Ghayth AlMahadin

Reported accuracies for machine-learning intrusion detection on ToN-IoT cluster between 97% and 99.9%. We ask what those figures measure, auditing ToN-IoT's complete network and Windows host-telemetry releases alongside NSL-KDD, CIC-IDS2017, UNSW-NB15 and BoT-IoT under one leakage-controlled protocol. First, six machine-configuration counters that no attack can causally alter (commit limit, resident kernel code, page-table entries) classify the Windows subsets at 94.84% and 96.88% accuracy, against majority-class baselines of 79.16% and 69.16%; each campaign occupies a disjoint capture window, so slowly varying machine state encodes the label. Session-assigned labelling is a known hazard in network data; here it reaches host telemetry through an ordinary performance counter no feature-selection heuristic would flag. Second, the 22,339,021-flow network release is 65.7% exact duplicate rows once identifiers are excluded; its ransomware class comprises 72,805 rows built from 322 distinct vectors, 99.56% redundant, against 93.32% for the KDD CUP 99 records whose redundancy retired it. Third, retaining addresses and ports raises Random Forest accuracy from 0.9694 to 0.9981 under an identical split. NSL-KDD serves as a positive control: 0.0% duplication, accuracy differing by 0.0017 between protocols, so the audit discriminates rather than flagging datasets indiscriminately. We release the tooling, corrected splits, reference results and three pre-benchmark checks.

Internet of things (IoT) devices enhance quality of life and industrial operations but pose significant security risks, necessitating intelligent intrusion detection systems (IDS) to combat evolving cyber threats. This paper proposes a novel IDS framework integrating bio-inspired heuristic feature selection, a generative adversarial network (GAN)-based data augmentation, and an ensemble classifier combining ResNet, AlexNet, and MobileNet. The methodology, tested on the botnet (BoT)-IoT dataset, follows four stages: preprocessing, feature augmentation, feature selection, and ensemble classification. Evaluated on benchmarks including CIC-IDS-2018, NSL-KDD, and UNSW-NB15, the model achieved accuracies of 98.2%, 99.1%, 97.6%, and 98.4%, respectively, with consistently high precision, recall, and F1-scores, demonstrating robust detection of diverse cyberattacks. Beyond accuracy, the framework optimizes processing time for large-scale IoT data, addressing scalability challenges in real-time threat mitigation. By synergizing feature optimization, synthetic data generation, and deep learning architectures, the solution enhances detection rates while minimizing computational overhead. Comparative analysis highlights its superior performance over existing methods, positioning it as a vital tool for securing IoT ecosystems against unauthorized access and malicious activities. The results underscore its potential to fortify IoT network security, balancing efficiency, adaptability, and computational feasibility for practical deployment in resource-constrained environments.

Deep learning has emerged as a potent tool for various tasks, such as image classification. However, in the medical domain, there exists a scarcity of data, which poses a challenge in obtaining a well-balanced and high-quality dataset. Commonly seen issues in the realm of renal health include conditions such as kidney stones, cysts, and tumors. This study is centered on the examination of deep learning models for the purpose of classifying renal computed tomography (CT)-scan pictures. State-of-the-art classification models, such as convolutional neural network (CNN) approaches, are employed to boost model performance and improve accuracy. The algorithm is comprised of six convolutional layers that progressively increase in complexity. Every layer in the network utilizes a uniform 3x3 kernel size and applies the rectified linear unit (ReLU) activation function. This is followed by a max-pooling layer that downsamples the feature maps using a 2x2 pool size. Following this, a flatten layer was implemented in order to preprocess the data for the fully linked layers. The consistent utilization of uniform kernel sizes and activation functions throughout all layers of the model facilitated the smooth extraction of complex features, thereby enhancing the model’s ability to accurately identify different kidney conditions. As a result, we achieved a high accuracy rate of 99.8%, precision is 99.8%, and F1 score of approximately 99.7%.

The integration of internet of things (IoT) and big data technologies is transforming the overall perspective of managing various sectors of modern life; with higher educational sectors being no exception of this transformation. This paper explores the idea of a “smart university” as an extension of the overarching “smart city” framework, emphasizing the blending of IoT and big data technologies within higher education institutions. The study investigates the incorporation of IoT technologies throughout university campuses, including intelligent classrooms, smart infrastructure, and device networking. Moreover, the paper delves into the substantial role played by big data analytics in processing and extracting meaningful insights from extensive data generated by IoT devices in a Smart University. The use of predictive analytics, machine learning algorithms, and data-driven decision-making contributes to personalized learning experiences, adaptive campus management, and proactive maintenance of university facilities. Furthermore, this paper not only emphasizes the potential benefits of deploying IoT and big data in a university setting but also addresses challenges related to security, privacy, and ethical considerations. By embracing a comprehensive approach to technology integration, universities can leverage the capabilities of IoT and big data to establish intelligent, interconnected, and flexible educational environments that align with the broader vision of a smart city.

The treatment of brain tumors poses significant challenges and contributes to a significant number of deaths on a global scale. The process of identifying brain tumors in medical practice involves the visual analysis of photographs by healthcare experts, who manually delineate the tumor locations. However, this approach is characterized by its time-consuming nature and susceptibility to errors. In recent years, scholars have put forth automated approaches to early detection of brain tumors. However, these techniques face challenges attributed to their limited precision and significant false-positive rates. There is a need for an effective methodology to identify and classify tumors, which involves extracting reliable features and achieving precise disease classification. This work presents a novel model architecture that is derived from the EfficientNetB3. The suggested framework has been trained and assessed on a dataset consisting of 7,023 magnetic resonance images. The findings of this study indicate that the fused feature vector exhibits superior performance compared to the individual vectors. Furthermore, the technique that was provided showed superior performance compared to the currently available systems and attained a 100% accuracy rate. As a result, it is viable to employ this technique within a clinical environment for the purpose of categorizing brain tumors based on magnetic resonance images scans.

This Review explores the transformative impact of artificial intelligence (AI) on the early detection of skin cancer, with a specific focus on melanoma, a potentially lethal form of the disease. Beginning with an overview of traditional diagnostic methods and their limitations, this paper delves into the evolution of AI within dermatology, emphasizing its application in image analysis and pattern recognition. A comprehensive examination of AI algorithms for melanoma detection, including machine learning and deep learning models, is provided. This Review critically assesses the performance metrics, training datasets, and comparative analyses with traditional methods. Addressing challenges such as data quality, interpretability, and ethical considerations, this paper outlines future directions, emphasizing ongoing research, algorithm improvements, and integration with clinical practices. Case studies and success stories highlight the real-world impact of AI in dermatology. This Review concludes by summarizing key findings and underlining the pivotal role of AI in revolutionizing early melanoma detection, with implications for personalized medicine and enhanced patient outcomes.

The rise of Vehicular Ad-hoc Networks (VANETs) has led to the growing significance in intelligent transportation systems. This research suggests a deep learning model for anomaly detection based on GRU over VANET network traffic to address this challenge. Consumer electronics technologies can be successfully introduced to the market in one of two ways: either there is a clear benefit for the customer from using this technology, or it is required by a regulatory order that prevents the use of alternatives. It is possible to detect unknown assaults and DoS floods using traffic anomalies. Users can keep track of the security features of multimedia services by using Traffic Anomaly Detection, which provides an overview of traffic anomaly detection analysis. Anomaly detection methods fall into three categories: unsupervised, semi-supervised, and supervised. The right anomaly detection technique basically depends on the labels that are present in the dataset. To further improve the accuracy of proposed model, a new semi-supervised technique for detecting VANET network activity anomalies called SEMI-GRU has been proposed. The results demonstrate that proposed GRU-based deep learning model outperforms existing methods in detecting network anomalies with low false positive rates.