Biblio
In AI Matters Volume 4, Issue 2, and Issue 4, we raised the notion of the possibility of an AI Cosmology in part in response to the "AI Hype Cycle" that we are currently experiencing. We posited that our current machine learning and big data era represents but one peak among several previous peaks in AI research in which each peak had accompanying "Hype Cycles". We associated each peak with an epoch in a possible AI Cosmology. We briefly explored the logic machines, cybernetics, and expert system epochs. One of the objectives of identifying these epochs was to help establish that we have been here before. In particular we've been in the territory where some application of AI research finds substantial commercial success which is then closely followed by AI fever and hype. The public's expectations are heightened only to end in disillusionment when the applications fall short. Whereas it is sometimes somewhat of a challenge even for AI researchers, educators, and practitioners to know where the reality ends and hype begins, the layperson is often in an impossible position and at the mercy of pop culture, marketing and advertising campaigns. We suggested that an AI Cosmology might help us identify a single standard model for AI that could be the foundation for a common shared understanding of what AI is and what it is not. A tool to help the layperson understand where AI has been, where it's going, and where it can't go. Something that could provide a basic road map to help the general public navigate the pitfalls of AI Hype.
K-anonymity is a popular model used in microdata publishing to protect individual privacy. This paper introduces the idea of ball tree and projection area density partition into k-anonymity algorithm.The traditional kd-tree implements the division by forming a super-rectangular, but the super-rectangular has the area angle, so it cannot guarantee that the records on the corner are most similar to the records in this area. In this paper, the super-sphere formed by the ball-tree is used to address this problem. We adopt projection area density partition to increase the density of the resulting recorded points. We implement our algorithm with the Gotrack dataset and the Adult dataset in UCI. The experimentation shows that the k-anonymity algorithm based on ball-tree and projection area density partition, obtains more anonymous groups, and the generalization rate is lower. The smaller the K is, the more obvious the result advantage is. The result indicates that our algorithm can make data usability even higher.
Expected and unexpected risks in cloud computing, which included data security, data segregation, and the lack of control and knowledge, have led to some dilemmas in several fields. Among all of these dilemmas, the privacy problem is even more paramount, which has largely constrained the prevalence and development of cloud computing. There are several privacy protection algorithms proposed nowadays, which generally include two categories, Anonymity algorithm, and differential privacy mechanism. Since many types of research have already focused on the efficiency of the algorithms, few of them emphasized the different orientation and demerits between the two algorithms. Motivated by this emerging research challenge, we have conducted a comprehensive survey on the two popular privacy protection algorithms, namely K-Anonymity Algorithm and Differential Privacy Algorithm. Based on their principles, implementations, and algorithm orientations, we have done the evaluations of these two algorithms. Several expectations and comparisons are also conducted based on the current cloud computing privacy environment and its future requirements.
In the process of informationization and networking of smart grids, the original physical isolation was broken, potential risks increased, and the increasingly serious cyber security situation was faced. Therefore, it is critical to develop accuracy and efficient anomaly detection methods to disclose various threats. However, in the industry, mainstream security devices such as firewalls are not able to detect and resist some advanced behavior attacks. In this paper, we propose a time series anomaly detection model, which is based on the periodic extraction method of discrete Fourier transform, and determines the sequence position of each element in the period by periodic overlapping mapping, thereby accurately describe the timing relationship between each network message. The experiments demonstrate that our model can detect cyber attacks such as man-in-the-middle, malicious injection, and Dos in a highly periodic network.
Sparse and low rank matrix decomposition is a method that has recently been developed for estimating different components of hyperspectral data. The rank component is capable of preserving global data structures of data, while a sparse component can select the discriminative information by preserving details. In order to take advantage of both, we present a novel decision fusion based on joint low rank and sparse component (DFJLRS) method for hyperspectral imagery in this paper. First, we analyzed the effects of different components on classification results. Then a novel method adopts a decision fusion strategy which combines a SVM classifier with the information provided by joint sparse and low rank components. With combination of the advantages, the proposed method is both representative and discriminative. The proposed algorithm is evaluated using several hyperspectral images when compared with traditional counterparts.
The rapid growth of Android malware apps poses a great security threat to users thus it is very important and urgent to detect Android malware effectively. What's more, the increasing unknown malware and evasion technique also call for novel detection method. In this paper, we focus on API feature and develop a novel method to detect Android malware. First, we propose a novel selection method for API feature related with the malware class. However, such API also has a legitimate use in benign app thus causing FP problem (misclassify benign as malware). Second, we further explore structure relationships between these APIs and map to a matrix interpreted as the hand-refined API-based feature graph. Third, a CNN-based classifier is developed for the API-based feature graph classification. Evaluations of a real-world dataset containing 3,697 malware apps and 3,312 benign apps demonstrate that selected API feature is effective for Android malware classification, just top 20 APIs can achieve high F1 of 94.3% under Random Forest classifier. When the available API features are few, classification performance including FPR indicator can achieve effective improvement effectively by complementing our further work.
Since the neural networks are utilized to extract information from an image, Gatys et al. found that they could separate the content and style of images and reconstruct them to another image which called Style Transfer. Moreover, there are many feed-forward neural networks have been suggested to speeding up the original method to make Style Transfer become practical application. However, this takes a price: these feed-forward networks are unchangeable because of their fixed parameters which mean we cannot transfer arbitrary styles but only single one in real-time. Some coordinated approaches have been offered to relieve this dilemma. Such as a style-swap layer and an adaptive normalization layer (AdaIN) and soon. Its worth mentioning that we observed that the AdaIN layer only aligns the means and variance of the content feature maps with those of the style feature maps. Our method is aimed at presenting an operational approach that enables arbitrary style transfer in real-time, reserving more statistical information by histogram matching, providing more reliable texture clarity and more humane user control. We achieve performance more cheerful than existing approaches without adding calculation, complexity. And the speed comparable to the fastest Style Transfer method. Our method provides more flexible user control and trustworthy quality and stability.
This paper presents an overview of the H2020 project VESSEDIA [9] aimed at verifying the security and safety of modern connected systems also called IoT. The originality relies in using Formal Methods inherited from high-criticality applications domains to analyze the source code at different levels of intensity, to gather possible faults and weaknesses. The analysis methods are mostly exhaustive an guarantee that, after analysis, the source code of the application is error-free. This paper is structured as follows: after an introductory section 1 giving some factual data, section 2 presents the aims and the problems addressed; section 3 describes the project's use-cases and section 4 describes the proposed approach for solving these problems and the results achieved until now; finally, section 5 discusses some remaining future work.
In monolithic operating system (OS), any error of system software can be exploit to destroy the whole system. The situation becomes much more severe in cloud environment, when the kernel and the hypervisor share the same address space. The security of guest Virtual Machines (VMs), both sensitive data and vital code, can no longer be guaranteed, once the hypervisor is compromised. Therefore, it is essential to deploy some security approaches to secure VMs, regardless of the hypervisor is safe or not. Some approaches propose microhypervisor reducing attack surface, or a new software requiring a higher privilege level than hypervisor. In this paper, we propose a novel approach, named HyperPS, which separates the fundamental and crucial privilege into a new trusted environment in order to monitor hypervisor. A pivotal condition for HyperPS is that hypervisor must not be allowed to manipulate any security-sensitive system resources, such as page tables, system control registers, interaction between VM and hypervisor as well as VM memory mapping. Besides, HyperPS proposes a trusted environment which does not rely on any higher privilege than the hypervisor. We have implemented a prototype for KVM hypervisor on x86 platform with multiple VMs running Linux. KVM with HyperPS can be applied to current commercial cloud computing industry with portability. The security analysis shows that this approach can provide effective monitoring against attacks, and the performance evaluation confirms the efficiency of HyperPS.
The Joint Test Action Group (JTAG) standards define test and debug architectures that are ingrained within much of today's commercial silicon. In particular, the IEEE Std. 1149.1 (Standard Test Access Port and Boundary Scan Architecture) forms the foundation of on-chip embedded instrumentation that is used extensively for everything from prototype board bring-up to firmware triage to field and depot system repair. More recently, JTAG is being used in-system as a hardware/firmware mechanism for Built-In Test (BIT), addressing No Fault Found (NFF) and materiel availability issues. Its power and efficacy are a direct outcome of being a ubiquitously available, embedded on-die instrument that is inherent in most electronic devices. While JTAG is indispensable for all aspects of test and debug, it suffers from a lack of inherent security. Unprotected, it can represent a security weakness, exposing a back-door vulnerability through which hackers can reverse engineer, extract sensitive data from, or disrupt systems. More explicitly, JTAG can be used to: - Read and write from system memory - Pause execution of firmware (by setting breakpoints) - Patch instructions or data in memory - Inject instructions directly into the pipeline of a target chip (without modifying memory) - Extract firmware (for reverse engineering/vulnerability research) - Execute private instructions to activate other engines within the chip As a low-level means of access to a powerful set of capabilities, the JTAG interface must be safeguarded against unauthorized intrusions and attacks. One method used to protect platforms against such attacks is to physically fuse off the JTAG Test Access Ports, either at the integrated circuit or the board level. But, given JTAG's utility, alternative approaches that allow for both security and debug have become available, especially if there is a hardware root of trust on the platform. These options include chip lock and key registers, challenge-response mechanisms, secure key systems, TDI/TDO encryption, and other authentication/authorization techniques. This paper reviews the options for safe access to JTAG-based debug and test embedded instrumentation.
The battlefield environment differs from the natural environment in terms of irregular communications and the possibility of destroying communication and medical units by enemy forces. Information that can be collected in a war environment by soldiers is important information and must reach top-level commanders in time for timely decisions making. Also, ambulance staff in the battlefield need to enter the data of injured soldiers after the first aid, so that the information is available for the field hospital staff to prepare the needs for incoming injured soldiers.In this research, we propose two transaction techniques to handle these issues and use different concurrency control protocols, depending on the nature of the transaction and not a one concurrency control protocol for all types of transactions. Message transaction technique is used to collect valuable data from the battlefield by soldiers and allows top-level commanders to view it according to their permissions by logging into the system, to help them make timely decisions. In addition, use the capabilities of DBMS tools to organize data and generate reports, as well as for future analysis. Medical service unit transactional workflow technique is used to provides medical information to the medical authorities about the injured soldiers and their status, which helps them to prepare the required needs before the wounded soldiers arrive at the hospitals. Both techniques handle the disconnection problem during transaction processing.In our approach, the transaction consists of four phases, reading, editing, validation, and writing phases, and its processing is based on the optimistic concurrency control protocol, and the rules of actionability that describe how a transaction behaves if a value-change is occurred on one or more of its attributes during its processing time by other transactions.
With the rapid development of the Internet, preserving the security of confidential data has become a challenging issue. An effective method to this end is to apply steganography techniques. In this paper, we propose an efficient steganography algorithm which applies edge detection and MPC algorithm for data concealment in digital images. The proposed edge detection scheme partitions the given image, namely cover image, into blocks. Next, it identifies the edge blocks based on the variance of their corner pixels. Embedding the confidential data in sharp edges causes less distortion in comparison to the smooth areas. To diminish the imposed distortion by data embedding in edge blocks, we employ LSB and MPC algorithms. In the proposed scheme, the blocks are split into some groups firstly. Next, a full tree is constructed per group using the LSBs of its pixels. This tree is converted into another full tree in some rounds. The resultant tree is used to modify the considered LSBs. After the accomplishment of the data embedding process, the final image, which is called stego image, is derived. According to the experimental results, the proposed algorithm improves PSNR with at least 5.4 compared to the previous schemes.
Signal processing in encrypted domain has become an important mean to protect privacy in an untrusted network environment. Due to the limitations of the underlying encryption methods, many useful algorithms that are sophisticated are not well implemented. Considering that QR decomposition is widely used in many fields, in this paper, we propose to implement QR decomposition in homomorphic encrypted domain. We firstly realize some necessary primitive operations in homomorphic encrypted domain, including division and open square operation. Gram-Schmidt process is then studied in the encrypted domain. We propose the implementation of QR decomposition in the encrypted domain by using the secure implementation of Gram-Schmidt process. We conduct experiments to demonstrate the effectiveness and analyze the performance of the proposed outsourced QR decomposition.
These days deep learning is the fastest-growing area in the field of Machine Learning. Convolutional Neural Networks are currently the main tool used for the image analysis and classification purposes. Although great achievements and perspectives, deep neural networks and accompanying learning algorithms have some relevant challenges to tackle. In this paper, we have focused on the most frequently mentioned problem in the field of machine learning, that is relatively poor generalization abilities. Partial remedies for this are regularization techniques e.g. dropout, batch normalization, weight decay, transfer learning, early stopping and data augmentation. In this paper we have focused on data augmentation. We propose to use a method based on a neural style transfer, which allows to generate new unlabeled images of high perceptual quality that combine the content of a base image with the appearance of another one. In a proposed approach, the newly created images are described with pseudo-labels, and then used as a training dataset. Real, labeled images are divided into the validation and test set. We validated proposed method on a challenging skin lesion classification case study. Four representative neural architectures are examined. Obtained results show the strong potential of the proposed approach.
As trust becomes increasingly important in the software domain. Due to its complex composite concept, people face great challenges, especially in today's dynamic and constantly changing internet technology. In addition, measuring the software trustworthiness correctly and effectively plays a significant role in gaining users trust in choosing different software. In the context of security, trust is previously measured based on the vulnerability time occurrence to predict the total number of vulnerabilities or their future occurrence time. In this study, we proposed a new unified index called "loss speed index" that integrates the most important variables of software security such as vulnerability occurrence time, number and severity loss, which are used to evaluate the overall software trust measurement. Based on this new definition, a new model called software trustworthy security growth model (STSGM) has been proposed. This paper also aims at filling the gap by addressing the severity of vulnerabilities and proposed a vulnerability severity prediction model, the results are further evaluated by STSGM to estimate the future loss speed index. Our work has several features such as: (1) It is used to predict the vulnerability severity/type in future, (2) Unlike traditional evaluation methods like expert scoring, our model uses historical data to predict the future loss speed of software, (3) The loss metric value is used to evaluate the risk associated with different software, which has a direct impact on software trustworthiness. Experiments performed on real software vulnerability datasets and its results are analyzed to check the correctness and effectiveness of the proposed model.
Using security primitives, a novel scheme for licensing hardware intellectual properties (HWIPs) on Field Programmable Gate Arrays (FPGAs) in public clouds is proposed. The proposed scheme enforces a pay-per-use model, allows HWIP's installation only on specific on-cloud FPGAs, and efficiently protects the HWIPs from being cloned, reverse engineered, or used without the owner's authorization by any party, including a cloud insider. It also provides protection for the users' designs integrated with the HWIP on the same FPGA. This enables cloud tenants to license HWIPs in the cloud from the HWIP vendors at a relatively low price based on usage instead of paying the expensive unlimited HWIP license fee. The scheme includes a protocol for FPGA authentication, HWIP secure decryption, and usage by the clients without the need for the HWIP vendor to be involved or divulge their secret keys. A complete prototype test-bed implementation showed that the proposed scheme is very feasible with relatively low resource utilization. Experiments also showed that a HWIP could be licensed and set up in the on-cloud FPGA in 0.9s. This is 15 times faster than setting up the same HWIP from outside the cloud, which takes about 14s based on the average global Internet speed.
Mobile phones have become nowadays a commodity to the majority of people. Using them, people are able to access the world of Internet and connect with their friends, their colleagues at work or even unknown people with common interests. This proliferation of the mobile devices has also been seen as an opportunity for the cyber criminals to deceive smartphone users and steel their money directly or indirectly, respectively, by accessing their bank accounts through the smartphones or by blackmailing them or selling their private data such as photos, credit card data, etc. to third parties. This is usually achieved by installing malware to smartphones masking their malevolent payload as a legitimate application and advertise it to the users with the hope that mobile users will install it in their devices. Thus, any existing application can easily be modified by integrating a malware and then presented it as a legitimate one. In response to this, scientists have proposed a number of malware detection and classification methods using a variety of techniques. Even though, several of them achieve relatively high precision in malware classification, there is still space for improvement. In this paper, we propose a text mining all repeated pattern detection method which uses the decompiled files of an application in order to classify a suspicious application into one of the known malware families. Based on the experimental results using a real malware dataset, the methodology tries to correctly classify (without any misclassification) all randomly selected malware applications of 3 categories with 3 different families each.