Biblio

Exploring the intrinsic interconnections between the knowledge encoded in PRe-trained Deep Neural Networks (PR-DNNs) of heterogeneous tasks sheds light on their mutual transferability, and consequently enables knowledge transfer from one task to another so as to reduce the training effort of the latter. In this paper, we propose the DEeP Attribution gRAph (DEPARA) to investigate the transferability of knowledge learned from PR-DNNs. In DEPARA, nodes correspond to the inputs and are represented by their vectorized attribution maps with regards to the outputs of the PR-DNN. Edges denote the relatedness between inputs and are measured by the similarity of their features extracted from the PR-DNN. The knowledge transferability of two PR-DNNs is measured by the similarity of their corresponding DEPARAs. We apply DEPARA to two important yet under-studied problems in transfer learning: pre-trained model selection and layer selection. Extensive experiments are conducted to demonstrate the effectiveness and superiority of the proposed method in solving both these problems. Code, data and models reproducing the results in this paper are available at https://github.com/zju-vipa/DEPARA.

Threat Actor Attribution is the task of identifying an attacker responsible for an attack. This often requires expert analysis and involves a lot of time. There had been attempts to detect a threat actor using machine learning techniques that use information obtained from the analysis of malware samples. These techniques will only be able to identify the attack, and it is trivial to guess the attacker because various attackers may adopt an attack method. A state-of-the-art method performs attribution of threat actors from text reports using Machine Learning and NLP techniques using Threat Intelligence reports. We use the same set of Threat Reports of Advanced Persistent Threats (APT). In this paper, we propose a Deep Learning architecture to attribute Threat actors based on threat reports obtained from various Threat Intelligence sources. Our work uses Neural Networks to perform the task of attribution and show that our method makes the attribution more accurate than other techniques and state-of-the-art methods.

Security maintenance of Supervisory Control and Data Acquisition (SCADA) systems has been a point of interest during recent years. Numerous research works have been dedicated to the design of intrusion detection systems for securing SCADA communications. Nevertheless, these data-driven techniques are usually dependant on the quality of the monitored data. In this work, we propose a novel feature selection approach, called MSFS, to tackle undesirable quality of data caused by feature redundancy. In contrast to most feature selection techniques, the proposed method models each class in a different subspace, where it is optimally discriminated. This has been accomplished by resorting to ensemble learning, which enables the usage of multiple feature sets in the same feature space. The proposed method is then utilized to perform intrusion detection in smaller subspaces, which brings about efficiency and accuracy. Moreover, a comparative study is performed on a number of advanced feature selection algorithms. Furthermore, a dataset obtained from the SCADA system of a gas pipeline is employed to enable a realistic simulation. The results indicate the proposed approach extensively improves the detection performance in terms of classification accuracy and standard deviation.

A Supervisory Control and Data Acquisition (SCADA) system used in controlling or monitoring purpose in industrial process automation system is the process of collecting data from instruments and sensors located at remote sites and transmitting data at a central site. Most of the existing works on SCADA system focused on simulation-based study which cannot always mimic the real world situations. We propose a novel methodology that analyzes SCADA logs on offline basis and helps to detect process-related threats. This threat takes place when an attacker performs malicious actions after gaining user access. We conduct our experiments on a real-life SCADA system of a Power transmission utility. Our proposed methodology will automate the analysis of SCADA logs and systemically identify undesired events. Moreover, it will help to analyse process-related threats caused by user activity. Several test study suggest that our approach is powerful in detecting undesired events that might caused by possible malicious occurrence.

Embedded systems often consist of deeply intertwined hardware and software components. At the same time, they are often used in safety-critical applications, where an error may result in enormous costs or even loss of human lives. Existing verification techniques that show the absence of errors do not scale well for complex integrated HW/SW systems. In this paper, we present a dependence analysis and automated partitioning approach for the formal analysis of HW/SW codesigns that are modeled in SystemC. The key idea of our approach is threefold: first, we partition a given system into loosely coupled submodels. Second, we analyze the dependences between these submodels and compute an abstract verification interface for each of them, which captures all possible influences of all other submodels. Third, we verify global properties of the overall system by verifying them separately for each subsystem. We demonstrate that our approach significantly reduces verification times and increases scalability with results for an anti-lock braking system.

Data mining visualization is an important aspect of big data visualization and analysis. The impact of the nature-inspired algorithm along with the impact of computing traditions for the complete visualization of the storage and data communication needs have been studied. This paper also explores the possibilities of the hybridization of data mining in terms of association of cloud computing. It also explores the data analytical view in the exploration of these approaches in terms of data storage in big data. Based on these aspects the methodological advancement along with the problem statements has been analyzed. This will help in the exploration of computational capability along with the new insights in this domain.

Internet of Things devices and data sources areseeing increased use in various application areas. The pro-liferation of cheaper sensor hardware has allowed for widerscale data collection deployments. With increased numbers ofdeployed sensors and the use of heterogeneous sensor typesthere is increased scope for collecting erroneous, inaccurate orinconsistent data. This in turn may lead to inaccurate modelsbuilt from this data. It is important to evaluate this data asit is collected to determine its validity. This paper presents ananalysis of data quality as it is represented in Internet of Things(IoT) systems and some of the limitations of this representation. The paper discusses the use of trust as a heuristic to drive dataquality measurements. Trust is a well-established metric that hasbeen used to determine the validity of a piece or source of datain crowd sourced or other unreliable data collection techniques. The analysis extends to detail an appropriate framework forrepresenting data quality effectively within the big data modeland why a trust backed framework is important especially inheterogeneously sourced IoT data streams.

In many organizations, permissioned blockchain networks are currently transitioning from a proof-of-concept stage to production use. A crucial part of this transition is ensuring awareness of potential threats to network operations. Due to the plethora of software components involved in distributed ledgers, threats may be difficult or impossible to detect without a structured monitoring approach. To this end, we conduct a survey of attacks on permissioned blockchains and develop a set of threat indicators. To gather these indicators, a data processing pipeline is proposed to aggregate log information from relevant blockchain components, enriched with data from external sources. To evaluate the feasibility of monitoring current blockchain frameworks, we determine relevant data sources in Hyperledger Fabric. Our results show that the required data is mostly available, but also highlight significant improvement potential with regard to threat intelligence, chaincode scanners and built-in metrics.

One important aspect in protecting Cyber Physical System (CPS) is ensuring that the proper control and measurement signals are propagated within the control loop. The CPS research community has been developing a large set of check blocks that can be integrated within the control loop to check signals against various types of attacks (e.g., false data injection attacks). Unfortunately, it is not possible to integrate all these “checks” within the control loop as the overhead introduced when checking signals may violate the delay constraints of the control loop. Moreover, these blocks do not completely operate in isolation of each other as dependencies exist among them in terms of their effectiveness against detecting a subset of attacks. Thus, it becomes a challenging and complex problem to assign the proper checks, especially with the presence of a rational adversary who can observe the check blocks assigned and optimizes her own attack strategies accordingly. This paper tackles the inherent state-action space explosion that arises in securing CPS through developing DeepBLOC (DB)-a framework in which Deep Reinforcement Learning algorithms are utilized to provide optimal/sub-optimal assignments of check blocks to signals. The framework models stochastic games between the adversary and the CPS defender and derives mixed strategies for assigning check blocks to ensure the integrity of the propagated signals while abiding to the real-time constraints dictated by the control loop. Through extensive simulation experiments and a real implementation on a water purification system, we show that DB achieves assignment strategies that outperform other strategies and heuristics.

In general, Cloud storage is considered as a distributed model. Here, the data is usually stored on remote servers to properly maintain, back up and make it accessible to clients over a network, whenever required. Cloud storage providers keep the data and processes to oversee it on capacity servers based on secure virtualization methods. A security framework is proposed for auditing the cloud data, which makes use of the proposed blockchain technology. This ensures to efficiently maintain the data integrity. The blockchain structure inspects the mutation of operational information and thereby ensures the data security. Usually, the data auditing scheme is widely used in a Third Party Auditor (TPA), which is a centralized entity that the client is forced to trust, even if the credibility is not guaranteed. To avoid the participation of TPA, a decentralised scheme is suggested, where it uses a smart contract for auditing the cloud data. The working of smart contracts is based on blockchain. Ethereum is used to deploy a smart contract thereby eliminating the need of a foreign source in the data auditing process.

With the continuous progress of the information age, e-commerce, the Internet of things and other emerging Internet areas are gradually emerging. Massive amount of structured data auditing becomes a major issue. Log files and other data can be uploaded to the cloud via the Internet to guard against potential threats. Difficulty now is how to realize the data in the field of data audit query online, interactive and impromptu. There are two main methods of data warehouse, respectively is zhang table reduction method and basic data verification method. In the age of big data, data quantity increases gradually, so that the audit speed, design of the data storage and so on will be more or less problematic. If the audit task is not completed in time, it will result in the failure to store the audit data, which will cause losses to enterprises and the government. This paper focuses on the data cube physical model and distributed technical analysis, through the establishment of a set of efficient distributed and online auditing system, so as to make the data fast and efficient auditing.

The deep learning-based compressive Sensing (CS) has shown substantial improved performance and in run-time reduction with signal sampling and reconstruction. In most cases, moreover, these techniques suffer from disrupting artefacts or high-frequency contents at low sampling ratios. Similarly, this occurs in the multi-resolution sampling method, which further collects more components with lower frequencies. A promising innovation combining CS with convolutionary neural network has eliminated the sparsity constraint yet recovery persists slow. We propose a Deep wavelet based compressive sensing with multi-resolution framework provides better improvement in reconstruction as well as run time. The proposed model demonstrates outstanding quality on test functions over previous approaches.

In this paper, we propose a decision support process that is designed to help network and security operators in understanding the complexity of a current security situation and decision making concerning ongoing cyber-attacks and threats. The process focuses on enterprise missions and uses a graph-based mission decomposition model that captures the missions, underlying hosts and services in the network, and functional and security requirements between them. Knowing the vulnerabilities and attacker's position in the network, the process employs logical attack graphs and Bayesian network to infer the probability of the disruption of the confidentiality, integrity, and availability of the missions. Based on the probabilities of disruptions, the process suggests the most resilient mission configuration that would withstand the current security situation.

In recent days, cloud computing is one of the emerging fields. It is a platform to maintain the data and privacy of the users. To process and regulate the data with high security, the access control methods are used. The cloud environment always faces several challenges such as robustness, security issues and so on. Conventional methods like Cipher text-Policy Attribute-Based Encryption (CP-ABE) are reflected in providing huge security, but still, the problem exists like the non-existence of attribute revocation and minimum efficient. Hence, this research work particularly on the attribute-based mechanism to maximize efficiency. Initially, an objective coined out in this work is to define the attributes for a set of users. Secondly, the data is to be re-encrypted based on the access policies defined for the particular file. The re-encryption process renders information to the cloud server for verifying the authenticity of the user even though the owner is offline. The main advantage of this work evaluates multiple attributes and allows respective users who possess those attributes to access the data. The result proves that the proposed Data sharing scheme helps for Revocation under a fine-grained attribute structure.

Increasing penetration of distributed energy resources (DERs) in distribution networks expands the cyberattack surface. Moreover, the widely used standard protocols for communicating DER inverters such as Modbus is more vulnerable to data-integrity attacks and denial of service (DoS) attacks because of its native clear-text packet format. This paper proposes a distributed intrusion detection system (D-IDS) architecture and algorithms for detecting anomalies on the DER Modbus communication. We devised a model-based approach to define physics-based threshold bands for analog data points and transaction-based threshold bands for both the analog and discrete data points. The proposed IDS algorithm uses the model- based approach to develop Modbus-specific IDS rule sets, which can enhance the detection accuracy of the anomalies either by data-integrity attacks or maloperation on cyber-physical DER Modbus devices. Further, the IDS algorithm autogenerates the Modbus-specific IDS rulesets in compliance with various open- source IDS rule syntax formats, such as Snort and Suricata, for seamless integration and mitigation of semantic/syntax errors in the development and production environment. We considered the IEEE 13-bus distribution grid, including DERs, as a case study. We conducted various DoS type attacks and data-integrity attacks on the hardware-in-the-loop (HIL) CPS DER testbed at ISU to evaluate the proposed D-IDS. Consequently, we computed the performance metrics such as IDS detection accuracy, IDS detection rate, and end-to-end latency. The results demonstrated that 100% detection accuracy, 100% detection rate for 60k DoS packets, 99.96% detection rate for 80k DoS packets, and 0.25 ms end-to-end latency between DERs to Control Center.

We propose an Artificial Intelligence (AI)/Deep Learning (DL)-based image analysis framework for hardware assurance of digital integrated circuits (ICs). Our aim is to examine and verify various hardware information from analyzing the Scanning Electron Microscope (SEM) images of an IC. In our proposed framework, we apply DL-based methods at all essential steps of the analysis. To the best of our knowledge, this is the first such framework that makes heavy use of DL-based methods at all essential analysis steps. Further, to reduce time and effort required in model re-training, we propose and demonstrate various automated or semi-automated training data preparation methods and demonstrate the effectiveness of using synthetic data to train a model. By applying our proposed framework to analyzing a set of SEM images of a large digital IC, we prove its efficacy. Our DL-based methods are fast, accurate, robust against noise, and can automate tasks that were previously performed mainly manually. Overall, we show that DL-based methods can largely increase the level of automation in hardware assurance of digital ICs and improve its accuracy.

This letter presents a novel median filtering forensics approach, based on a convolutional neural network (CNN) with an adaptive filtering layer (AFL), which is built in the discrete cosine transform (DCT) domain. Using the proposed AFL, the CNN can determine the main frequency range closely related with the operational traces. Then, to automatically learn the multi-scale manipulation features, a multi-scale convolutional block is developed, exploring a new multi-scale feature fusion strategy based on the maxout function. The resultant features are further processed by a convolutional stream with pooling and batch normalization operations, and finally fed into the classification layer with the Softmax function. Experimental results show that our proposed approach is able to accurately detect the median filtering manipulation and outperforms the state-of-the-art schemes, especially in the scenarios of low image resolution and serious compression loss.

Many important security informatics problems require consideration of dynamical phenomena for their solution; examples include predicting the behavior of individuals in social networks and distinguishing malicious and innocent computer network activities based on activity traces. While information theory offers powerful tools for analyzing dynamical processes, to date the application of information-theoretic methods in security domains has focused on static analyses (e.g., cryptography, natural language processing). This paper leverages information-theoretic concepts and measures to quantify the similarity of pairs of stochastic dynamical systems, and shows that this capability can be used to solve important problems which arise in security applications. We begin by presenting a concise review of the information theory required for our development, and then address two challenging tasks: 1.) characterizing the way influence propagates through social networks, and 2.) distinguishing malware from legitimate software based on the instruction sequences of the disassembled programs. In each application, case studies involving real-world datasets demonstrate that the proposed techniques outperform standard methods.

Since remote ages, queues and delays have been a rather exasperating reality of human daily life. Today, they pursue us everywhere: in technical, social, socio-technical, and even control systems, dramatically deteriorating their performance. In this variety, it is the computer systems that are sure to cause the growing anxiety in our digital era. Although for our everyday Internet surfing, experiencing long-lasting and annoying delays is an unpleasant but not dangerous situation, for industrial control systems, especially those dealing with critical infrastructures, such behavior is unacceptable. The article presents a deterministic approach to solving some digital control system problems associated with delays and backlogs. Being based on Network calculus, in contrast to statistical methods of Queuing theory, it provides worst-case results, which are eminently desirable for critical infrastructures. The article covers the basics of a theory of deterministic queuing systems Network calculus, its evolution regarding the relationship between backlog bound and delay, and a technique for handling empirical data. The problems being solved by the deterministic approach: standard calculation of network performance measures, estimation of database maximum updating time, and cybersecurity assessment including such issues as the CIA triad representation, operational technology influence, and availability understanding focusing on its correlation with a delay are thoroughly discussed as well.

In the last couple of years, the move to cyberspace provides a fertile environment for ransomware criminals like ever before. Notably, since the introduction of WannaCry, numerous ransomware detection solution has been proposed. However, the ransomware incidence report shows that most organizations impacted by ransomware are running state of the art ransomware detection tools. Hence, an alternative solution is an urgent requirement as the existing detection models are not sufficient to spot emerging ransomware treat. With this motivation, our work proposes "DeepGuard," a novel concept of modeling user behavior for ransomware detection. The main idea is to log the file-interaction pattern of typical user activity and pass it through deep generative autoencoder architecture to recreate the input. With sufficient training data, the model can learn how to reconstruct typical user activity (or input) with minimal reconstruction error. Hence, by applying the three-sigma limit rule on the model's output, DeepGuard can distinguish the ransomware activity from the user activity. The experiment result shows that DeepGuard effectively detects a variant class of ransomware with minimal false-positive rates. Overall, modeling the attack detection with user-behavior permits the proposed strategy to have deep visibility of various ransomware families.

Today, Internet of Things (IoT) devices mostly operate in enclosed, proprietary environments. To unfold the full potential of IoT applications, a unifying and permissionless environment is crucial. All IoT devices, even unknown to each other, would be able to trade services and assets across various domains. In order to realize those applications, uniquely resolvable identities are essential. However, quantifiable trust in identities and their authentication are not trivially provided in such an environment due to the absence of a trusted authority. This research presents a new identity and trust framework for IoT devices, based on Distributed Ledger Technology (DLT). IoT devices assign identities to themselves, which are managed publicly and decentralized on the DLT's network as Self Sovereign Identities (SSI). In addition to the Identity Management System (IdMS), the framework provides a Web of Trust (WoT) approach to enable automatic trust rating of arbitrary identities. For the framework we used the IOTA Tangle to access and store data, achieving high scalability and low computational overhead. To demonstrate the feasibility of our framework, we provide a proof-of-concept implementation and evaluate the set objectives for real world applicability as well as the vulnerability against common threats in IdMSs and WoTs.

While internet technologies are developing day by day, threats against them are increasing at the same speed. One of the most serious and common types of attacks is Distributed Denial of Service (DDoS) attacks. The DDoS intrusion detection approach proposed in this study is based on fuzzy logic and entropy. The network is modeled as a graph and graphics-based features are used to distinguish attack traffic from non-attack traffic. Fuzzy clustering is applied based on these properties to indicate the tendency of IP addresses or port numbers to be in the same cluster. Based on this uncertainty, attack and non-attack traffic were modeled. The detection stage uses the fuzzy relevance function. This algorithm was tested on real data collected from Boğaziçi University network.

Human behaviors are often prohibited, or permitted by social norms. Therefore, if autonomous agents interact with humans, they also need to reason about various legal rules, social and ethical social norms, so they would be trusted and accepted by humans. Inverse Reinforcement Learning (IRL) can be used for the autonomous agents to learn social norm-compliant behavior via expert demonstrations. However, norms are context-sensitive, i.e. different norms get activated in different contexts. For example, the privacy norm is activated for a domestic robot entering a bathroom where a person may be present, whereas it is not activated for the robot entering the kitchen. Representing various contexts in the state space of the robot, as well as getting expert demonstrations under all possible tasks and contexts is extremely challenging. Inspired by recent work on Modularized Normative MDP (MNMDP) and early work on context-sensitive RL, we propose a new IRL framework, Context-Sensitive Norm IRL (CNIRL). CNIRL treats states and contexts separately, and assumes that the expert determines the priority of every possible norm in the environment, where each norm is associated with a distinct reward function. The agent chooses the action to maximize its cumulative rewards. We present the CNIRL model and show that its computational complexity is scalable in the number of norms. We also show via two experimental scenarios that CNIRL can handle problems with changing context spaces.

In this study, it was aimed to recognize the emotional state from facial images using the deep learning method. In the study, which was approved by the ethics committee, a custom data set was created using videos taken from 20 male and 20 female participants while simulating 7 different facial expressions (happy, sad, surprised, angry, disgusted, scared, and neutral). Firstly, obtained videos were divided into image frames, and then face images were segmented using the Haar library from image frames. The size of the custom data set obtained after the image preprocessing is more than 25 thousand images. The proposed convolutional neural network (CNN) architecture which is mimics of LeNet architecture has been trained with this custom dataset. According to the proposed CNN architecture experiment results, the training loss was found as 0.0115, the training accuracy was found as 99.62%, the validation loss was 0.0109, and the validation accuracy was 99.71%.

Avoiding security vulnerabilities is very important for embedded systems. Dynamic Information Flow Tracking (DIFT) is a powerful technique to analyze SW with respect to security policies in order to protect the system against a broad range of security related exploits. However, existing DIFT approaches either do not exist for Virtual Prototypes (VPs) or fail to model complex hardware/software interactions.In this paper, we present a novel approach that enables early and accurate DIFT of binaries targeting embedded systems with custom peripherals. Leveraging the SystemC framework, our DIFT engine tracks accurate data flow information alongside the program execution to detect violations of security policies at run-time. We demonstrate the effectiveness and applicability of our approach by extensive experiments.