Graph neural networks (GNNs) are a powerful class of models that learn effective representations over dependencies between entities in a graph. Despite the achievement of GNNs in capturing both long-and short-distance semantics, challenges persist within text classification problems, such as multilingual sentiment analysis. Most existing GNNs methods are unable to capture word ordering and do not effectively support inductive learning with new data. In this work, to handle these challenges, we propose Ind-MSA, an Inductive Multilingual Sentiment Analysis approach. Our initial step involves constructing a single heterogeneous text graph, leveraging diverse information to effectively model the multilingual corpus. Subsequently, the learned word representations are employed to train a Bi-LSTM with attention mechanism, enhancing the proposed approach by incorporating sequential information into the analysis. The proposed approach can capture both short-and long-distance semantics while supporting the word ordering, which is extremely critical in sentiment analysis. Comprehensive experiments on various distinct datasets reveal that Ind-MSA leads to state-of-the-art results, significantly outperforming methods which focus on local consecutive word sequences and global word co-occurrence.

Mobile CrowdSensing (MCS) is a new and attractive paradigm that offers an alternative solution to data collection. This paradigm has recently been used to aggregate data for Intelligent Transportation Systems (ITS) towards a smart city, as an alternative to traditional data collection associated with expensive sensors. However, as many users can collaborate on the sensing task, data reliability remains a major challenge for ITS, especially for visual content such as images. As a result, image forgery detection becomes an essential task in MCS applications, intending to guarantee image reliability. Its goal is to identify manipulated content which requires robust techniques. Many researchers addressed this using deep learning models and handcrafted features. With attention mechanisms’ success in the last few years, more researchers have adopted these techniques for accurate image forgery detection, motivated by the slogan "attention is all you need" [1]. However, as the proposed techniques rely usually on pixel-level supervision, and sometimes with the intervention of handcrafted features, it is therefore unclear whether attention mechanisms are able to focus and localize forged regions with no further guidance. This paper addresses this problem by studying the effectiveness of the attention mechanism in localizing forged regions. To this end, we designed a customized Convolutional Neural Network architecture, incorporating the proposed Dual Focus Attention Module. Next, the Grad-CAM technique [2] was employed to compare the ground truths of the forged images with the heat maps generated by Grad-CAM, to provide insights into the addressed problem.

Mobile Crowdsensing (MCS) is an effective sensing paradigm that promotes data acquisition and sharing among mobile devices by engaging a crowd of participants. However, MCS systems face two major challenges, dealing with unreliable participants, and preventing privacy leakage. This paper presents BlockCrowd, a privacy-preserving trust and reputation management system for MCS based on blockchain to address these challenges. First, smart contracts are used to enhance user engagement by creating a transparent and trustworthy system while preserving privacy. Second, a reputation management scheme is designed based on the quality of users' contributions to resist malicious activities. Finally, the task publishing and worker selection processes are based on user reputation metrics. BlockCrowd is implemented on the Ethereum blockchain using the solidity programming language. Security and performance evaluations revealed that BlockCrowd can effectively identify malicious participants while ensuring security and privacy.

Accurate short-term traffic flow prediction is crucial for effective urban traffic management. However, selecting the most suitable prediction model and relevant features poses a significant challenge. Moreover, predicting traffic flow on a road using only historical data, adjacent roads, or all roads in the study area can compromise the model’s accuracy or execution time. This paper tackles this challenge by proposing an enhanced Vector Auto regression VAR-based prediction method. We suggest selecting relevant roads for the model using spatio-temporal correlation analysis. Subsequently, we conduct a comparative study between our methodology’s results and those obtained from other temporal and spatio-temporal traffic forecasting methods, including historical average, K-nearest neighbors (KNN), support vector machine for regression (SVR), and autoregressive model (AR). Model performance is evaluated by considering both the impact of normal and abnormal traffic conditions, as well as the selected training days: weekdays and weekends. The study utilizes a traffic dataset collected from an area of Xuancheng city in China. The proposed enhanced VAR outperforms the other methods for short-term forecasting horizons ($\approx$ from 5 to 25 minutes), under both normal and abnormal traffic conditions.

With the advancement of technology, the collection of data used in Intelligent Transportation Systems has become increasingly easy, notably with the emergence of the Mobile CrowdSensing paradigm. This paradigm could provide insights into traffic situation, road condition, pedestrians' behaviours, public transportation situation, and so on. Through the use of the powerful sensors in the mobile, different types of data can be generated, such as Gyroscope data, light sensors data, Magnetometer data, GPS data, Accelerometer data, videos, and images. However, the use of MCS raises issues of data reliability, as it involves the participation of several mobile owners with different 'levels' of trustworthiness. An important related issue is image forgery, i.e. the manipulation of images in order to deliberately provide misleading information, which constitutes a threat to the decision-makers of the MSC-based applications in Intelligent Transportation Systems. In this paper, an overview on image forgery is presented. We examine the workflows, approaches and techniques employed for image forgery detection and localization. Then, we provide a brief review of some image splicing localization techniques. Finally, we provide a comparative analysis of handcrafted feature extraction-based techniques and deep learning-based techniques. The aim of this paper is to draw attention to the image forgery problem that could threaten any Mobile CrowdSensing application dedicated to image collection, and to provide an overview of different existing techniques that could be used to overcome this problem.

With the growing number of vehicles and increasing traffic complexity, the development of Intelligent Transportation Systems (ITS) has become essential in managing traffic flow, reducing congestion, and improving road safety. As a result, numerous research initiatives and projects have been launched in recent years in the Middle East and North African (MENA) countries. This paper presents a comprehensive review of the ITS projects funded in this region. We then evaluate these projects based on three criteria: security and data privacy, data analysis, and sensing techniques. The main objective is to identify common threads among these projects and explore potential opportunities for developing an open-source platform that can benefit the research community in the MENA region with similar research objectives.

The growth of Mobile Crowd Sensing has had an important impact on various industries, particularly Intelligent Transportation Systems, where it has been an efficient method for gathering data. Despite its advantages, there are significant privacy issues associated with MCS-based applications, which can result in the unauthorized exposure of users’ identities and locations. For this purpose, we explore the integration of MCS in the context of ITS, and its privacy concerns. This paper outlines the structure of MCS systems and provides an overview of their use in the context of ITS. It then discusses the privacy issues and measures related to MCS and conducts a literature review of privacy protection mechanisms using proposed privacy measures as evaluation criteria. Finally, the paper identifies open research challenges and future directions for the integration of privacy-preserving approaches in MCS applications within the ITS context, offering a comprehensive understanding and highlighting areas for further research.

This paper presents the findings of a field study conducted in Rabat, the capital of Morocco, utilizing the Wardriving technique to assess Wi-Fi network security. The study encompasses approximately 10,000 Wi-Fi networks situated in residential and administrative neighborhoods in Rabat. Through our comprehensive analysis, we observed that a substantial 89.42% of the networks use WPA2, suggesting that Wi-Fi security in Morocco compares favorably to that of developed countries. We also found that most networks don’t use default configurations, this indicates a proactive approach by network administrators to implement robust security measures. Moreover, our investigation revealed a balanced distribution of channels 1, 6, and 11, illustrating that network operators are mindful of potential interferences, particularly on channel 6, and have taken measures to mitigate such interferences effectively. Based on our results, we draw a positive conclusion that the Wi-Fi situation in the examined neighborhoods of Rabat is highly encouraging. The study highlights the efforts made by network administrators to secure their Wi-Fi infrastructures and optimize network performance, contributing to a safer and more reliable wireless environment for users. This research serves as a valuable reference for understanding the state of Wi-Fi security in Rabat and provides insights into the overall Wi-Fi landscape in the city. The data-driven conclusions can aid policymakers, businesses, and individuals in further enhancing Wi-Fi security practices to ensure the continued growth and stability of wireless connectivity in the region.

Data breaches have begun to take on new dimensions and their prediction is becoming of great importance to organizations. Prior work has addressed this issue mainly from a technical perspective and neglected other interfering aspects such as the social media dimension. To fill this gap, we propose STRisk which is a predictive system where we expand the scope of the prediction task by bringing into play the social media dimension. We study over 3800 US organizations including both victim and non-victim organizations. For each organization, we design a profile composed of a variety of externally measured technical indicators and social factors. In addition, to account for unreported incidents, we consider the non-victim sample to be noisy and propose a noise correction approach to correct mislabeled organizations. We then build several machine learning models to predict whether an organization is exposed to experience a hacking breach. By exploiting both technical and social features, we achieve a Area Under Curve (AUC) score exceeding 98%, which is 12% higher than the AUC achieved using only technical features. Furthermore, our feature importance analysis reveals that open ports and expired certificates are the best technical predictors, while spreadability and agreeability are the best social predictors.

We propose Hercules, a parallel tree-based technique for exact similarity search on massive disk-based data series collections. We present novel index construction and query answering algorithms that leverage different summarization techniques, carefully schedule costly operations, optimize memory and disk accesses, and exploit the multi-threading and SIMD capabilities of modern hardware to perform CPU-intensive calculations. We demonstrate the superiority and robustness of Hercules with an extensive experimental evaluation against state-of-the-art techniques, using many synthetic and real datasets, and query workloads of varying difficulty. The results show that Hercules performs up to one order of magnitude faster than the best competitor (which is not always the same). Moreover, Hercules is the only index that outperforms the optimized scan on all scenarios, including the hard query workloads on disk-based datasets.

Understanding the properties exhibited by Spatial-temporal evolution of cyber attacks improve cyber threat intelligence. In addition, better understanding on threats patterns is a key feature for cyber threats prevention, detection, and management and for enhancing defenses. In this work, we study different aspects of emerging threats in the wild shared by 160,000 global participants form all industries. First, we perform an exploratory data analysis of the collected cyber threats. We investigate the most targeted countries, most common malwares and the distribution of attacks frequency by localisation. Second, we extract attacks’ spreading patterns at country level. We model these behaviors using transition graphs decorated with probabilities of switching from a country to another. Finally, we analyse the extent to which cyber threats have been affected by the COVID-19 outbreak and sanitary measures imposed by governments to prevent the virus from spreading.

Data series are a special type of multidimensional data present in numerous domains, where similarity search is a key operation that has been extensively studied in the data series literature. In parallel, the multidimensional community has studied approximate similarity search techniques. We propose a taxonomy of similarity search techniques that reconciles the terminology used in these two domains, we describe modifications to data series indexing techniques enabling them to answer approximate similarity queries with quality guarantees, and we conduct a thorough experimental evaluation to compare approximate similarity search techniques under a unified framework, on synthetic and real datasets in memory and on disk. Although data series differ from generic multidimensional vectors (series usually exhibit correlation between neighboring values), our results show that data series techniques answer approximate queries with strong guarantees and an excellent empirical performance, on data series and vectors alike. These techniques outperform the state-of-the-art approximate techniques for vectors when operating on disk, and remain competitive in memory.

Increasingly large data series collections are becoming commonplace across many different domains and applications. A key operation in the analysis of data series collections is similarity search, which has attracted lots of attention and effort over the past two decades. Even though several relevant approaches have been proposed in the literature, none of the existing studies provides a detailed evaluation against the available alternatives. The lack of comparative results is further exacerbated by the non-standard use of terminology, which has led to confusion and misconceptions. In this paper, we provide definitions for the different flavors of similarity search that have been studied in the past, and present the first systematic experimental evaluation of the efficiency of data series similarity search techniques. Based on the experimental results, we describe the strengths and weaknesses of each approach and give recommendations for the best approach to use under typical use cases. Finally, by identifying the shortcomings of each method, our findings lay the ground for solid further developments in the field.

In this paper, we investigate the use of Text Classification techniques to extract contextual information from user reviews for Context Aware Recommendation. We conduct several experiments to identify the best Text Representation settings and the best classification algorithm for our dataset. We carry out our experiments on hotel reviews. We focus on extracting the trip type, as contextual information, from these reviews. Results show that the Naïve Bayes classifier yields the best results with up to 72.2% in terms of F1-measure. To extract context from user reviews with text classification techniques, we recommend to use raw text rather than employing stemming, to use the normalized frequency based weighting rather than the presence based one, to remove terms that occur once in the data set, and to combine unigrams, bigrams and trigrams.

With the emergence of Web 2.0, Sentiment Analysis is receiving more and more attention. Several interesting works were performed to address different issues in Sentiment Analysis. Nevertheless, the problem of Unbalanced Data Sets was not enough tackled within this research area. This paper presents the study we have carried out to address the problem of unbalanced data sets in supervised sentiment classification in a multi-lingual context. We propose three different methods to under-sample the majority class documents. These methods are Remove Similar, Remove Farthest and Remove by Clustering. Our goal is to compare the effectiveness of the proposed methods with the common random under-sampling. We also aim to evaluate the behavior of the classifiers toward different under-sampling rates. We use three different common classifiers, namely Naïve Bayes, Support Vector Machines and k-Nearest Neighbors. The experiments are carried out on two Arabic data sets and an English data set. We show that the four under-sampling methods are typically competitive. Naïve Bayes is shown as insensitive to unbalanced data sets. But Support Vector Machines seems to be highly sensitive to unbalanced data sets; k-Nearest Neighbors shows a slight sensitivity to imbalance in comparison with Support Vector Machines.