Search Results (1,997)

Network traffic classification (NTC) has attracted great attention in many applications such as secure communications, intrusion detection systems. The existing NTC methods based on supervised learning rely on sufficient labeled datasets in the training phase, but for most traffic datasets, it is difficult to obtain label information in practical applications. Although unsupervised learning does not rely on labels, its classification accuracy is not high, and the number of data classes is difficult to determine. This paper proposes an unsupervised NTC method based on adversarial training and deep clustering with improved network traffic classification (NTC) and lower computational complexity in comparison with the traditional clustering algorithms. Here, the training process does not require data labels, which greatly reduce the computational complexity of the network traffic classification through pretraining. In the pretraining stage, an autoencoder (AE) is used to reduce the dimension of features and reduce the complexity of the initial high-dimensional network traffic data features. Moreover, we employ the adversarial training model and a deep clustering structure to further optimize the extracted features. The experimental results show that our proposed method has robust performance, with a multiclassification accuracy of 92.2%, which is suitable for classification with a large number of unlabeled data in actual application scenarios. This paper only focuses on breakthroughs in the algorithm stage, and future work can be focused on the deployment and adaptation in practical environments. Full article

In the rapidly evolving landscape of internet usage, ensuring robust cybersecurity measures has become a paramount concern across diverse fields. Among the numerous cyber threats, denial of service (DoS) and distributed denial of service (DDoS) attacks pose significant risks, as they can render websites and servers inaccessible to their intended users. Conventional intrusion detection methods encounter substantial challenges in effectively identifying and mitigating these attacks due to their widespread nature, intricate patterns, and computational complexities. However, by harnessing the power of deep learning-based techniques, our proposed dense channel-spatial attention model exhibits exceptional accuracy in detecting and classifying DoS and DDoS attacks. The successful implementation of our proposed framework addresses the challenges posed by imbalanced data and exhibits its potential for real-world applications. By leveraging the dense channel-spatial attention mechanism, our model can precisely identify and classify DoS and DDoS attacks, bolstering the cybersecurity defenses of websites and servers. The high accuracy rates achieved across different datasets reinforce the robustness of our approach, underscoring its efficacy in enhancing intrusion detection capabilities. As a result, our framework holds promise in bolstering cybersecurity measures in real-world scenarios, contributing to the ongoing efforts to safeguard against cyber threats in an increasingly interconnected digital landscape. Comparative analysis with current intrusion detection methods reveals the superior performance of our model. We achieved accuracy rates of 99.38%, 99.26%, and 99.43% for Bot-IoT, CICIDS2017, and UNSW_NB15 datasets, respectively. These remarkable results demonstrate the capability of our approach to accurately detect and classify various types of DoS and DDoS assaults. By leveraging the inherent strengths of deep learning, such as pattern recognition and feature extraction, our model effectively overcomes the limitations of traditional methods, enhancing the accuracy and efficiency of intrusion detection systems. Full article

Federated Learning is identified as a reliable technique for distributed training of ML models. Specifically, a set of dispersed nodes may collaborate through a federation in producing a jointly trained ML model without disclosing their data to each other. Each node performs local model training and then shares its trained model weights with a server node, usually called Aggregator in federated learning, as it aggregates the trained weights and then sends them back to its clients for another round of local training. Despite the data protection and security that FL provides to each client, there are still well-studied attacks such as membership inference attacks that can detect potential vulnerabilities of the FL system and thus expose sensitive data. In this paper, in order to prevent this kind of attack and address private data leakage, we introduce FREDY, a differential private federated learning framework that enables knowledge transfer from private data. Particularly, our approach has a teachers–student scheme. Each teacher model is trained on sensitive, disjoint data in a federated manner, and the student model is trained on the most voted predictions of the teachers on public unlabeled data which are noisy aggregated in order to guarantee the privacy of each teacher’s sensitive data. Only the student model is publicly accessible as the teacher models contain sensitive information. We show that our proposed approach guarantees the privacy of sensitive data against model inference attacks while it combines the federated learning settings for the model training procedures. Full article

In the quest to optimize user experience, network, and service, providers continually seek to deliver high-quality content tailored to individual preferences. However, predicting user perception of quality remains a challenging task, given the subjective nature of human perception and the plethora of technical attributes that contribute to the overall viewing experience. Thus, we introduce a Fuzzy Logic-bAsed ModEl for Video Quality Assessment (FLAME-VQA), leveraging the LIVE-YT-HFR database containing 480 video sequences and subjective ratings of their quality from 85 test subjects. The proposed model addresses the challenges of assessing user perception by capturing the intricacies of individual preferences and video attributes using fuzzy logic. It operates with four input parameters: video frame rate, compression rate, and spatio-temporal information. The Spearman Rank–Order Correlation Coefficient (SROCC) and Pearson Correlation Coefficient (PCC) show a high correlation between the output and the ground truth. For the training, test, and complete dataset, SROCC equals 0.8977, 0.8455, and 0.8961, respectively, while PCC equals 0.9096, 0.8632, and 0.9086, respectively. The model outperforms comparative models tested on the same dataset. Full article

Blockchain represents a decentralized and distributed ledger technology, ensuring transparent and secure transaction recording across networks. This innovative technology offers several benefits, including increased security, trust, and transparency, making it suitable for a wide range of applications. In the last few years, there has been a growing interest in investigating the potential of Blockchain technology to enhance diverse fields, such as e-learning. In this research, we undertook a systematic literature review to explore the potential of Blockchain technology in enhancing the e-learning domain. Our research focused on four main questions: (1) What potential characteristics of Blockchain can contribute to enhancing e-learning? (2) What are the existing Blockchain projects dedicated to e-learning? (3) What are the limitations of existing projects? (4) What are the future trends in Blockchain-related research that will impact e-learning? The results showed that Blockchain technology has several characteristics that could benefit e-learning. We also discussed immutability, transparency, decentralization, security, and traceability. We also identified several existing Blockchain projects dedicated to e-learning and discussed their potential to revolutionize learning by providing more transparency, security, and effectiveness. However, our research also revealed many limitations and challenges that could be addressed to achieve Blockchain technology’s potential in e-learning. Full article

Wireless sensor network (WSN) environment monitoring and smart city applications present challenges for maintaining network connectivity when, for example, dynamic events occur. Such applications can benefit from recent technologies such as software-defined networks (SDNs) and network virtualization to support network flexibility and offer validation for a physical network. This paper aims to present a testbed-based, software-defined wireless sensor network (SDWSN) for IoT applications with a focus on promoting the approach of virtual network testing and analysis prior to physical network implementation to monitor and repair any network failures. Herein, physical network implementation employing hardware boards such as Texas Instruments CC2538 (TI CC2538) and TI CC1352R sensor nodes is presented and designed based on virtual WSN- based clustering for stationary and dynamic networks use cases. The key performance indicators such as evaluating node (such as a gateway node to the Internet) connection capability based on packet drop and energy consumption virtually and physically are discussed. According to the test findings, the proposed software-defined physical network benefited from “prior-to-implementation” analysis via virtualization, as the performance of both virtual and physical networks is comparable. Full article

Autism spectrum disorder (ASD) has been associated with conditions like depression, anxiety, epilepsy, etc., due to its impact on an individual’s educational, social, and employment. Since diagnosis is challenging and there is no cure, the goal is to maximize an individual’s ability by reducing the symptoms, and early diagnosis plays a role in improving behavior and language development. In this paper, an autism screening analysis for toddlers and adults has been performed using fair AI (feature engineering, SMOTE, optimizations, etc.) and deep learning methods. The analysis considers traditional deep learning methods like Multilayer Perceptron (MLP), Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM), and also proposes two hybrid deep learning models, i.e., CNN–LSTM with Particle Swarm Optimization (PSO), and a CNN model combined with Gated Recurrent Units (GRU–CNN). The models have been validated using multiple performance metrics, and the analysis confirms that the proposed models perform better than the traditional models. Full article

Fifth-generation (5G) networks require efficient radio resource management (RRM) which should dynamically adapt to the current network load and user needs. Monitoring and forecasting network performance requirements and metrics helps with this task. One of the parameters that highly influences radio resource management is the profile of user traffic generated by various 5G applications. Forecasting such mobile network profiles helps with numerous RRM tasks such as network slicing and load balancing. In this paper, we analyze a dataset from a mobile network operator in Portugal that contains information about volumes of traffic in download and upload directions in one-hour time slots. We apply two statistical models for forecasting download and upload traffic profiles, namely, seasonal autoregressive integrated moving average (SARIMA) and Holt-Winters models. We demonstrate that both models are suitable for forecasting mobile network traffic. Nevertheless, the SARIMA model is more appropriate for download traffic (e.g., MAPE [mean absolute percentage error] of 11.2% vs. 15% for Holt-Winters), while the Holt-Winters model is better suited for upload traffic (e.g., MAPE of 4.17% vs. 9.9% for SARIMA and Holt-Winters, respectively). Full article

Unmanned Aerial Vehicles (UAVs) equipped with communication technologies have gained significant attention as a promising solution for providing wireless connectivity in remote, disaster-stricken areas lacking communication infrastructure. However, enabling UAVs to provide communications (e.g., UAVs acting as flying base stations) in real scenarios requires the integration of various technologies and algorithms. In particular 3D path planning algorithms are crucial in determining the optimal path free of obstacles so that UAVs in isolation or forming networks can provide wireless coverage in a specific region. Considering that most of the existing proposals in the literature only address path planning in a 2D environment, this paper systematically studies existing path-planning solutions in UAVs in a 3D environment in which optimization models (optimal and heuristics) have been applied. This paper analyzes 37 articles selected from 631 documents from a search in the Scopus database. This paper also presents an overview of UAV-enabled communications systems, the research questions, and the methodology for the systematic mapping study. In the end, this paper provides information about the objectives to be minimized or maximized, the optimization variables used, and the algorithmic strategies employed to solve the 3D path planning problem. Full article

Cloud computing has become ubiquitous in the enterprise environment as its on-demand model realizes technical and economic benefits for users. Cloud users demand a level of reliability, availability, and quality of service. Improvements to reliability generally come at the cost of additional replication. Existing approaches have focused on the replication of virtual environments as a method of improving the reliability of cloud services. As cloud systems move towards microservices-based architectures, a more granular approach to replication is now possible. In this paper, we propose a cloud orchestration approach that balances the potential cost of failure with the spot market running cost, optimizing the resource usage of the cloud system. We present the results of empirical testing we carried out using a simulator to compare the outcome of our proposed approach to a control algorithm based on a static reliability requirement. Our empirical testing showed an improvement of between 37% and 72% in total cost over the control, depending on the specific characteristics of the cloud models tested. We thus propose that in clouds where the cost of failure can be reasonably approximated, our approach may be used to optimize the cloud redundancy configuration to achieve a lower total cost. Full article

Wireless sensor networks (WSN) are useful in medicine for monitoring the vital signs of elderly patients. These sensors allow for remote monitoring of a patient’s state of health, making it easier for elderly patients, and allowing to avoid or at least to extend the interval between visits to specialized health centers. The proposed system is a low-cost WSN deployed at the elderly patient’s home, monitoring the main areas of the house and sending daily recommendations to the patient. This study measures the impact of the proposed sensor network on nine vital sign metrics based on a person’s sleep patterns. These metrics were taken from 30 adults over a period of four weeks, the first two weeks without the sensor system while the remaining two weeks with continuous monitoring of the patients, providing security for their homes and a perception of well-being. This work aims to identify relationships between parameters impacted by the sensor system and predictive trends about the level of improvement in vital sign metrics. Moreover, this work focuses on adapting a reactive algorithm for energy and performance optimization for the sensor monitoring system. Results show that sleep metrics improved statistically based on the recommendations for use of the sensor network; the elderly adults slept more and more continuously, and the higher their heart rate, respiratory rate, and temperature, the greater the likelihood of the impact of the network on the sleep metrics. The proposed energy-saving algorithm for the WSN succeeded in reducing energy consumption and improving resilience of the network. Full article

Generative AI (artificial intelligence) refers to algorithms and models, such as OpenAI’s ChatGPT, that can be prompted to generate various types of content. In this narrative review, we present a selection of representative examples of generative AI applications in medicine and healthcare. We then briefly discuss some associated issues, such as trust, veracity, clinical safety and reliability, privacy, copyrights, ownership, and opportunities, e.g., AI-driven conversational user interfaces for friendlier human-computer interaction. We conclude that generative AI will play an increasingly important role in medicine and healthcare as it further evolves and gets better tailored to the unique settings and requirements of the medical domain and as the laws, policies and regulatory frameworks surrounding its use start taking shape. Full article

The complex and interconnected infrastructure of smart cities offers several opportunities for attackers to exploit vulnerabilities and carry out cyberattacks that can have serious consequences for the functioning of cities’ critical infrastructures. This study aims to address this phenomenon and characterize the dimensions of security risks in smart cities and present mitigation proposals to address these risks. The study adopts a qualitative methodology through the identification of 62 European research projects in the field of cybersecurity in smart cities, which are underway during the period from 2022 to 2027. Compared to previous studies, this work provides a comprehensive view of security risks from the perspective of multiple universities, research centers, and companies participating in European projects. The findings of this study offer relevant scientific contributions by identifying 7 dimensions and 31 sub-dimensions of cybersecurity risks in smart cities and proposing 24 mitigation strategies to face these security challenges. Furthermore, this study explores emerging cybersecurity issues to which smart cities are exposed by the increasing proliferation of new technologies and standards. Full article

Nowadays, virtual reality is a new and rapidly developing technology that provides the opportunity for a new, more immersive form of data visualization. Evaluating digitized pathological serial sections and establishing the appropriate diagnosis is one of the key task of the pathologist in the daily work. The type of tools used by pathologists in the evaluation of samples has not changed much in recent decades. On the other hand, the amount of information required to establish an accurate diagnosis has been significantly increased. Nowadays, pathologists are working with the help of multiple high-resolution desktop monitors. Instead of the large screens, the use of virtual reality can serve as an alternative solution, which provides virtualized working space for pathologists during routine sample evaluation. In our research, we defined a new immersive working environment for pathologists. In our proposed solution we visualize several type of digitized medical image data with the corresponding meta data in 3D, and we also defined virtualized functions that support the evaluation process. The main aim of this paper is to present the new possibilities provided by 3D visualization and virtual reality in digital pathology. The paper presents a new virtual reality-based examination environment, as well as software functionalities that are essential for 3D pathological tissue evaluation. Full article