Defense Model Research Articles

Retraction Note: Defense model for preserving the wind turbine records in cloud using fog computing with coupling based cryptography

Retraction Note: Defense model for preserving the wind turbine records in cloud using fog computing with coupling based cryptography

Black-box attacks and defense for DNN-based power quality classification in smart grid

Machine learning (ML) models are widely used in smart grid, but they are vulnerable to adversarial examples that are maliciously crafted using the input data. Therefore, the use of these models in smart grid can cause significant damage. Although there are many studies on ML models attack and defense, there are few studies based on smart grid environment. In this paper, we first utilize deep neural network (DNN) to classify power quality (PQ) signals. Then an algorithm based on the Projected Gradient Descent (PGD) is proposed to attack the classification model, which employs multiple iterations to find the optimal perturbation. Finally, a defense model based on Multi-Level-Denoising Autoencoder (MLD-AE) is proposed to defend against the black-box attacks. The experimental results show that, compared with the Fast Gradient Sign Method (FGSM), the proposed black-box attack algorithm can significantly reduce the classification ability of the model with a smaller disturbance. Furthermore, compared with adversarial training defense methods, the proposed defense model can improve the power system’s ability to resist black-box attacks without reducing the ability of the classification model to classify clean PQ signals, and has excellent scalability.

Vulnerability assessment and federated intrusion detection of Air Taxi enabled smart cities

A Sustainable Smart City (SSC) has become the largest industrial sector that encompasses advanced cognitive, information, and computing methods for providing better big data handling to enhance the quality of life and urban operations with optimized resource utilization. Unmanned Aerial Vehicles (UAVs) are used as Air Taxis (ATs) in SSC, bypassing traffic congestion. However, the integration of ATs in SSC is very challenging, seeking high safety and security for the passengers. Such an autonomous system carries vital information, becoming vulnerable to furtive False Data Injection Attacks (FDIA), which have not received considerable research attention heretofore. To address the problem, this article establishes a model-free FDIA employing a recurrent deterministic policy gradient to exploit the data integrity of AT’s sensors, leading to collisions. Thereafter, a decentralized threat detection model, a deep optimized attentive federated aggregation learning is deployed, to counteract the threat and protect the data privacy of each AT. The suggested method is parallelizable and can accurately detect an unobservable FDIA in a distributed manner. The simulation findings show that the proposed attack–defense model investigates in-depth system’s vulnerabilities and tackles the risk with a higher detection rate and reduced computing cost while maintaining privacy preservation that surpasses modern approaches.

A Review of Artificial Intelligence in Security and Privacy: Research Advances, Applications, Opportunities, and Challenges

Artificial intelligence has the potential to address many societal, economic, and environmental challenges, but only if AI-enabled gadgets are kept secure. Many artificial intelligence (AI) models produced in recent years can be hacked by utilizing cutting-edge techniques. This issue has sparked intense research into adversarial AI to develop machine and deep learning models that can withstand various types of attacks. We provide a detailed summary of artificial intelligence in this paper to prove how adversarial attacks against AI applications can be mounted, covering topics such as confrontational knowledge and capabilities, existing methods for actually producing adversarial examples, and existing cyber defense models. In addition, we investigated numerous cyber countermeasures that could defend AI applications against these attacks and offered a systematic approach for demonstrating war strategies against machine learning and artificial intelligence. To safeguard AI applications, we emphasize the importance of understanding the intentions and methods of possible attackers. In the end, we list the biggest problems and most interesting research areas in the field of AI privacy and security.

Black box attack and network intrusion detection using machine learning for malicious traffic

Network intrusion detection systems (NIDS) play an important role in maintaining network security. In recent years, machine learning (ML) technology has been increasingly used in NIDS, increasing detection accuracy and aggravating network evasion risk. In order to detect the robustness of existing anomaly detection algorithms based on ML, we design and implement a black box attack method to evade network intrusion detection in this paper. The method includes two parts: first, we apply the Generative Adversarial Network (GAN) features to generate adversarial samples to generate data packets that can evade anomaly detection. Then, we insert the forged data packets into the original malicious traffic so that the anomaly detectors cannot detect the malicious traffic. We test on the latest intrusion detection datasets. Experimental results show that this method has a good evasion effect on all tested anomaly detectors, and the evasion rate of half of the anomaly detectors exceeds 90%. Through experiments, we find that the feature extraction of the existing anomaly detection algorithm has a loophole, so we propose a new defense model to make up for this loophole. Experiments confirm that the model can significantly reduce the evasion rate.

Compound adversarial examples in deep neural networks

Although deep learning has made great progress in many fields, they are still vulnerable to adversarial examples. Many methods for generating adversarial examples have been proposed, which either contain adversarial perturbation or patch. In this paper, we explore the method that creates compound adversarial examples including both perturbation and patch. We show that fusing two weak attack modes can produce more powerful adversarial examples, where the patch covers only 1% of the pixels at random location in the image, and the perturbation changes only by 2/255 in the original pixel value (scale to 0–1). For both targeted attack and untargeted attack, compound attack can improve the generative efficiency of adversarial examples, and can attain higher attack success rate with fewer iteration steps. The compound adversarial examples successfully attack the models with defensive mechanisms that previously can defend perturbation attack or patch attack. Furthermore, the compound adversarial examples show good transferability on normal trained classifiers and adversarial trained classifiers. Experimental results on a series of widely used classifiers and defense models show that the proposed compound adversarial examples have strong robustness, high effectiveness, and good transferability.

Strategic Planning Pitfalls in Society 5.0: A Systematic Literature Review

Although strategic planning relevance has its ups and downs, it has shown that it can always be a reference point to guide managers and management. This paper aims to clarify if the strategic planning pitfalls identified in the mid-90s still apply to today's reality. Also, to validate strategic planning's importance in society 5.0 while exploring its barriers from a dynamic capability’s perspective. A bibliometric analysis focused on strategic planning pitfalls and the use of indexed journals/citations to validate the study relevance. To achieve this purpose, this study performs a systematic literature review and content analysis of 68 articles from 37 journals. The main results of strategic planning research revealed: (1) The relevance of strategic planning in society 5.0, (2) the appliance of dynamic capabilities creating a defensive business model, and (3) the two essential dimensions which can impact most strategic planning, and (4) the new pitfalls in today’s context. Guided by dynamic capabilities theory, this paper seeks to promote insights related to the importance of strategic planning, adds value to the body of knowledge, and identifies contingency considerations that help explain the results on the relationship between strategic planning and the present context.
 
 Received: 1 July 2022 / Accepted: 29 August 2022 / Published: 2 September 2022

Toward feature space adversarial attack in the frequency domain

Recent researchers have shown that deep neural networks (DNNs) are vulnerable to adversarial exemplars, making them unsuitable for security-critical applications. Transferability of adversarial examples is crucial for attacking black-box models, which facilitates adversarial attacks in more practical scenarios. We propose a novel adversarial attack with high transferability. Unlike existing attacks that directly modify the input pixels, our attack is executed in the feature space. More specifically, we corrupt the abstract features by maximizing the feature distance between the adversarial example and clean images with a perceptual similarity network, inducing model misclassification. In addition, we apply a spectral transformation to the input, thus narrowing the search space in the frequency domain to enhance the transferability of adversarial examples. The disruption of crucial features in a specific frequency component achieves greater transferability. Extensive evaluations illustrate that our approach is easily compatible with many existing frameworks for transfer attacks and can significantly improve the baseline performance of black-box attacks. Moreover, we can obtain a higher fooling rate even if the model has a defense technique. We achieve a maximum black-box fooling rate of 61.70% on the defense model. Our work indicates that existing pixel space defense techniques are difficult to guarantee the robustness of the feature space, and the feature space from a frequency perspective is promising for developing more robust models.

Machine Learning to Improve Situational Awareness in Beyond Visual Range Air Combat

This article presents an artificial intelligence model using artificial neural networks that provide parameters to improve the situational awareness of a Beyond Visual Range (BVR) air combat pilot. In this combat modality, it is necessary to make decisions based on information from sensors, mainly radars. Furthermore, since information regarding enemy aircraft systems is sometimes unknown, pilots' decisions are usually based on beliefs regarding the opponent. The presented model proposes to deal with such characteristics, generating behaviors for entities represented in a constructive simulation environment, i.e., simulated people operating simulated systems. We created BVR air combat simulations between two aircraft, with only one missile each, through Latin Hypercube Sampling (LHS) to choose input variables to cover almost homogeneously all their ranges. The aircraft have similar behaviors, and their parameters may change only at the beginning of the simulation. The simulation environment generated ten thousand air combat scenarios, varying thirty-six input parameters, for the analysis proposed in the case study. From this data, we could create supervised machine learning models that substantially improve the BVR air combat pilot's situational awareness regarding offensive situations, in which the reference aircraft employs a missile against a target, or defensive positions, in contrast to when the same reference aircraft tries to avoid a possible enemy's missile launched in its direction. The offensive and defensive models were consistent with the accuracy of 0.930 and 0.924 and the F1-score of 0.717 and 0.678, respectively. Thus, the contribution of this work is to use machine learning algorithms to generate responses concerning the tactical state to improve the pilot's situational awareness and, therefore, the in-flight decision-making process.

Research on SQL Injection Attack and Defense Technology of Power Dispatching Data Network: Based on Data Mining

In the process of SQL injection attack and defense of power dispatching data network, in order to ensure the accuracy of identification and defense, it is often necessary to build a rule base. However, the scale of the temporarily constructed rule base is limited, which is easy to cause false positives and omissions. Based on the traditional defense, data mining technology is introduced to design a SQL injection attack and defense technology for power dispatching data network. First, the SQL injection attack is analyzed, and the attack flow diagram is obtained as the basis of attack identification. Using the ITFIDF algorithm in data mining technology, combined with the dataset distribution diagram of conventional words and sensitive characters, the SQL injection attack is detected. Design the SQL injection attack early warning mechanism, establish the SQL injection attack defense model, and optimize the workflow of the defense model. The performance test results show that compared with the traditional defense technology, the proposed technology has certain advantages in false alarm rate and running time.

Mitigation of Black-Box Attacks on Intrusion Detection Systems-Based ML

Intrusion detection systems (IDS) are a very vital part of network security, as they can be used to protect the network from illegal intrusions and communications. To detect malicious network traffic, several IDS based on machine learning (ML) methods have been developed in the literature. Machine learning models, on the other hand, have recently been proved to be effective, since they are vulnerable to adversarial perturbations, which allows the opponent to crash the system while performing network queries. This motivated us to present a defensive model that uses adversarial training based on generative adversarial networks (GANs) as a defense strategy to offer better protection for the system against adversarial perturbations. The experiment was carried out using random forest as a classifier. In addition, both principal component analysis (PCA) and recursive features elimination (Rfe) techniques were leveraged as a feature selection to diminish the dimensionality of the dataset, and this led to enhancing the performance of the model significantly. The proposal was tested on a realistic and recent public network dataset: CSE-CICIDS2018. The simulation results showed that GAN-based adversarial training enhanced the resilience of the IDS model and mitigated the severity of the black-box attack.

Towards Uncoated Stainless-Steel Bipolar Plates in Automotive PEM Fuel Cells

Bipolar plates (BPPs) have been identified as a key component in improving the economic viability of proton exchange membrane fuel cells (PEMFCs), as they contribute approximately 80 % of weight and 30 % of total cost in a fuel cell stack [1]. Stainless steel is considered an excellent material for mass-producing BPPs with complex flow field structures in a cost-effective manner. However, the perceived risk of corrosion and subsequent metal ion contamination as well as the inherently high contact resistance of stainless-steel surfaces have compelled manufacturers to add protective coatings to BPPs.In this study, we show new insights into the actual metal ion dissolution mechanisms in operating fuel cell systems and present experimental results which demonstrate that stainless-steel bipolar plates can be used in real life fuel cell applications without protective coatings. The bipolar plates were evaluated, in an operating fuel cell, using a dynamic load cycling protocol based on the New European Drive Cycle with increased humidity levels which simulates realistic conditions known to facilitate metal ion dissolution mechanisms [2]. The post analysis showed no signs of surface dissolution on any of the tested bipolar plates which had been subjected to the endurance/stress hybrid tests. Furthermore, only trace amounts of critical metals were found in the MEAs after testing, i.e., an order of magnitude below the radical scavenger (Mn) concentration. The trace metal contamination was comparable to a reference test without bipolar plates (graphite current collector), rendering metal input from the balance of plant the most likely source of contamination. The observable changes in cell performance and voltage degradation was unrelated to the presence of bipolar plates. A selection of experimental data is shown in Figure 1.As a theoretical explanation for the apparent stability of bare stainless-steel plates, three inter-linked phenomena were identified, akin to a three line of defence model [3]. The first and arguably most important of these phenomena is ionic decoupling between electrode and bipolar plate surface. The gas diffusion layer (GDL) strongly inhibits transport of ions from the electrode to the bipolar plate, which leads to a largely unpolarized bipolar plate and a stable surface potential even during load changes of the fuel cell. These stable potentials at the anode and cathode bipolar plates are only influenced by the immediate surroundings of the plate surfaces and are generally much milder than the encountered potentials at the electrodes. Secondly, metal dissolution from the surface of stainless steel follows a passive-transpassive-passive pattern. This mechanism describes a dissolution process, where a change in conditions at the surface, which destabilizes a previously established passive film, will only lead to limited amounts of (transpassive) metal dissolution, until a new stable passive film of different structure establishes itself. Finally, an operating regime which mitigates start/stop effects and hydrogen starvation during load changes will ensure very stable conditions at the bipolar plates’ surfaces, which strongly mitigates the risk of transpassive dissolution. As a final demonstration of the industrial feasibility of uncoated stainless-steel plates, an experimental alloy was tested under the same conditions. Results showed similar resistance to metal dissolution paired with low contact resistance at the level of C-coated or graphite plates. These results help to legitimize the efforts to industrialize uncoated stainless-steel plates since only the contact resistance challenge remains.[1] B. Pollet, S. Kocha, I. Staffell, ”Current status of automotive fuel cells for sustainable transport,” Current Opinion in Electrochemistry, vol. 16, pp. 90-95, 2019[2] T. Novalin, B. Eriksson, S. Proch, U. Bexell, C. Moffatt, J. Westlinder, C. Lagergren, G. Lindbergh, R. Wreland Lindström, “Trace-metal contamination in proton exchange membrane fuel cells caused by laser-cutting stains on carbon-coated metallic bipolar plates,” Int. J. Hydrogen Energy, Volume 46, Issue 26, 2021[3] T. Novalin, B. Eriksson, S. Proch, U. Bexell, C. Moffatt, J. Westlinder, C. Lagergren, G. Lindbergh, R. Wreland Lindström, “Concepts for preventing metal dissolution from stainless-steel bipolar plates in PEM fuel cells,” Energy Conversion and Management, 2022 Figure 1

LOGICDEF: An Interpretable Defense Framework against Adversarial Examples via Inductive Scene Graph Reasoning

Deep vision models have provided new capability across a spectrum of applications in transportation, manufacturing, agriculture, commerce, and security. However, recent studies have demonstrated that these models are vulnerable to adversarial attack, exposing a risk-of-use in critical applications where untrusted parties have access to the data environment or even directly to the sensor inputs. Existing adversarial defense methods are either limited to specific types of attacks or are too complex to be applied to practical vision models. More importantly, these methods rely on techniques that are not interpretable to humans. In this work, we argue that an effective defense should produce an explanation as to why the system is attacked, and by using a representation that is easily readable by a human user, e.g. a logic formalism. To this end, we propose logic adversarial defense (LogicDef), a defense framework that utilizes the scene graph of the image to provide a contextual structure for detecting and explaining object classification. Our framework first mines inductive logic rules from the extracted scene graph, and then uses these rules to construct a defense model that alerts the user when the vision model violates the consistency rules. The defense model is interpretable and its robustness is further enhanced by incorporating existing relational commonsense knowledge from projects such as ConceptNet. In order to handle the hierarchical nature of such relational reasoning, we use a curriculum learning approach based on object taxonomy, yielding additional improvements to training and performance.

Making Adversarial Examples More Transferable and Indistinguishable

Fast gradient sign attack series are popular methods that are used to generate adversarial examples. However, most of the approaches based on fast gradient sign attack series cannot balance the indistinguishability and transferability due to the limitations of the basic sign structure. To address this problem, we propose a method, called Adam Iterative Fast Gradient Tanh Method (AI-FGTM), to generate indistinguishable adversarial examples with high transferability. Besides, smaller kernels and dynamic step size are also applied to generate adversarial examples for further increasing the attack success rates. Extensive experiments on an ImageNet-compatible dataset show that our method generates more indistinguishable adversarial examples and achieves higher attack success rates without extra running time and resource. Our best transfer-based attack NI-TI-DI-AITM can fool six classic defense models with an average success rate of 89.3% and three advanced defense models with an average success rate of 82.7%, which are higher than the state-of-the-art gradient-based attacks. Additionally, our method can also reduce nearly 20% mean perturbation. We expect that our method will serve as a new baseline for generating adversarial examples with better transferability and indistinguishability.

Intrusion of Attacks in Puppet and Zombie Attacking and Defence Model Using BW-DDOS

The Internet is helpless against data transmission disseminated refusal of administration (BW-DDoS) assaults, wherein many hosts send countless parcels to cause clog and disturb authentic traffic. While adding a protection part against ill-disposed assaults, it is vital to convey different guard strategies paired to accomplish a decent inclusion of different assaults, BW-DDoS assaults have utilized generally rough, wasteful, beast force components; future assaults may be essentially more compelling and destructive. To meet the rising dangers, further developed guards are necessary. Distributed refusal of administration (DDoS) and adversarial assaults represent a genuine danger to the Internet. We examine the Internets vulnerability to Bandwidth Distributed Denial of Service (BW-DDoS) assaults, where many hosts send an immense number of bundles surpassing organization limits and causing clogs and misfortunes, along these lines disturbing genuine traffic. TCP and different conventions utilize clog control instruments that answer misfortunes and postponements by diminishing network usage, henceforth, their presentation might be debased pointedly because of such assaults. Aggressors might disturb connectivity to servers, organizations, independent frameworks, or entire nations or areas; such go after were at that point sent off in several conflicts. In this paper, we overview BW-DDoS assaults and safeguards. We contend that up until this point, BW-DDoS utilized relatively crude, wasteful, animal power systems; future assaults might be altogether more compelling, and thus much more destructive. We examine as of now sent and proposed protections. We contend that to meet the expanding threats, more progressed guards ought to be conveyed. This might include a few proposed systems (not yet conveyed), as well as new methodologies. This article is an outline and will be exceptionally compelling to per users who need to learn about data transmission DDoS assaults and guards.

Improving transferability of adversarial examples by saliency distribution and data augmentation

Although deep neural networks (DNNs) have advanced performance in many application scenarios, they are vulnerable to the attacks of adversarial examples that are crafted by adding imperceptible perturbations. Most of the existing adversarial attacks rely on the structure and parameters information of the attacked network. Thus, the generated adversarial examples are poor in transferability to attack black-box defense models, which makes them difficult to be used in real-world applications. In this paper, we propose an approach based on saliency distribution and data augmentation to generate transferable adversarial examples against the defense models. By optimizing perturbations over non-saliency regions, the generated adversarial examples are less sensitive to the attacked source models and have better transferability. Further, by utilizing data augmentation in generating adversarial examples, the overfitting problem on source models is alleviated in targeted attacks. Extensive experiments show that the proposed approach can generate adversarial examples with higher transferability. The source code is available at https://github.com/dongysxd/SDM-FGSM-Attack.

Continuous Authentication against Collusion Attacks.

As mobile devices become more and more popular, users gain many conveniences. It has also made smartphone makers install new software and prebuilt hardware on their products, including many kinds of sensors. With improved storage and computing power, users also become accustomed to storing and interacting with personally sensitive information. Due to convenience and efficiency, mobile devices use gait authentication widely. In recent years, protecting the information security of mobile devices has become increasingly important. It has become a hot research area because smartphones are vulnerable to theft or unauthorized access. This paper proposes a novel attack model called a collusion attack. Firstly, we study the imitation attack in the general state and its results and propose and verify the feasibility of our attack. We propose a collusion attack model and train participants with quantified action specifications. The results demonstrate that our attack increases the attacker’s false match rate only using an acceleration sensor in some systems sensor. Furthermore, we propose a multi-cycle defense model based on acceleration direction changes to improve the robustness of smartphone-based gait authentication methods against such attacks. Experimental results show that our defense model can significantly reduce the attacker’s success rate.

The Effect of the Three Lines of Defense Model on the Performance of State-Owned Enterprises Moderated by the Audit Committee

This study's objective is to examine the impact of adopting three lines of defense on the performance of State-Owned Enterprises (BUMN) in Indonesia, as moderated by an audit committee. The method employed is quantitative. In Indonesia, the population consists of State-Owned Enterprises. Practitioners, observers, and academics compose the sample. The utilized data is primary data. An online questionnaire was used to collect data. 128 (one hundred twenty-eight) individuals completed an online questionnaire that was sent via email and Whatsapp. The results indicate that line 1 (risk owner) has a positive effect on firm performance, line 2 (risk manager) has a negative effect on firm performance, and line 3 (Internal Audit Unit) has no effect on firm performance. The Audit Committee does not moderate the influence of line 3 (Internal Audit Unit) on firm performance, while it moderates the influence of line 2 (risk management unit) on firm performance.

Off-ball behavior in association football: A data-driven model to measure changes in individual defensive pressure

ABSTRACT This study describes an approach to evaluate the off-ball behaviour of attacking players in association football. The aim was to implement a defensive pressure model to examine an offensive player’s ability to create separation from a defender using 1411 high-intensity off-ball actions including 988 Deep Runs (DRs) DRs and 423 Change of Directions (CODs). Twenty-two official matches (14 competitive matches and 8 friendlies) of the German National Team were included in the research. To validate the effectiveness of the pressure model, each pass (n = 25,418) was evaluated for defensive pressure on the receiver at the moment of the pass and for the pass completion rate (R = −.34, p < .001). Next, after assessing the inter-rater reliability (Fleiss Kappa of 80 for DRs and 78 for CODs), three expert raters annotated all DRs and CODs that met the pre-set criteria. A time-series analysis of each DR and COD was calculated to the nearest 0.1 second, finding a slight increase in pressure from the start to the end of the off-ball actions as defenders re-established proximity to the attacker after separation was created. A linear mixed model using run type (DR or COD) as a fixed effect with the local maximum as a fixed effect on a continuous scale resulted in p < 0.001, d = 4.81, CI = 0.63 to 0.67 for the greatest decrease in pressure, p < 0.001, d = 0.143, CI = 9.18 to 10.61 for length of the longest decrease in pressure, and p < 0.001, d = 1.13, CI = 0.90 to 1.11 for the fastest rate of decrease in pressure. As these values pertain to the local maximum, situations with greater starting pressure on the attacker often led to greater subsequent decreases. Furthermore, there was a significant (p < .0001) difference between offensive and defensive positions and the number of off-ball actions. Results suggest the model can be applied to quantify and visualise the pressure exerted on non-ball-possessing players. This approach can be combined with other methods of match analysis, providing practitioners with new opportunities to measure tactical performance in football.

De-risking procurement by digitalising spend monitoring

Managing procurement risk through a conventional adoption of the three lines of defence model does not work effectively. This is primarily because digitalisation is restricted to the first line only through spend management systems, while the second and third lines rely on people armed with checklists undertaking periodic sample-based audits. This allows sub-optimal purchases, non-compliance, process abuse and fraud to slip through undetected and cause irreversible damage. A novel way forward would be to digitalise spend monitoring and create a second line of defence against such issues. Digitalising spend monitoring would entail several actions for it to make a significant dent in reducing procurement risk. The paper discusses how, first, an enterprise needs to study its past behaviour and identify what is abnormal and worth monitoring. Second, the enterprise needs to build an ability to correlate information from both within and outside its boundary so as to take the right decisions. Third, the enterprise needs to digitise a catalogue of controls that can be configured to continuously monitor and flag live transactions based on past and current observations. Fourth, the enterprise needs to build an ability to get actionable recommendations on anomalous transactions and intercept them immediately. Finally, the enterprise needs to ensure that this spend monitoring setup is continuously learning from its users’ actions and tuning itself to improve its accuracy. Meeting all the above criteria will ensure that such a system has a good signal-to-noise ratio and appeals to several stakeholders simultaneously within an enterprise. Global business service (GBS) owners and operations managers could use it to screen transactions at various approval tollgates in the procure-to-pay process and reduce the risk of waste, non-compliance and process abuse. Sourcing directors and category managers could use it to tune sourcing strategies and monitor their effectiveness to reduce supplier and category risk. Risk and compliance officers could use it to continuously audit 100 per cent of transactions and reduce the risk of non-compliance or fraud.