Biblio

There have been a growing number of interests in using the convolutional neural network(CNN) in image forensics, where some excellent methods have been proposed. Training the randomly initialized model from scratch needs a big amount of training data and computational time. To solve this issue, we present a new method of training an image forensic model using prior knowledge transferred from the existing steganalysis model. We also find out that CNN models tend to show poor performance when tested on a different database. With knowledge transfer, we are able to easily train an excellent model for a new database with a small amount of training data from the new database. Performance of our models are evaluated on Bossbase and BOW by detecting five forensic types, including median filtering, resampling, JPEG compression, contrast enhancement and additive Gaussian noise. Through a series of experiments, we demonstrate that our proposed method is very effective in two scenario mentioned above, and our method based on transfer learning can greatly accelerate the convergence of CNN model. The results of these experiments show that our proposed method can detect five different manipulations with an average accuracy of 97.36%.

Caching query results is an efficient technique for Web search engines. A state-of-the-art approach named Static-Dynamic Cache (SDC) is widely used in practice. Replacement policy is the key factor on the performance of cache system, and has been widely studied such as LIRS, ARC, CLOCK, SKLRU and RANDOM in different research areas. In this paper, we discussed replacement policies for static-dynamic cache and conducted the experiments on real large scale query logs from two famous commercial Web search engine companies. The experimental results show that ARC replacement policy could work well with static-dynamic cache, especially for large scale query results cache.

MANETs have been focusing the interest of researchers for several years. The new scenarios where MANETs are being deployed make that several challenging issues remain open: node scalability, energy efficiency, network lifetime, Quality of Service (QoS), network overhead, data privacy and security, and effective routing. This latter is often seen as key since it frequently constrains the performance of the overall network. Location-based routing protocols provide a good solution for scalable MANETs. Although several location-based routing protocols have been proposed, most of them rely on error-free positions. Only few studies have focused so far on how positioning error affects the routing performance; also, most of them consider outdated solutions. This paper is aimed at filling this gap, by studying the impact of the error in the position of the nodes of two location-based routing protocols: DYMOselfwd and AODV-Line. These protocols were selected as they both aim at reducing the routing overhead. Simulations considering different mobility patterns in a dense network were conducted, so that the performance of these protocols can be assessed under ideal (i.e. error-less) and realistic (i.e. with error) conditions. The results show that AODV-Line builds less reliable routes than DYMOselfwd in case of error in the position information, thus increasing the routing overhead.

Recent studies have shown that adding explicit social trust information to social recommendation significantly improves the prediction accuracy of ratings, but it is difficult to obtain a clear trust data among users in real life. Scholars have studied and proposed some trust measure methods to calculate and predict the interaction and trust between users. In this article, a method of social trust relationship extraction based on hellinger distance is proposed, and user similarity is calculated by describing the f-divergence of one side node in user-item bipartite networks. Then, a new matrix factorization model based on implicit social relationship is proposed by adding the extracted implicit social relations into the improved matrix factorization. The experimental results support that the effect of using implicit social trust to recommend is almost the same as that of using actual explicit user trust ratings, and when the explicit trust data cannot be extracted, our method has a better effect than the other traditional algorithms.

In the present time, there has been a huge increase in large data repositories by corporations, governments, and healthcare organizations. These repositories provide opportunities to design/improve decision-making systems by mining trends and patterns from the data set (that can provide credible information) to improve customer service (e.g., in healthcare). As a result, while data sharing is essential, it is an obligation to maintaining the privacy of the data donors as data custodians have legal and ethical responsibilities to secure confidentiality. This research proposes a 2-layer privacy preserving (2-LPP) data sanitization algorithm that satisfies ε-differential privacy for publishing sanitized data. The proposed algorithm also reduces the re-identification risk of the sanitized data. The proposed algorithm has been implemented, and tested with two different data sets. Compared to other existing works, the results obtained from the proposed algorithm show promising performance.

This paper studies physical layer security in a wireless ad hoc network with numerous legitimate transmitter-receiver pairs and passive eavesdroppers. A hybrid full-/half-duplex receiver deployment strategy is proposed to secure legitimate transmissions, by letting a fraction of legitimate receivers work in the full-duplex (FD) mode sending jamming signals to confuse eavesdroppers upon their own information receptions, and other receivers work in the half-duplex mode just receiving desired signals. This paper aims to properly choose the fraction of the FD receivers to enhance network security. Tractable expressions for the connection outage probability and the secrecy outage probability of a typical legitimate link are first derived, based on which the network-wide secrecy throughput is maximized. Some insights into the optimal fraction are further developed. It is concluded that the fraction of the FD receivers triggers a non-trivial trade-off between reliability and secrecy, and the optimal fraction significantly improves the network security performance.

Third-party IME (Input Method Editor) apps are often the preference means of interaction for Android users' input. In this paper, we first discuss the insecurity of IME apps, including the Potentially Harmful Apps (PHA) and malicious IME apps, which may leak users' sensitive keystrokes. The current defense system, such as I-BOX, is vulnerable to the prefix-substitution attack and the colluding attack due to the post-IME nature. We provide a deeper understanding that all the designs with the post-IME nature are subject to the prefix-substitution and colluding attacks. To remedy the above post-IME system's flaws, we propose a new idea, pre-IME, which guarantees that "Is this touch event a sensitive keystroke?" analysis will always access user touch events prior to the execution of any IME app code. We designed an innovative TrustZone-based framework named IM-Visor which has the pre-IME nature. Specifically, IM-Visor creates the isolation environment named STIE as soon as a user intends to type on a soft keyboard, then the STIE intercepts, translates and analyzes the user's touch input. If the input is sensitive, the translation of keystrokes will be delivered to user apps through a trusted path. Otherwise, IM-Visor replays non-sensitive keystroke touch events for IME apps or replays non-keystroke touch events for other apps. A prototype of IM-Visor has been implemented and tested with several most popular IMEs. The experimental results show that IM-Visor has small runtime overheads.

Context suggestion refers to the task of recommending appropriate contexts to the users to improve the user experience. The suggested contexts could be time, location, companion, category, and so forth. In this paper, we particularly focus on the task of suggesting appropriate contexts to a user on a specific item. We evaluate the indirect context suggestion approaches over a movie data collected from user surveys, in comparison with direct context prediction approaches. Our experimental results reveal that indirect context suggestion is better and tensor factorization is generally the best way to suggest contexts to a user when given an item.

Insider threat is a significant security risk for information system, and detection of insider threat is a major concern for information system organizers. Recently existing work mainly focused on the single pattern analysis of user single-domain behavior, which were not suitable for user behavior pattern analysis in multi-domain scenarios. However, the fusion of multi-domain irrelevant features may hide the existence of anomalies. Previous feature learning methods have relatively a large proportion of information loss in feature extraction. Therefore, this paper proposes a hybrid model based on the deep belief network (DBN) to detect insider threat. First, an unsupervised DBN is used to extract hidden features from the multi-domain feature extracted by the audit logs. Secondly, a One-Class SVM (OCSVM) is trained from the features learned by the DBN. The experimental results on the CERT dataset demonstrate that the DBN can be used to identify the insider threat events and it provides a new idea to feature processing for the insider threat detection.

Remote Application Programming Interfaces (APIs) are technology enablers for major distributed system trends such as mobile and cloud computing and the Internet of Things. In such settings, message-based APIs dominate over procedural and object-oriented ones. It is hard to design such APIs so that they are easy and efficient to use for client developers. Maintaining their runtime qualities while preserving backward compatibility is equally challenging for API providers. For instance, finding a well suited granularity for services and their operations is a particularly important design concern in APIs that realize service-oriented software architectures. Due to the fallacies of distributed computing, the forces for message-based APIs and service interfaces differ from those for local APIs – for instance, network latency and security concerns deserve special attention. Existing pattern languages have dealt with local APIs in object-oriented programming, with remote objects, with queue-based messaging and with service-oriented computing platforms. However, patterns or equivalent guidance for the structural design of request and response messages in message-based remote APIs is still missing. In this paper, we outline such a pattern language and introduce five basic interface representation patterns to promote platform-independent design advice for common remote API technologies such as RESTful HTTP and Web services (WSDL/SOAP). Known uses and examples of the patterns are drawn from public Web APIs, as well as application development and software integration projects the authors have been involved in.

Modbus over TCP/IP is one of the most popular industrial network protocol that are widely used in critical infrastructures. However, vulnerability of Modbus TCP protocol has attracted widely concern in the public. The traditional intrusion detection methods can identify some intrusion behaviors, but there are still some problems. In this paper, we present an innovative approach, SD-IDS (Stereo Depth IDS), which is designed for perform real-time deep inspection for Modbus TCP traffic. SD-IDS algorithm is composed of two parts: rule extraction and deep inspection. The rule extraction module not only analyzes the characteristics of industrial traffic, but also explores the semantic relationship among the key field in the Modbus TCP protocol. The deep inspection module is based on rule-based anomaly intrusion detection. Furthermore, we use the online test to evaluate the performance of our SD-IDS system. Our approach get a low rate of false positive and false negative.

Internet of things (IoT) is internetworking of various physical devices to provide a range of services and applications. IoT is a rapidly growing field, on an account of this; the security measurements for IoT should be at first concern. In the modern day world, the most emerging cyber-attack threat for IoT is ransomware attack. Ransomware is a kind of malware with the aim of rendering a victim's computer unusable or inaccessible, and then asking the user to pay a ransom to revert the destruction. In this paper we are evaluating ransomware attacks statistics for the past 2 years and the present year to estimate growth rate of the most emerging ransomware families from the last 3 years to evaluate most threatening ransomware attacks for IoT. Growth rate results shows that the number of attacks for Cryptowall and locky ransomware are notably increasing therefore, these ransomware families are potential threat to IoT. Moreover, we present a Cryptowall ransomware attack detection model based on the communication and behavioral study of Cryptowall for IoT environment. The proposed model observes incoming TCP/IP traffic through web proxy server then extracts TCP/IP header and uses command and control (C&C) server black listing to detect ransomware attacks.

Most of Wireless Sensor Networks (WSNs) are usually deployed in hostile environments where the communications conditions are not stable and not reliable. Hence, there is a need to design an effective distributed schemes to enable the sensors cooperating in order to recover the sensed data. In this paper, we establish a novel cooperative data exchange (CDE) scheme using instantly decodable network coding (IDNC) across the sensor nodes. We model the problem using the cooperative game theory in partition form. We develop also a distributed merge-and-split algorithm in order to form dynamically coalitions that maximize their utilities in terms of both energy consumption and IDNC delay experienced by all sensors. Indeed, the proposed algorithm enables these sensors to self-organize into stable clustered network structure where all sensors do not have incentives to change the cluster he is part of. Simulation results show that our cooperative scheme allows nodes not only to reduce the energy consumption, but also the IDNC completion time.

With the advance of fifth generation (5G) networks, network density needs to grow significantly in order to meet the required capacity demands. A massive deployment of small cells may lead to a high cost for providing fiber connectivity to each node. Consequently, many small cells are expected to be connected through wireless links to the umbrella eNodeB, leading to a mesh backhaul topology. This backhaul solution will most probably be composed of high capacity point-to-point links, typically operating in the millimeter wave (mmWave) frequency band due to its massive bandwidth availability. In this paper, we propose a mathematical model that jointly solves the user association and backhaul routing problem in the aforementioned context, aiming at the energy efficiency maximization of the network. Our study considers the energy consumption of both the access and backhaul links, while taking into account the capacity constraints of all the nodes as well as the fulfillment of the service-level agreements (SLAs). Due to the high complexity of the optimal solution, we also propose an energy efficient heuristic algorithm (Joint), which solves the discussed joint problem, while inducing low complexity in the system. We numerically evaluate the algorithm performance by comparing it not only with the optimal solution but also with reference approaches under different traffic load scenarios and backhaul parameters. Our results demonstrate that Joint outperforms the state-of-the-art, while being able to find good solutions, close to optimal, in short time.

Autoencoders have been successful in learning meaningful representations from image datasets. However, their performance on text datasets has not been widely studied. Traditional autoencoders tend to learn possibly trivial representations of text documents due to their confoundin properties such as high-dimensionality, sparsity and power-law word distributions. In this paper, we propose a novel k-competitive autoencoder, called KATE, for text documents. Due to the competition between the neurons in the hidden layer, each neuron becomes specialized in recognizing specific data patterns, and overall the model can learn meaningful representations of textual data. A comprehensive set of experiments show that KATE can learn better representations than traditional autoencoders including denoising, contractive, variational, and k-sparse autoencoders. Our model also outperforms deep generative models, probabilistic topic models, and even word representation models (e.g., Word2Vec) in terms of several downstream tasks such as document classification, regression, and retrieval.

This paper is the first work to perform spatio-temporal mapping of human activity using the visual content of geo-tagged videos. We utilize a recent deep-learning based video analysis framework, termed hidden two-stream networks, to recognize a range of activities in YouTube videos. This framework is efficient and can run in real time or faster which is important for recognizing events as they occur in streaming video or for reducing latency in analyzing already captured video. This is, in turn, important for using video in smart-city applications. We perform a series of experiments to show our approach is able to map activities both spatially and temporally.

Side-channel risks of Intel SGX have recently attracted great attention. Under the spotlight is the newly discovered page-fault attack, in which an OS-level adversary induces page faults to observe the page-level access patterns of a protected process running in an SGX enclave. With almost all proposed defense focusing on this attack, little is known about whether such efforts indeed raise the bar for the adversary, whether a simple variation of the attack renders all protection ineffective, not to mention an in-depth understanding of other attack surfaces in the SGX system. In the paper, we report the first step toward systematic analyses of side-channel threats that SGX faces, focusing on the risks associated with its memory management. Our research identifies 8 potential attack vectors, ranging from TLB to DRAM modules. More importantly, we highlight the common misunderstandings about SGX memory side channels, demonstrating that high frequent AEXs can be avoided when recovering EdDSA secret key through a new page channel and fine-grained monitoring of enclave programs (at the level of 64B) can be done through combining both cache and cross-enclave DRAM channels. Our findings reveal the gap between the ongoing security research on SGX and its side-channel weaknesses, redefine the side-channel threat model for secure enclaves, and can provoke a discussion on when to use such a system and how to use it securely.

Security threats such as jamming and route manipulation can have significant consequences on the performance of modern wireless networks. To increase the efficacy and stealthiness of such threats, a number of extremely challenging, next-generation cross-layer attacks have been recently unveiled. Although existing research has thoroughly addressed many single-layer attacks, the problem of detecting and mitigating cross-layer attacks still remains unsolved. For this reason, in this paper we propose a novel framework to analyze and address cross-layer attacks in wireless networks. Specifically, our framework consists of a detection and a mitigation component. The attack detection component is based on a Bayesian learning detection scheme that constructs a model of observed evidence to identify stealthy attack activities. The mitigation component comprises a scheme that achieves the desired trade-off between security and performance. We specialize and evaluate the proposed framework by considering a specific cross-layer attack that uses jamming as an auxiliary tool to achieve route manipulation. Simulations and experimental results obtained with a testbed made up by USRP software-defined radios demonstrate the effectiveness of the proposed methodology.

Using mobile sinks to collect sensed data in WSNs (Wireless Sensor Network) is an effective technique for significantly improving the network lifetime. We investigate the problem of collecting sensed data using a mobile sink in a WSN with unreachable regions such that the network lifetime is maximized and the total tour length is minimized, and propose a polynomial-time heuristic, an ILP-based (Integer Linear Programming) heuristic and an MINLP-based (Mixed-Integer Non-Linear Programming) algorithm for constructing a shortest path routing forest for the sensor nodes in unreachable regions, two energy-efficient heuristics for partitioning the sensor nodes in reachable regions into disjoint clusters, and an efficient approach to convert the tour construction problem into a TSP (Travelling Salesman Problem). We have performed extensive simulations on 100 instances with 100, 150, 200, 250 and 300 sensor nodes in an urban area and a forest area. The simulation results show that the average lifetime of all the network instances achieved by the polynomial-time heuristic is 74% of that achieved by the ILP-based heuristic and 65% of that obtained by the MINLP-based algorithm, and our tour construction heuristic significantly outperforms the state-of-the-art tour construction heuristic EMPS.

To detect data races that harm production systems, program analysis must target production runs. However, sound and precise data race detection adds too much run-time overhead for use in production systems. Even existing approaches that provide soundness or precision incur significant limitations. This work addresses the need for soundness (no missed races) and precision (no false races) by introducing novel, efficient production-time analyses that address each need separately. (1) Precise data race detection is useful for developers, who want to fix bugs but loathe false positives. We introduce a precise analysis called RaceChaser that provides low, bounded run-time overhead. (2) Sound race detection benefits analyses and tools whose correctness relies on knowledge of all potential data races. We present a sound, efficient approach called Caper that combines static and dynamic analysis to catch all data races in observed runs. RaceChaser and Caper are useful not only on their own; we introduce a framework that combines these analyses, using Caper as a sound filter for precise data race detection by RaceChaser. Our evaluation shows that RaceChaser and Caper are efficient and effective, and compare favorably with existing state-of-the-art approaches. These results suggest that RaceChaser and Caper enable practical data race detection that is precise and sound, respectively, ultimately leading to more reliable software systems.

The number of resource-limited wireless devices utilized in many areas of Internet of Things is growing rapidly; there is a concern about privacy and security. Various lightweight block ciphers are proposed; this work presents a modified lightweight block cipher algorithm. A Linear Feedback Shift Register is used to replace the key generation function in the XTEA1 Algorithm. Using the same evaluation conditions, we analyzed the software implementation of the modified XTEA using FELICS (Fair Evaluation of Lightweight Cryptographic Systems) a benchmarking framework which calculates RAM footprint, ROM occupation and execution time on three largely used embedded devices: 8-bit AVR microcontroller, 16-bit MSP microcontroller and 32-bit ARM microcontroller. Implementation results show that it provides less software requirements compared to original XTEA. We enhanced the security level and the software performance.

Denial of service (DOS) attacks are a serious threat to network security. These attacks are often sourced from virtual machines in the cloud, rather than from the attacker's own machine, to achieve anonymity and higher network bandwidth. Past research focused on analyzing traffic on the destination (victim's) side with predefined thresholds. These approaches have significant disadvantages. They are only passive defenses after the attack, they cannot use the outbound statistical features of attacks, and it is hard to trace back to the attacker with these approaches. In this paper, we propose a DOS attack detection system on the source side in the cloud, based on machine learning techniques. This system leverages statistical information from both the cloud server's hypervisor and the virtual machines, to prevent network packages from being sent out to the outside network. We evaluate nine machine learning algorithms and carefully compare their performance. Our experimental results show that more than 99.7% of four kinds of DOS attacks are successfully detected. Our approach does not degrade performance and can be easily extended to broader DOS attacks.