Intrusion Detection Systems (IDSs) are the cornerstone of cybersecurity, monitoring network traffic to find abnormal suspicious activities. Traditional IDSs usually face challenges in adapting to the cyber threats that evolve day by day, leading to very high false positive rates and missed detections. This study focuses on enhancing the performance of an IDS system by integrating deep learning techniques with time series data. The efficiency of RNN, CNN, and LSTM networks was evaluated in detecting intrusions in real-time. The experimental results showed that hybrid models, especially the CNN+RNN+LSTM combination, performed best with a 0.86 F1 score, 0.92 precision, and 0.79 recall, indicating that hybrid deep learning methods can improve detection accuracy while reducing false alarms, opening a resilient future for cybersecurity.

Explainable Artificial Intelligence (XAI) is critical for bridging the gap between opaque machine decision-making and human comprehension. Despite advances, most XAI systems deliver static explanations that fail to account for users' diverse expertise levels. This study proposes an adaptive XAI framework that personalizes explanations according to individual user expertise. Grounded in cognitive load theory and trust calibration principles, the system dynamically adjusts explanation complexity, depth, and modality. Through a mixed-methods experimental design involving 150 participants classified as novices, intermediates, and experts, results show that adaptive explanations significantly enhance understanding (+27%), trust calibration (+19%), and decision-making accuracy (+22%) compared to static explanations. The findings provide strong empirical support for user-centered XAI models and offer actionable design guidelines for adaptive explainability in critical sectors such as healthcare, finance, and education.

Fractional order differential equations often possess inherent symmetries that play a crucial role in governing their dynamics in a variety of scientific fields. In this work, we consider numerical solutions for fractional-order linear delay differential equations. The numerical solution is obtained via the Laplace transform technique. The quadrature approximation of the Bromwich integral provides the foundation for several commonly employed strategies for inverting the Laplace transform. The key factor for quadrature approximation is the contour deformation, and numerous contours have been proposed. However, the highly convergent trapezoidal rule has always been the most common quadrature rule. In this work, the Gauss–Hermite quadrature rule is used as a substitute for the trapezoidal rule. Plotting figures of absolute error and comparing results to other methods from the literature illustrate how effectively the suggested approach works. Functional analysis was used to examine the existence of the solution and the Ulam–Hyers (UH) stability of the considered equation.

The growing difficulty of ransomware attacks highlights the urgent need for intelligent, real-time detection strategies that can mitigate threats before data encryption occurs. Traditional single-source detection methods often fail to identify rapidly evolving ransomware variants during their early encryption stages. This paper presents a hybrid ransomware detection framework that integrates behavioral system Application Programming Interface (API) call sequences, network traffic features from Canadian Institute for Cybersecurity Intrusion Detection System dataset (CICIDS), and static Portable Executable (PE) attributes from the Elastic Malware Benchmark for Empowering Researchers dataset (EMBER). A time-windowed data modeling technique is used to capture early-stage ransomware behavior within the first few seconds of execution. The proposed framework employs a hybrid Recurrent Neural Network (RNN) and Bidirectional Long Short-Term Memory (BiLSTM) deep learning model to learn both sequential and contextual patterns across the fused feature space. Experimental results demonstrate high accuracy (0.96) and early detection capability within 3 s, significantly outperforming traditional machine learning baselines. The system effectively detects known and unknown ransomware families by leveraging a multi-perspective view of ransomware activity. This study demonstrates the potential of hybrid feature integration and time-sensitive learning in enhancing early ransomware detection, contributing to the development of robust, real-time cybersecurity defense systems.