The rapid growth of sophisticated Android malware (AM) threats is significant, as Android devices often store private and sensitive personal and financial information. These threats allow stealing of data, interference with device functioning, and network compromise. One of the greatest difficulties in efficient interception systems is ensuring a high level of detection accuracy for distinguishable AM variants. This study focuses on developing a robust Android malware detection model via machine learning (ML) and deep learning (DL) techniques. The model combines ML classifiers, which consist of logistic regression (LR) and decision trees (DTs), and a DL classifier, an artificial neural network (ANN). The model was implemented via an open-source data mining program called Orange. The NATICUSdroid dataset was used to train and test the model, which was measured in terms of accuracy, precision, recall, F-measure and AUC. The experimental findings revealed that the ANN performed the best (accuracy: 98.0%, precision/recall/F-measure: 98.0%, AUC: 0.997) and was better than the LR (accuracy: 96.1%, AUC: 0.989) and DT (accuracy: 96.0%, AUC: 0.971) methods. The results highlight the high potential of DL-based approaches, especially ANNs, to detect Android malware and reinforce their suitability for enhancing mobile security systems.

Autism spectrum disorder (ASD) poses a complex challenge to researchers and practitioners, with its multifaceted etiology and varied manifestations. Timely intervention is critical in enhancing the developmental outcomes of individuals with ASD. This paper underscores the paramount significance of early detection and diagnosis as a pivotal precursor to effective intervention. To this end, integrating advanced technological tools, specifically eye-tracking technology and deep learning algorithms, is investigated for its potential to discriminate between children with ASD and their typically developing (TD) peers. By employing these methods, the research aims to contribute to refining early detection strategies and support mechanisms. This study introduces innovative deep learning models grounded in convolutional neural network (CNN) and recurrent neural network (RNN) architectures, employing an eye-tracking dataset for training. Of note, performance outcomes have been realised, with the bidirectional long short-term memory (BiLSTM) achieving an accuracy of 96.44%, the gated recurrent unit (GRU) attaining 97.49%, the CNN-LSTM hybridising to 97.94%, and the LSTM achieving the most remarkable accuracy result of 98.33%. These outcomes underscore the efficacy of the applied methodologies and the potential of advanced computational frameworks in achieving substantial accuracy levels in ASD detection and classification.

The increasing sophistication of cyber threats necessitates the development of advanced security mechanisms to protect modern networks. Among these threats, spoofing attacks pose a significant risk by enabling malicious actors to impersonate legitimate entities. To address this challenge, we propose a novel Transformer-based deep learning framework designed for the effective detection of spoofing attacks. The core of our novel model is a Transformer neural network, enhanced with a custom attention mechanism to improve feature extraction and classification accuracy. To enhance model interpretability and foster trust in AI-driven security systems, we integrate Explainable AI (XAI) techniques, specifically SHAP analysis, allowing for a deeper understanding of feature contributions in decision-making. The proposed model utilized the CIC IoMT2024 dataset, a benchmark with limited prior research on spoofing attack detection. Further, our approach incorporates comprehensive data preprocessing techniques and employs over-sampling using the synthetic minority oversampling technique (smote) and cleaning using (tomek) these techniques are integrated into links smotetomek to mitigate class imbalance, ensuring a more representative training dataset. The proposed framework is evaluated using benchmark dataset datasets, demonstrating high binary classification performance in spoofing attacks through key metrics such as accuracy, confusion matrix analysis, and other classification benchmarks. The proposed model archived an exact result with Accuracy 99.71%. The findings highlight the potential of Transformer-based architectures in cybersecurity applications, paving the way for real-time threat detection and adaptive defense mechanisms.

The proposed CapsNet Model excels in capturing spatial dependencies and hierarchical feature relationships, surpassing the capabilities of traditional convolutional architectures.

The nano-scale Internet of Things (nano-IoT) is ushering in a new era of applications in areas such as biomedical sensing, smart dust, and embedded environmental monitoring. Unfortunately, the utilization of nano-devices is barely feasible because of their very limited energy, computation, memory, and communication bandwidth, making existing cryptographic authentication methods infeasible. This study introduces an ultra-lightweight identity authentication protocol designed for nano-IoT systems with limited resources. The proposed protocol uses symmetric key cryptography, one-way hash functions, and XOR operations to achieve mutual authentication with low computational and communication costs. A highly modular architecture is proposed, consisting of 4 main stages: cryptographically secure pre-loading of credentials, a hash-based mutual authentication scheme, a stateless session update mechanism using hash chains and nonces, and a gateway-level revocation enforcement model. In contrast to PUF- or ECC-based protocols, the proposed protocol is extremely resistant to impersonation, replay, man-in-the-middle, desynchronization, and side-channel attacks without requiring special hardware. Simulations in nano-IoT settings show that the proposed protocol is more than 8 times faster and 11 times more energy-efficient than public-key-based algorithms, with high scalability and robust security. This study offers a solid basis for the secure bootstrapping of the forthcoming nano-IoT in dynamic, low-power, and latency-constrained environments.

Many of the images that can be accessed through web search engines or social media networking sites are rare and not high quality because they are endangered or disappear. There must be a way to increase the quality of these images and conduct experiments to reduce noise, remove blur, and make them sharper to reach high-quality surfaces. Approaches that seek to achieve better results compete to increase the efficiency of those low-resolution images and generate images with the same color (RGB) characteristics but with higher quality. Deep learning algorithms, especially the use of convolutional neural networks (CNNs), have achieved advanced results within this context. In this approach, we propose a powerful base model UIR for image recovery by using conventional neural networks (CNNs) added to conventional algorithms for ultrasupper-resolution from low-resolution images by extracting the feature map from a low-resolution image ILR as overlapping superresolution ISR patches, in which every patch represents a high-dimensional vector. The missing features of the pixels that occur during the training process are subsequently compensated via the residual Swin Transformer block (RSTB). The results of quantitative evaluation experiments using PSNR(db)/SSIM metrics were superior to those of state-of-the-art methods on benchmark datasets (Set5, Set14, and BSD100). The selected images have a magnification of x2, resulting in values of (36.86(db)/0.9739, 36.10(db)/0.9656, 34.74(db)/0.9893) and x4, resulting in values of (34.44(db)/0.9784, 27.71(db)/0.8894, and 26.87(db)/0.9915, respectively. The results of the visual comparison also revealed that the texture of the surfaces is sharper, more expressive, less noisy, and blurry than those of the other methods.