation is kept secret. In the presence of stochastic disruptions, a risk measure is incorporated into the defender. To deal with the proposed scenario-based simultaneous game in which the defender and the attacker have different objective functions, we develop an adaptive algorithm based on a maximum-likelihood sampling method to identify its mixed-strategy Nash equilibrium solution with respect to the conditional value-at-risk. In the experimental part, through comparative experiments, it is shown that the model of this article improves the decision-making benefit of the defender.

Retracted] Cascading Failure Dynamics against Intentional Attack for Interdependent Industrial Internet of Things

The emerging Industrial Internet of Things (IIoT) provides industries with an opportunity to collect, aggregate, and analyze data from sensors, including motion control, machine‐to‐machine communication, predictive maintenance, smart energy grid, big data analysis, and other smart connected medical systems. The physical systems and the cyber systems are organically integrated, forming an interdependent IIoT. This system provides us with enormous advantages, but at the same time, it also introduces the main safety challenges in the design and operation phase. To exploit the security threats of IIoT systems, in this paper, we propose a novel security‐by‐design approach for interdependent IIoT environments across two different levels, namely, theory modeling and runtime simulation. Our method theoretically analyzes the cascading failure dynamics of the intentional attack network. Simultaneously, we verified the theoretical results through simulations and gave the risk factors that affect the system’s security to mitigate potential security attack threats. Besides, we prove its applicability through comparative simulation experiments to study application environments that rely on IIoT, which shows that our method helps identify risk factors and mitigate IIoT attacks’ mechanism.

Power Grid‐Oriented Cascading Failure Vulnerability Identifying Method Based on Wireless Sensors

In our paper, we study the vulnerability in cascading failures of the real‐world network (power grid) under intentional attacks. Here, we use three indexes (B, K, k‐shell) to measure the importance of nodes; that is, we define three attacks, respectively. Under these attacks, we measure the process of cascade effect in network by the number of avalanche nodes, the time steps, and the speed of the cascade propagation. Also, we define the node’s bearing capacity as a tolerant parameter to study the robustness of the network under three attacks. Taking the power grid as an example, we have obtained a good regularity of the collapse of the network when the node’s affordability is low. In terms of time and speed, under the betweenness‐based attacks, the network collapses faster, but for the number of avalanche nodes, under the degree‐based attack, the number of the failed nodes is highest. When the nodes’ bearing capacity becomes large, the regularity of the network’s performances is not obvious. The findings can be applied to identify the vulnerable nodes in real networks such as wireless sensor networks and improve their robustness against different attacks.

Vulnerability Disclosure and Management for AI/ML Systems: A Working Paper with Policy Recommendations

Artificial intelligence systems, especially those dependent on machine learning (ML), can be vulnerable to intentional attacks that involve evasion, data poisoning, model replication, and exploitation of traditional software flaws to deceive, manipulate, compromise, and render them ineffective. Yet too many organizations adopting AI/ML systems are oblivious to their vulnerabilities. Applying the cybersecurity policies of vulnerability disclosure and management to AI/ML can heighten appreciation of the technologies’ vulnerabilities in real-world contexts and inform strategies to manage cybersecurity risk associated with AI/ML systems. Federal policies and programs to improve cybersecurity should expressly address the unique vulnerabilities of AI-based systems, and policies and structures under development for AI governance should expressly include a cybersecurity component.