As cyber threats in IoT networks become increasingly sophisticated, identifying Man-in-the-Middle (MitM) and Distributed Denial of Service (DDoS) attacks with high precision is critical. This study pioneers the application of a Transformer-based model, fortified with self-attention mechanisms, to serve as a state-of-the-art intrusion detection system. This research specifically excels in feature selection and sequence classification within IoT network traffic data. Data is transformed into sequential representations through preprocessing, facilitating the model's ability to capture temporal relationships and contextual understanding. The Transformer model, tailored for sequence classification, auto-learns essential features from the processed data. Multi-headed self-attention allows the model to focus on varied aspects of network data, enhancing detection capabilities. The model is fine-tuned using a cross-entropy loss function on a labeled dataset comprising MitM and DDoS instances, as well as benign traffic. During inference, the model assigns anomaly scores to sequences, flagging those with high scores as likely attacks. Trained and evaluated on the recently released Edge-IIoTset dataset, the model achieved an accuracy rate of 98.50%. In comparisons, it outperformed traditional classifiers like Support Vector Machines, Long Short-Term Memory, and Deep Recurrent Neural Networks. The model attained an F1 score of 99.39%, evidencing its efficacy in balancing false positives and negatives—an essential metric in network security. This study highlights the potential of deploying Transformer-based models in intrusion detection systems. The system's performance underscores a new paradigm in proactive cybersecurity measures, opening avenues for adaptive network defense mechanisms. Future research could optimize the model's architecture and expand its applications across cybersecurity domains.

For Student mental health has emerged as a critical concern in higher education, this study developed a web-based application that leverages a Generative AI model to address the need to identify student mental health status. Traditional methods of mental health monitoring and support may not be sufficient to address the growing needs of students. This paper explores the potential of generative AI to develop a proactive mental health monitoring system. By utilizing a standardized questionnaire based on the American Psychiatric Association's (APA) DSM-5-TR Self-Rated Level 1 Cross-Cutting Symptom Measure—Adult (DSM XC), the system collects and analyzes student mental health data to identify potential issues and provide timely interventions. This paper discusses the development and implementation of this AI-powered system, its potential benefits, and ethical considerations.

Traditional cybersecurity education often emphasizes theoretical knowledge but lacks practical experience, creating a significant skills gap. This research addresses this gap by proposing the design and development of an open-source cyber range specifically tailored for educators and security professionals, in order to further aid their preparedness towards defending against sophisticated cyber-attacks in this ever-evolving cyber threat landscape. Leveraging open-source technologies, the anticipated outcome is a user-friendly and effective open-source cyber range platform that enhances cybersecurity education and training for educators and security professionals, ultimately contributing to a more secure digital environment. Through a comprehensive literature review and comparative analysis, the project will also identify key features and functionalities required for effective cyber range training. Agile development methodologies will be employed as well, incorporating the STRIDE threat modelling framework to ensure robust security throughout the design phase. Some of the functionalities the developed platform will provide include scenario creation, attack simulation, performance evaluation, and reporting. Several labs simulating real-world security challenges will be developed for user testing and platform evaluation. Iterative testing would also be part of the process.

The growth of the urban population has fueled the development of smarter cities. Smart cities built on blockchain have proven to have increased resilience to security threats, increased reliability, and many other advantages. This paper examines how blockchain has been implemented in the smart city infrastructure to date. Looking at existing research between 2013 and 2022, we investigate how blockchain research has evolved in a decade, especially the triggers for the boom in research in 2017, and how the trends have changed to date. Although by 2018, researchers were trying different concepts and exploring the potential in different fields, the COVID-19 pandemic was the only strong external factor that caused a shift in the trend toward smart health. However, despite the pandemic, the smart health research spike was very brief as there was more research on smart economy and smart living. Therefore, as blockchain expands, there are a lot of challenges that need to be resolved. Thus, this paper also discusses the challenges associated with blockchain-based smart cities and possible solutions using AI convergence.

The rising dropout rate in universities poses a significant challenge for educational institutions. This study aims to identify the reasons behind student dropouts at universities. We utilized a dataset of undergraduates from 17 different courses, encompassing 36 attributes. To achieve maximum classification accuracy, we proposed a wrapper feature selection technique to identify the most important subset of features. Our approach employs the Whale Optimization Algorithm combined with a neural network model (WOA-NN). In this study, we compared the performance of this hybrid approach with six different machine learning algorithms (Random Forest, Logistic Regression, Naive Bayes, Support Vector Machine, Decision Tree, and K-Nearest Neighbors) and a standalone neural network model. Based on the results of our experiments, we recommend the WOA-NN wrapper technique for feature selection. The main objective of the feature selection process is to identify the smallest subset of features that achieves maximum classification accuracy. We trained the neural network model using both the full set of features and the selected subset. The performance of the neural network model with the selected features was significantly better than with the full feature set, resulting in an improvement in accuracy from 92.13\% to 96.28\%. This enhanced accuracy enabled us to identify key factors contributing to student dropout in higher education.