Biblio

Graph-based Semi-Supervised Learning (GSSL) is a practical solution to learn from a limited amount of labelled data together with a vast amount of unlabelled data. However, due to their reliance on the known labels to infer the unknown labels, these algorithms are sensitive to data quality. It is therefore essential to study the potential threats related to the labelled data, more specifically, label poisoning. In this paper, we propose a novel data poisoning method which efficiently approximates the result of label inference to identify the inputs which, if poisoned, would produce the highest number of incorrectly inferred labels. We extensively evaluate our approach on three classification problems under 24 different experimental settings each. Compared to the state of the art, our influence-driven attack produces an average increase of error rate 50% higher, while being faster by multiple orders of magnitude. Moreover, our method can inform engineers of inputs that deserve investigation (relabelling them) before training the learning model. We show that relabelling one-third of the poisoned inputs (selected based on their influence) reduces the poisoning effect by 50%. ACM Reference Format: Adriano Franci, Maxime Cordy, Martin Gubri, Mike Papadakis, and Yves Le Traon. 2022. Influence-Driven Data Poisoning in Graph-Based Semi-Supervised Classifiers. In 1st Conference on AI Engineering - Software Engineering for AI (CAIN’22), May 16–24, 2022, Pittsburgh, PA, USA. ACM, New York, NY, USA, 11 pages. https://doi.org/10.1145/3522664.3528606

This paper use the method of finite element analysis, and comparing and analyzing the split box and the integrated box from two aspects of modal analysis and static analysis. It is concluded that the integrated box has the characteristics of excellent vibration characteristics and high strength tolerance; At the same time, according to the S-N curve of the material and the load spectrum of the box, the fatigue life of the integrated box is 26.24 years by using the fatigue analysis software Fe-safe, which meets the service life requirements; The reliability analysis module PDS is used to calculate the reliability of the box, and the reliability of the integrated box is 96.5999%, which meets the performance requirements.

Cyber-Physical Systems (CPS) are complex systems of computational, physical, and human components integrated to achieve some function over one or more networks. The use of distributed simulation, or co-simulation, is one method often used to analyze the behavior and properties of these systems. High-Level Architecture (HLA) is an IEEE co-simulation standard that supports the development and orchestration of distributed simulations. However, a simple HLA federation constructed with the component simulations (i.e., federates) does not satisfy several requirements that arise in real-world use cases such as the shared use of limited physical and computational resources, the need to selectively hide information from participating federates, the creation of reusable federates and federations for supporting configurable shared services, achieving performant distributed simulations, organizing federations across different model types or application concerns, and coordinating federations across organizations with different information technology policies. This paper describes these core requirements that necessitate the use of multiple HLA federations and presents various mechanisms for constructing such integrated HLA federations. An example use case is implemented using a model-based rapid simulation integration framework called the Universal CPS Environment for Federation (UCEF) to illustrate these requirements and demonstrate techniques for integrating multiple HLA federations.

Even as Internet of Things (IoT) network security grows, concerns about the security of IoT devices have arisen. Although a few companies produce IP-connected gadgets for such ranging from small office, their security policies and implementations are often weak. They also require firmware updates or revisions to boost security and reduce vulnerabilities in equipment. A brownfield advance is necessary to verify systems where these helpless devices are present: putting in place basic security mechanisms within the system to render the system powerless possibly. Gadgets should cohabit without threatening their security in the same device. IoT network security has evolved into a platform that can segregate a large number of IoT devices, allowing law enforcement to compel the communication of defenseless devices in order to reduce the damage done by its unlawful transaction. IoT network security appears to be doable in well-known gadget types and can be deployed with minimum transparency.

The compressed sensing (CS) method can reconstruct images with a small amount of under-sampling data, which is an effective method for fast magnetic resonance imaging (MRI). As the traditional optimization-based models for MRI suffered from non-adaptive sampling and shallow” representation ability, they were unable to characterize the rich patterns in MRI data. In this paper, we propose a CS MRI method based on iterative shrinkage threshold algorithm (ISTA) and adaptive sparse sampling, called DSLS-ISTA-Net. Corresponding to the sampling and reconstruction of the CS method, the network framework includes two folders: the sampling sub-network and the improved ISTA reconstruction sub-network which are coordinated with each other through end-to-end training in an unsupervised way. The sampling sub-network and ISTA reconstruction sub-network are responsible for the implementation of adaptive sparse sampling and deep sparse representation respectively. In the testing phase, we investigate different modules and parameters in the network structure, and perform extensive experiments on MR images at different sampling rates to obtain the optimal network. Due to the combination of the advantages of the model-based method and the deep learning-based method in this method, and taking both adaptive sampling and deep sparse representation into account, the proposed networks significantly improve the reconstruction performance compared to the art-of-state CS-MRI approaches.

With the development of Internet of Things (IoT) technology, the transaction behavior of IoT devices has gradually increased, which also brings the problem of transaction data security and transaction processing efficiency. As one of the research hotspots in the field of data security, blockchain technology has been widely applied in the maintenance of transaction records and the construction of financial payment systems. However, the proportion of microtransactions in the Internet of Things poses challenges to the coupling of blockchain and IoT devices. This paper proposes a three-party scalable architecture based on “IoT device-edge server-blockchain”. In view of the characteristics of micropayment, the verification mechanism of the execution results of the off-chain transaction is designed, and the bridge node is designed in the off-chain architecture, which ensures the finality of the blockchain to the transaction. According to system evaluation, this scalable architecture improves the processing efficiency of micropayments on blockchain, while ensuring its decentration equal to that of blockchain. Compared with other blockchain-based IoT device payment schemes, our architecture is more excellent in activity.

Modular exponentiation, especially for very large integers of hundreds or thousands of bits, is a commonly used function in popular cryptosystems such as RSA. The complexity of this algorithm is partly driven by the very large word sizes, which require many - often millions - of primitive operations in a CPU implementation, or a large amount of logic when accelerated by an ASIC. FPGAs, with their many embedded DSP resources have started to be used as well. In almost all cases, the calculations have required multiple - occasionally many - clock cycles to complete. Recently, blockchain algorithms have required very low-latency implementations of modular multiplications, motivating new implementations and approaches.In this paper we show nine different high performance modular exponentiation for 1024-bit operands, using a 1024-bit modular multiplication as it’s core. Rather than just showing a number of completed designs, our paper shows the evolution of architectures which lead to different resource mix options. This will allow the reader to apply the examples to different FPGA targets which may have differing ratios of logic, memory, and embedded DSP blocks. In one design, we show a 1024b modular multiplier requiring 83K ALMs and 2372 DSPs, with a delay of 21.21ns.

The correctness of user traffic forwarding paths is an important goal of trusted transmission. Many network security issues are related to it, i.e., denial-of-service attacks, route hijacking, etc. The current path-aware network architecture can effectively overcome this issue through path verification. At present, the main problems of path verification are high communication and high computation overhead. To this aim, this article proposes a lightweight real-time verification mechanism of intradomain forwarding paths in autonomous systems to achieve a path verification architecture with no communication overhead and low computing overhead. The problem situation is that a packet finally reaches the destination, but its forwarding path is inconsistent with the expected path. The expected path refers to the packet forwarding path determined by the interior gateway protocols. If the actual forwarding path is different from the expected one, it is regarded as an incorrect forwarding path. This article focuses on the most typical intradomain routing environment. A few routers are set as the verification routers to block the traffic with incorrect forwarding paths and raise alerts. Experiments prove that this article effectively solves the problem of path verification and the problem of high communication and computing overhead.

Contrary to previous opinion, ‘frost shattering’ is not the only major contributor to rock weathering at mid latitudes and high elevations, more specifically along edges of bedrock escarpments. Lightning is also a significant contributor to land surface denudation. We can show this as lightning strikes on outcrops can dramatically alter the magnetic signature of rocks and is one of the main sources of noise in paleomagnetic studies. Igneous rocks in the highlands of Lesotho, southern Africa (\textgreater 3000 m elevation) provide an ideal study location, as flow lavas remain as prominent ridges that are relatively resistant to weathering. It is well known that lightning strikes can cause large remanent magnetization in rocks with little resultant variation in susceptibility. At two adjoining peaks in the Lesotho highlands, mapped freshly fractured rock correlates with areas of high magnetic intensity (remanent component), but little variation in susceptibility (related to the induced field), and is therefore a clear indicator of lightning damage. The majority of these mapped strike sites occur at the edges of topographic highs. Variations in magnetic intensity are correlated with the much lower resolution national lightning strikes dataset. These data confirm that high elevation edges of peak scarps are the focus of previous lightning strikes. This method of magnetic surveying compared with lightning strike data is a new method of confirming the locations of lightning strikes, and reduces the need for intensive paleomagnetic studies of the area to confirm remanence.

The botnet-based network assault are one of the most serious security threats overlay the Internet this day. Although significant progress has been made in this region of research in recent years, it is still an ongoing and challenging topic to virtually direction the threat of botnets due to their continuous evolution, increasing complexity and stealth, and the difficulties in detection and defense caused by the limitations of network and system architectures. In this paper, we propose a novel and efficient botnet detection method, and the results of the detection method are validated with the CTU-13 dataset.

We introduce AdaMix, an adaptive differentially private algorithm for training deep neural network classifiers using both private and public image data. While pre-training language models on large public datasets has enabled strong differential privacy (DP) guarantees with minor loss of accuracy, a similar practice yields punishing trade-offs in vision tasks. A few-shot or even zero-shot learning baseline that ignores private data can outperform fine-tuning on a large private dataset. AdaMix incorporates few-shot training, or cross-modal zero-shot learning, on public data prior to private fine-tuning, to improve the trade-off. AdaMix reduces the error increase from the non-private upper bound from the 167–311% of the baseline, on average across 6 datasets, to 68-92% depending on the desired privacy level selected by the user. AdaMix tackles the trade-off arising in visual classification, whereby the most privacy sensitive data, corresponding to isolated points in representation space, are also critical for high classification accuracy. In addition, AdaMix comes with strong theoretical privacy guarantees and convergence analysis.

State of the art Artificial Intelligence Assurance (AIA) methods validate AI systems based on predefined goals and standards, are applied within a given domain, and are designed for a specific AI algorithm. Existing works do not provide information on assuring subjective AI goals such as fairness and trustworthiness. Other assurance goals are frequently required in an intelligent deployment, including explainability, safety, and security. Accordingly, issues such as value loading, generalization, context, and scalability arise; however, achieving multiple assurance goals without major trade-offs is generally deemed an unattainable task. In this manuscript, we present two AIA pipelines that are model-agnostic, independent of the domain (such as: healthcare, energy, banking), and provide scores for AIA goals including explainability, safety, and security. The two pipelines: Adversarial Logging Scoring Pipeline (ALSP) and Requirements Feedback Scoring Pipeline (RFSP) are scalable and tested with multiple use cases, such as a water distribution network and a telecommunications network, to illustrate their benefits. ALSP optimizes models using a game theory approach and it also logs and scores the actions of an AI model to detect adversarial inputs, and assures the datasets used for training. RFSP identifies the best hyper-parameters using a Bayesian approach and provides assurance scores for subjective goals such as ethical AI using user inputs and statistical assurance measures. Each pipeline has three algorithms that enforce the final assurance scores and other outcomes. Unlike ALSP (which is a parallel process), RFSP is user-driven and its actions are sequential. Data are collected for experimentation; the results of both pipelines are presented and contrasted.

In this paper, a new model-free adaptive sliding mode load frequency control (LFC) scheme is designed for inter-connected power systems, where modeling is difficult and suffers from load change disturbances and denial of service (DoS) attacks. The proposed algorithm only uses real-time I/O data of the power system to achieve a high control performance. Firstly, the dynamic linearization strategy is used to build a data-based model of the power system, and intermittent DoS attacks are modeled by limiting their duration and frequency. Secondly, the model-free adaptive sliding mode control (MFASMC) scheme is designed based on optimization theory and sliding mode reaching law, and its stability is analyzed. Finally, the three-area interconnected power system was selected to test the presented MFASMC scheme. Simulation data shows the effectiveness of the LFC algorithm in this paper.

As cyber threats become highly damaging and complex, a new cybersecurity compliance certification model has been developed by the Department of Defense (DoD) to secure its Defense Industrial Base (DIB), and communication with its private partners. These partners or contractors are obligated by the Defense Federal Acquisition Regulations (DFARS) to be compliant with the latest standards in computer and data security. The Cybersecurity Maturity Model Certification (CMMC), and it is built upon existing DFARS 252.204-7012 and the NIST SP 800–171 controls. As of 2020, the DoD has incorporated DFARS and the National Institute of Standards and Technology (NIST) recommended security practices into what is now the CMMC. This paper presents the most commonly identified Security-Control-Deficiencies (SCD) faced, the attacks mitigated by addressing these SCD, and remediations applied to 127 DoD contractors in order to bring them into compliance with the CMMC guidelines. An analysis is done on what vulnerabilities are most prominent in the companies, and remediations applied to ensure these vulnerabilities are better avoided and the DoD supply-chain is more secure from attacks.

Data or information are being transferred at an enormous pace and hence protecting and securing this transmission of data are very important and have been very challenging. Cryptography and Steganography are the most broadly used techniques for safeguarding data by encryption of data and hiding the existence of data. A multi-layered secure transmission can be achieved by combining Cryptography with Steganography and by adding message authentication ensuring the confidentiality of the message. Different approach towards Steganography implementation is proposed using rotations and flips to prevent detection of encoded messages. Compression of multimedia files is set up for increasing the speed of encoding and consuming less storage space. The HMAC (Hash-based Authentication Code) algorithm is chosen for message authentication and integrity. The performance of the proposed Steganography methods is concluded using Histogram comparative analysis. Simulations have been performed to back the reliability of the proposed method.

This paper has a new network security evaluation method as an absorbing Markov chain-based assessment method. This method is different from other network security situation assessment methods based on graph theory. It effectively refinement issues such as poor objectivity of other methods, incomplete consideration of evaluation factors, and mismatching of evaluation results with the actual situation of the network. Firstly, this method collects the security elements in the network. Then, using graph theory combined with absorbing Markov chain, the threat values of vulnerable nodes are calculated and sorted. Finally, the maximum possible attack path is obtained by blending network asset information to determine the current network security status. The experimental results prove that the method fully considers the vulnerability and threat node ranking and the specific case of system network assets, which makes the evaluation result close to the actual network situation.

With the development of industrial informatization, information security in the power production industry is becoming more and more important. In the power production industry, as the critical information egress of the industrial control system, the information security of the Networked Control System is particularly important. This paper proposes a construction method for an information security platform of Networked Control System, which is used for research, testing and training of Networked Control System information security.

Most existing deep neural networks (DNNs) are inexplicable and fragile, which can be easily deceived by carefully designed adversarial example with tiny undetectable noise. This allows attackers to cause serious consequences in many DNN-assisted scenarios without human perception. In the field of speaker recognition, the attack for speaker recognition system has been relatively mature. Most works focus on white-box attacks that assume the information of the DNN is obtainable, and only a few works study gray-box attacks. In this paper, we study blackbox attacks on the speaker recognition system, which can be applied in the real world since we do not need to know the system information. By combining the idea of transferable attack and query attack, our proposed method NMI-FGSM-Tri can achieve the targeted goal by misleading the system to recognize any audio as a registered person. Specifically, our method combines the Nesterov accelerated gradient (NAG), the ensemble attack and the restart trigger to design an attack method that generates the adversarial audios with good performance to attack blackbox DNNs. The experimental results show that the effect of the proposed method is superior to the extant methods, and the attack success rate can reach as high as 94.8% even if only one query is allowed.

Hash-based data structures and algorithms are currently flourishing on the Internet. It is an effective way to store large amounts of information, especially for applications related to measurement, monitoring and security. At present, there are many hash table algorithms such as: Cuckoo Hash, Peacock Hash, Double Hash, Link Hash and D-left Hash algorithm. However, there are still some problems in these hash table algorithms, such as excessive memory space, long insertion and query operations, and insertion failures caused by infinite loops that require rehashing. This paper improves the kick-out mechanism of the Cuckoo Hash algorithm, and proposes a new hash table structure- Odd-Even Hash (OE Hash) algorithm. The experimental results show that OE Hash algorithm is more efficient than the existing Link Hash algorithm, Linear Hash algorithm, Cuckoo Hash algorithm, etc. OE Hash algorithm takes into account the performance of both query time and insertion time while occupying the least space, and there is no insertion failure that leads to rehashing, which is suitable for massive data storage.

This paper presents a study on the "Dynamic Load Altering Attacks" (D-LAAs), their effects on the dynamics of a transmission network, and provides a robust control protection scheme, based on polytopic uncertainties, invariance theory, Lyapunov arguments and graph theory. The proposed algorithm returns an optimal Energy Storage Systems (ESSs) placement, that minimizes the number of ESSs placed in the network, together with the associated control law that can robustly stabilize against D-LAAs. The paper provides a contextualization of the problem and a modelling approach for power networks subject to D-LAAs, suitable for the designed robust control protection scheme. The paper also proposes a reference scenario for the study of the dynamics of the control actions and their effects in different cases. The approach is evaluated by numerical simulations on large networks.

Distributed ledger technologies (DLTs) based on Directed Acyclic Graphs (DAGs) have been gaining much attention due to their performance advantage over the traditional blockchain. IOTA is an example of DAG-based DLT that has shown its significance in the Internet of Things (IoT) environment. Despite that, IOTA is vulnerable to double-spend attacks, which threaten the immutability of the ledger. In this paper, we propose an efficient yet simple method for detecting a parasite chain, which is one form of attempting a double-spend attack in the IOTA network. In our method, a score function measuring the importance of each transaction in the IOTA network is employed. Any abrupt change in the importance of a transaction is reflected in the 1st and 2nd order derivatives of this score function, and therefore used in the calculation of an anomaly score. Due to how the score function is formulated, this anomaly score can be used in the detection of a particular type of parasite chain, characterized by sudden changes in the in-degree of a transaction in the IOTA graph. The experimental results demonstrate that the proposed method is accurate and linearly scalable in the number of edges in the network.

This document addresses the issue of the actual security level of PDF documents. Two types of detection approaches are utilized to detect dangerous elements within malware: static analysis and dynamic analysis. Analyzing malware binaries to identify dangerous strings, as well as reverse-engineering is included in static analysis for t1he malware to disassemble it. On the other hand, dynamic analysis monitors malware activities by running them in a safe environment, such as a virtual machine. Each method has its own set of strengths and weaknesses, and it is usually best to employ both methods while analyzing malware. Malware detection could be simplified without sacrificing accuracy by reducing the number of malicious traits. This may allow the researcher to devote more time to analysis. Our worry is that there is no obvious need to identify malware with numerous functionalities when it isn't necessary. We will solve this problem by developing a system that will identify if the given file is infected with malware or not.

Many organizations use internal phishing campaigns to gauge awareness and coordinate training efforts based on those findings. Ongoing content design is important for phishing training tools due to the influence recency has on phishing susceptibility. Traditional approaches for content development require significant investment and can be prohibitively costly, especially during the requirements engineering phase of software development and for applications that are constantly evolving. While prior research primarily depends upon already known phishing cues curated by experts, our project, Phish Finders, uses crowdsourcing to explore phishing cues through the unique perspectives and thought processes of everyday users in a realistic yet safe online environment, Zooniverse. This paper contributes qualitative analysis of crowdsourced comments that identifies novel cues, such as formatting and typography, which were identified by the crowd as potential phishing indicators. The paper also shows that crowdsourcing may have the potential to scale as a requirements engineering approach to meet the needs of content labeling for improved training tool development.

Inertial Confinement Fusion(ICF) uses the inertia of the substance itself to confine the nest-temperature thermonuclear fuel plasma to achieve thermonuclear fusion and obtain fusion energy. In the design of the local-volume ignition target capsule, the ignition zone and the main combustion zone are separated by heavy medium. The ignition zone is located in the center of the system (the part of the fusion combustion). The mass is small and can be compressed to high density and the overall temperature is raised to the ignition state (local-volume ignition). The temperature increase and density increase of the local volume ignition are relatively decoupled in time. The multi-step enhanced shock wave heats the fuel temperature drop, after which the collision effect accelerates the metal shell layer by layer, and uses the inertia of high-Z metal shell with a larger residual mass to achieve effective compression of the fuel areal after the driving source ends for a long time. Local volume ignition has the advantages of no need to reshape the radiation driving pulse, resistance to the influence of hot electrons, less demanding compression symmetry, and large combustion gain.

The Manufacturer Usage Description (MUD) standard aims to reduce the attack surface for IoT devices by locking down their behavior to a formally-specified set of network flows (access control entries). Formal network behaviors can also be systematically and rigorously verified in any operating environment. Enforcing MUD flows and monitoring their activity in real-time can be relatively effective in securing IoT devices; however, its scope is limited to endpoints (domain names and IP addresses) and transport-layer protocols and services. Therefore, misconfigured or compromised IoTs may conform to their MUD-specified behavior but exchange unintended (or even malicious) contents across those flows. This paper develops PicP-MUD with the aim to profile the information content of packet payloads (whether unencrypted, encoded, or encrypted) in each MUD flow of an IoT device. That way, certain tasks like cyber-risk analysis, change detection, or selective deep packet inspection can be performed in a more systematic manner. Our contributions are twofold: (1) We analyze over 123K network flows of 6 transparent (e.g., HTTP), 11 encrypted (e.g., TLS), and 7 encoded (e.g., RTP) protocols, collected in our lab and obtained from public datasets, to identify 17 statistical features of their application payload, helping us distinguish different content types; and (2) We develop and evaluate PicP-MUD using a machine learning model, and show how we achieve an average accuracy of 99% in predicting the content type of a flow.