Biblio

In this paper, stock selection strategy design based on machine learning and multi-factor analysis is a research hotspot in quantitative investment field. Four machine learning algorithms including support vector machine, gradient lifting regression, random forest and linear regression are used to predict the rise and fall of stocks by taking stock fundamentals as input variables. The portfolio strategy is constructed on this basis. Finally, the stock selection strategy is further optimized. The empirical results show that the multifactor quantitative stock selection strategy has a good stock selection effect, and yield performance under the support vector machine algorithm is the best. With the increase of the number of factors, there is an inverse relationship between the fitting degree and the yield under various algorithms.

With the advent of technology and owing to mankind’s reliance on technology, it is of utmost importance to safeguard people’s data and their identity. Biometrics have for long played an important role in providing that layer of security ranging from small scale uses such as house locks to enterprises using them for confidentiality purposes. In this paper we will provide an insight into behavioral biometrics that rely on identifying and measuring human characteristics or behavior. We review different types of behavioral parameters such as keystroke dynamics, gait, footstep pressure signals and more.

Sample-then-lock construction is a reusable fuzzy extractor for low-entropy sources. When applied on iris recognition scenarios, many subsets of an iris-code are used to lock the cryptographic key. The security of this construction relies on the entropy of subsets of iris codes. Simhadri et al. reported a security level of 32 bits on iris sources. In this paper, we propose two kinds of attacks to crack existing sample-then-lock schemes. Exploiting the low-entropy subsets, our attacks can break the locked key and the enrollment iris-code respectively in less than 220 brute force attempts. To protect from these proposed attacks, we design an improved sample-then-lock scheme. More precisely, our scheme employs stability and discriminability to select high-entropy subsets to lock the genuine secret, and conceals genuine locker by a large amount of chaff lockers. Our experiment verifies that existing schemes are vulnerable to the proposed attacks with a security level of less than 20 bits, while our scheme can resist these attacks with a security level of more than 100 bits when number of genuine subsets is 106.

With the proliferation of malware, the detection and classification of malware have been hot topics in the academic and industrial circles of cyber security, and the generation of malware signatures is one of the important research directions. In this paper, we propose NBP-MS, a method of signature generation that is based on network traffic generated by malware. Specifically, we utilize the network traffic generated by malware to perform fine-grained profiling of its network behaviors first, and then cluster all the profiles to generate network behavior signatures to classify malware, providing support for subsequent analysis and defense.

Trojan attacks threaten deep neural networks (DNNs) by poisoning them to behave normally on most samples, yet to produce manipulated results for inputs attached with a particular trigger. Several works attempt to detect whether a given DNN has been injected with a specific trigger during the training. In a parallel line of research, the lottery ticket hypothesis reveals the existence of sparse sub-networks which are capable of reaching competitive performance as the dense network after independent training. Connecting these two dots, we investigate the problem of Trojan DNN detection from the brand new lens of sparsity, even when no clean training data is available. Our crucial observation is that the Trojan features are significantly more stable to network pruning than benign features. Leveraging that, we propose a novel Trojan network detection regime: first locating a “winning Trojan lottery ticket” which preserves nearly full Trojan information yet only chance-level performance on clean inputs; then recovering the trigger embedded in this already isolated sub-network. Extensive experiments on various datasets, i.e., CIFAR-10, CIFAR-100, and ImageNet, with different network architectures, i.e., VGG-16, ResNet-18, ResNet-20s, and DenseNet-100 demonstrate the effectiveness of our proposal. Codes are available at https://github.com/VITA-Group/Backdoor-LTH.

Recently, with the increasing number of large-scale remote sensing images, the demand for large-scale remote sensing image object classification is growing and attracting the interest of many researchers. Hashing, because of its low memory requirements and high time efficiency, has widely solve the problem of large-scale remote sensing image. Supervised hashing methods mainly leverage the label information of remote sensing image to learn hash function, however, the similarity of the original feature space cannot be well preserved, which can not meet the accurate requirements for object classification of remote sensing image. To solve the mentioned problem, we propose a novel method named Optimized Projection Supervised Discrete Hashing(OPSDH), which jointly learns a discrete binary codes generation and optimized projection constraint model. It uses an effective optimized projection method to further constraint the supervised hash learning and generated hash codes preserve the similarity based on the data label while retaining the similarity of the original feature space. The experimental results show that OPSDH reaches improved performance compared with the existing hash learning methods and demonstrate that the proposed method is more efficient for operational applications.

The fusion of infrared and visible images using traditional multi-scale decomposition methods often leads to the loss of detailed information or the blurring of image edges, which is because the contour information and the detailed information within the contour cannot be retained simultaneously in the fusion process. To obtain high-quality fused images, a hybrid multi-scale decomposition fusion method using co-occurrence and Gaussian filters is proposed in this research. At first, by making full use of the smoothing effect of the Gaussian filter and edge protection characteristic of the co-occurrence filter, source images are decomposed into multiple hierarchical structures with different characteristics. Then, characteristics of sub-images at each level are analyzed, and the corresponding fusion rules are designed for images at different levels. At last, the final fused image obtained by combining fused sub-images of each level has rich scene information and clear infrared targets. Compared with several traditional multi-scale fusion algorithms, the proposed method has great advantages in some objective evaluation indexes.

The network security problem brought by the cloud computing has become an important issue to be dealt with in information construction. Since anomaly detection and attack detection in cloud environment need to find the vulnerability through the reverse analysis of data flow, it is of great significance to carry out the reverse analysis of unknown network protocol in the security application of cloud environment. To solve this problem, an improved mining method on bitstream protocol association rules with unknown type and format is proposed. The method combines the location information of the protocol framework to make the frequent extraction process more concise and accurate. In addition, for the frame separation problem of unknown protocol, we design a hierarchical clustering algorithm based on Jaccard distance and a frame field delimitation method based on the proximity of information entropy between bytes. The experimental results show that this technology can correctly resolve the protocol format and realize the purpose of anomaly detection in cloud computing, and ensure the security of cloud services.

This paper describes the architecture and the fundamental methodology of an anomaly detector, which by continuously monitoring Simple Network Management Protocol data and by processing it as complex-events, is able to timely recognize patterns of faults and relevant cyber-attacks. This solution has been applied in the context of smart grids, and in particular as part of a security and resilience component of the Information and Communication Technologies (ICT) Gateway, a middleware-based architecture that correlates and fuses measurement data from different sources (e.g., Inverters, Smart Meters) to provide control coordination and to enable grid observability applications. The detector has been evaluated through experiments, where we selected some representative anomalies that can occur on the ICT side of the energy distribution infrastructure: non-malicious faults (indicated by patterns in the system resources usage), as well as effects of typical cyber-attacks directed to the smart grid infrastructure. The results show that the detection is promisingly fast and efficient.

Clustering artworks is difficult for several reasons. On the one hand, recognizing meaningful patterns in accordance with domain knowledge and visual perception is extremely difficult. On the other hand, applying traditional clustering and feature reduction techniques to the highly dimensional pixel space can be ineffective. To address these issues, we propose to use a deep convolutional embedding model for digitized painting clustering, in which the task of mapping the raw input data to an abstract, latent space is jointly optimized with the task of finding a set of cluster centroids in this latent feature space. Quantitative and qualitative experimental results show the effectiveness of the proposed method. The model is also capable of outperforming other state-of-the-art deep clustering approaches to the same problem. The proposed method can be useful for several art-related tasks, in particular visual link retrieval and historical knowledge discovery in painting datasets.

With the rise of the so-called Adversarial Attacks, there is an increased concern on model security. In this paper we present two different contributions: novel measures of robustness (based on adversarial attacks) and a novel adversarial attack. The key idea behind these metrics is to obtain a measure that could compare different architectures, with independence of how the input is preprocessed (robustness against different input sizes and value ranges). To do so, a novel adversarial attack is presented, performing a delayed elastic-net adversarial attack (constraints are only used whenever a successful adversarial attack is obtained). Experimental results show that our approach obtains state-of-the-art adversarial samples, in terms of minimal perturbation distance. Finally, a benchmark of ImageNet pretrained models is used to conduct experiments aiming to shed some light about which model should be selected whenever security is a role factor.

Malware classification is a problem of great significance in the domain of information security. This is because the classification of malware into respective families helps in determining their intent, activity, and level of threat. In this paper, we propose a novel deep learning approach to malware classification. The proposed method converts malware executables into image-based representations. These images are then classified into different malware families using an autoencoder enhanced deep convolutional neural network (AE-DCNN). In particular, we propose a novel training mechanism wherein a DCNN classifier is trained with the help of an encoder. We conjecture that using an encoder in the proposed way provides the classifier with the extra information that is perhaps lost during the forward propagation, thereby leading to better results. The proposed approach eliminates the use of feature engineering, reverse engineering, disassembly, and other domain-specific techniques earlier used for malware classification. On the standard Malimg dataset, we achieve a 10-fold cross-validation accuracy of 99.38% and F1-score of 99.38%. Further, due to the texture-based analysis of malware files, the proposed technique is resilient to several obfuscation techniques.

Integrated Gradients as an attribution method for deep neural network models offers simple implementability. However, it suffers from noisiness of explanations which affects the ease of interpretability. The SmoothGrad technique is proposed to solve the noisiness issue and smoothen the attribution maps of any gradient-based attribution method. In this paper, we present SmoothTaylor as a novel theoretical concept bridging Integrated Gradients and SmoothGrad, from the Taylor's theorem perspective. We apply the methods to the image classification problem, using the ILSVRC2012 ImageNet object recognition dataset, and a couple of pretrained image models to generate attribution maps. These attribution maps are empirically evaluated using quantitative measures for sensitivity and noise level. We further propose adaptive noising to optimize for the noise scale hyperparameter value. From our experiments, we find that the SmoothTaylor approach together with adaptive noising is able to generate better quality saliency maps with lesser noise and higher sensitivity to the relevant points in the input space as compared to Integrated Gradients.

Multi-factor user authentication (MFUA) becomes increasingly popular due to its superior security comparing with single-factor user authentication. However, existing MFUAs require multiple interactions between users and different authentication components when sensing the multiple factors, leading to extra overhead and bad use experiences. In this paper, we propose a secure and user-friendly MFUA system, namely BioDraw, which utilizes four categories of biometrics (impedance, geometry, composition, and behavior) of human hand plus the pattern-based password to identify and authenticate users. A user only needs to draw a pattern on a RFID tag array, while four biometrics can be simultaneously collected. Particularly, we design a gradient-based pattern recognition algorithm for pattern recognition and then a CNN-LSTM-based classifier for user recognition. Furthermore, to guarantee the systemic security, we propose a novel anti-spoofing scheme, called Binary ALOHA, which utilizes the inhabit randomness of RFID systems. We perform extensive experiments over 21 volunteers. The experiment result demonstrates that BioDraw can achieve a high authentication accuracy (with a false reject rate less than 2%) and is effective in defending against various attacks.

This article describes the principles of construction, training and use of neural networks. The features of the neural network approach are indicated, as well as the range of tasks for which it is most preferable. Algorithms of functioning, software implementation and results of work of an artificial neural network are presented.

The rapid growth of artificial intelligence has contributed a lot to the technology world. As the traditional algorithms failed to meet the human needs in real time, Machine learning and deep learning algorithms have gained great success in different applications such as classification systems, recommendation systems, pattern recognition etc. Emotion plays a vital role in determining the thoughts, behaviour and feeling of a human. An emotion recognition system can be built by utilizing the benefits of deep learning and different applications such as feedback analysis, face unlocking etc. can be implemented with good accuracy. The main focus of this work is to create a Deep Convolutional Neural Network (DCNN) model that classifies 5 different human facial emotions. The model is trained, tested and validated using the manually collected image dataset.

A THz image edge detection approach based on wavelet and neural network is proposed in this paper. First, the source image is decomposed by wavelet, the edges in the low-frequency sub-image are detected using neural network method and the edges in the high-frequency sub-images are detected using wavelet transform method on the coarsest level of the wavelet decomposition, the two edge images are fused according to some fusion rules to obtain the edge image of this level, it then is projected to the next level. Afterwards the final edge image of L-1 level is got according to some fusion rule. This process is repeated until reaching the 0 level thus to get the final integrated and clear edge image. The experimental results show that our approach based on fusion technique is superior to Canny operator method and wavelet transform method alone.

As the assets of people are growing, security and surveillance have become a matter of great concern today. When a criminal activity takes place, the role of the witness plays a major role in nabbing the criminal. The witness usually states the gender of the criminal, the pattern of the criminal's dress, facial features of the criminal, etc. Based on the identification marks provided by the witness, the criminal is searched for in the surveillance cameras. Surveillance cameras are ubiquitous and finding criminals from a huge volume of surveillance video frames is a tedious process. In order to automate the search process, proposed a novel smart methodology using deep learning. This method takes gender, shirt pattern, and spectacle status as input to find out the object as person from the video log. The performance of this method achieves an accuracy of 87% in identifying the person in the video frame.

The last decade has witnessed an increase in online human interactions, covering all aspects of personal and professional activities. Identification of people based on their behavior rather than physical traits is a growing industry, spanning diverse spheres such as online education, e-commerce and cyber security. One prominent behavior is the expression of opinions, commonly as a reaction to images posted online. Visual aesthetic is a soft, behavioral biometric that refers to a person's sense of fondness to a certain image. Identifying individuals using their visual aesthetics as discriminatory features is an emerging domain of research. This paper introduces a new method for aesthetic feature dimensionality reduction using gene expression programming. The advantage of this method is that the resulting system is capable of using a tree-based genetic approach for feature recombination. Reducing feature dimensionality improves classifier accuracy, reduces computation runtime, and minimizes required storage. The results obtained on a dataset of 200 Flickr users evaluating 40000 images demonstrates a 94% accuracy of identity recognition based solely on users' aesthetic preferences. This outperforms the best-known method by 13.5%.

Deep neural networks (DNNs) provide good performance for image recognition, speech recognition, and pattern recognition. However, a poisoning attack is a serious threat to DNN's security. The poisoning attack is a method to reduce the accuracy of DNN by adding malicious training data during DNN training process. In some situations such as a military, it may be necessary to drop only a chosen class of accuracy in the model. For example, if an attacker does not allow only nuclear facilities to be selectively recognized, it may be necessary to intentionally prevent UAV from correctly recognizing nuclear-related facilities. In this paper, we propose a selective poisoning attack that reduces the accuracy of only chosen class in the model. The proposed method reduces the accuracy of a chosen class in the model by training malicious training data corresponding to a chosen class, while maintaining the accuracy of the remaining classes. For experiment, we used tensorflow as a machine learning library and MNIST and CIFAR10 as datasets. Experimental results show that the proposed method can reduce the accuracy of the chosen class to 43.2% and 55.3% in MNIST and CIFAR10, while maintaining the accuracy of the remaining classes.

Warnings happen a lot in real transmission networks. These warnings can affect people's lives. It is significant to analyze the alarm association rules in the network. Many algorithms can help solve this problem but not considering the actual physical significance. Therefore, in this study, we mine the association rules in warning weblogs based on a sequential mining algorithm (GSP) with topology structure. We define a topology constraint from network physical connection data. Under the topology constraint, network nodes have topology relation if they are directly connected or have a common adjacency node. In addition, due to the large amount of data, we implement the hash-tree search method to improve the mining efficiency. The theoretical solution is feasible and the simulation results verify our method. In simulation, the topology constraint improves the accuracy for 86%-96% and decreases the run time greatly at the same time. The hash-tree based mining results show that hash tree efficiency improvements are in 3-30% while the number of patterns remains unchanged. In conclusion, using our method can mine association rules efficiently and accurately in warning weblogs.

Nearest neighbor search (NNS) is one of the current popular research directions, which widely used in machine learning, pattern recognition, image detection and so on. In the low dimension data, based on tree search method can get good results. But when the data dimension goes up, that will produce a curse of dimensional. The proposed Locality-Sensitive Hashing algorithm (LSH) greatly improves the efficiency of nearest neighbor query for high dimensional data. But the algorithm relies on the building a large number of hash table, which makes the space complexity very high. C2LSH based on dynamic collision improves the disadvantage of LSH, but its disadvantage is that it needs to detect the collision times of a large number of data points which Increased query time. Therefore, Based on LSH algorithm, later researchers put forward many improved algorithms, but still not ideal.In this paper, we put forward Locality-Sensitive Hashing Scheme Based on Heap Sort of Hash Bucket (HSLSH) algorithm aiming at the shortcomings of LSH and C2LSH. Its main idea is to take advantage of the efficiency of heapsort in massive data sorting to improve the efficiency of nearest neighbor query. It only needs to rely on a small number of hash functions can not only overcome the shortcoming of LSH need to build a large number of hash table, and avoids defects of C2LSH. Experiments show that our algorithm is more than 20% better than C2LSH in query accuracy and 40% percent lower in query time.

Supervisory control and data acquisition (SCADA) systems are becoming increasingly susceptible to cyber-physical attacks on both physical and cyber layers of critical information infrastructure. Failure Mode and Effects Analysis (FMEA) have been widely used as a structured method to prioritize all possible vulnerable areas (failure modes) for design review of security of information systems. However, traditional RPN based FMEA has some inherent problems. Besides, there is a lacking of application of FMEA for security in SCADAs under vague and uncertain environment. Thus, the main purpose of this study was to propose a new evaluation model, which not only intends to recover above mentioned problems, but also intends to evaluate, prioritize and correct security risk of SCADA system's threat modes. A numerical case study was also conducted to demonstrate that the proposed new evaluation model is not only capable of addressing FMEA's inherent problems but also is best suited for a semi-quantitative high level analysis of a secure SCADA's failure modes in the early design phases.

SQL injection attacks are a kind of the greatest security risks on Web applications. Much research has been done to detect SQL injection attacks by rule matching and syntax tree. However, due to the complexity and variety of SQL injection vulnerabilities, these approaches fail to detect unknown and variable SQL injection attacks. In this paper, we propose a model, ATTAR, to detect SQL injection attacks using grammar pattern recognition and access behavior mining. The most important idea of our model is to extract and analyze features of SQL injection attacks in Web access logs. To achieve this goal, we first extract and customize Web access log fields from Web applications. Then we design a grammar pattern recognizer and an access behavior miner to obtain the grammatical and behavioral features of SQL injection attacks, respectively. Finally, based on two feature sets, machine learning algorithms, e.g., Naive Bayesian, SVM, ID3, Random Forest, and K-means, are used to train and detect our model. We evaluated our model on these two feature sets, and the results show that the proposed model can effectively detect SQL injection attacks with lower false negative rate and false positive rate. In addition, comparing the accuracy of our model based on different algorithms, ID3 and Random Forest have a better ability to detect various kinds of SQL injection attacks.

Access to information and data is becoming an essential part of nearly every aspect of modern business operation. Unfortunately, accessing information systems comes with increased chances of intrusion and unauthorized access. Acquiring and maintaining evidence from a computer or networks in the current high-tech world is essential in any comprehensive forensic investigation. Software and hardware tools are used to easily manage the evidence and view all relevant files. In an effort to enhance computer access security, keystroke authentication, is one of the biometric solutions that were proposed as a solution for enhancing users' identification. This research proposes using user's keystroke errors to determine guilt during forensics investigations, where it was found that individuals keystroke patters are repeatable and variant from those of others, and that keystroke patterns are impossible to steal or imitate. So, in this paper, we investigate the effectiveness of relying on ``user's mistakes'' as another behavioral biometric keystroke dynamic.