Pacific States University

A brain tumour is an abnormal or mass growth of cells in the brain. It can be benign or malignant. This problem, if not identified in its early stages, can be fatal to the patient. Detection of Brain tumours in their early stages has become a major challenge in the realm of healthcare. This research uses about 4600 images of brain tumour. Initially, the data labels are encoded and the images are normalized. Pre-existing Neural Network models such as ResNet-50 and AlexNet are trained to predict the presence of Brain tumour. The respective architectures have been used instead of the pre-trained model. In this study, the batch sizes of the defined models are varied and the performance is compared based on the model loss and accuracy. The AlexNet model resulted in a testing accuracy of 0.96, 0.95 and 0.96 for the batch sizes 16,32 and 64 respectively whereas a testing accuracy of only 0.56 resulted for a batch size of 8. But ResNet-50 model, performed well for all batch sizes of 8, 16, 32 and 64 yielding an accuracy of 0.94,0.69,0.97 and 0.96 respectively depicting itself to be a suitable model for brain tumour classification.

The blistering development of cyber threats as far as advanced malware, adversarial attacks and zero-day exploits are concerned have revealed the severe weaknesses of traditional machine learning-based defence systems. Adversarial Machine Learning (AML) has become a promising paradigm that can result in the increased robustness and resilience of the cybersecurity frameworks by allowing the system to anticipate, detect, and mitigate adversarial manipulations. The paper provides a detailed AML-based cyber defence framework, which combines adversarial training, anomaly detection, and robust optimization algorithms to enhance the security of the system against evasion and poisoning. The suggested architectural design is based on the deep neural networks that are enhanced with adversarial sample generation and defensive distillation to enhance the detection accuracy under the adversarial circumstances. Also, there is a dynamic feedback mechanism to allow perpetual learning and adaptation to changing threat environments. The experimental analyses prove that the suggested method has a substantial positive impact on the detection performance, lower false positive rates, and is more stable to adversarial perturbations than the conventional models. The results demonstrate the possibilities of AML as a vital enabler to nextgeneration cyber defence systems to ensure reliability, scalability, and resiliency in complex and dynamic digital ecosystems.

Modern cloud, edge, and cyber-physical systems are based on large-scale distributed systems, which pose a larger attacker target due to their inherent heterogeneity, dynamic topology, and decentralized control surfaces, and their high scalability complicates traditional security enforcement. Traditional perimeter based and stationary defence systems are becoming less effective in combating against complex and multi-vector cyber threats to distributed environments. The current paper depicts a robust cybersecurity architecture that is aimed to provide adaptive resistance to threats, fault tolerance, and assurance of continuous trust in large-scale distributed system. The suggested architecture combines the concepts of zero-trust, adaptive trust assessment, distributed anomaly detection, and policy-based automated response to obtain real-time security coordination between geographically dispersed nodes. The architecture supports dynamically adjusting to changing attack patterns and ensures the availability and performance of the system by taking advantage of behaviour-aware monitoring, risk-based access control, and self-healing security workflows. An overall assessment is done based on the latency, scalability, resilience, and security-overhead parameters under different threats intensity and system scale. Empirical evidence shows that the given framework can greatly enhance attack detection precision, decrease the mean time to mitigation, and overall resiliency of the system in comparison with the traditional distributed security models. The results create a scalable and futural-fit base to achieve the next-generation distributed infrastructures, such as multi-cloud, edge–IoT, and mission-critical cyber-physical systems.

Histopathology-based diagnosis of lung and colorectal cancer is labor-intensive and subject to observer variability, motivating accurate and interpretable computer-aided diagnosis. In this paper, we propose a lightweight Histopathology Attention Network (HistoAttn-Net) for five-class lung and colon tissue classification on the LC25000 dataset. The model uses a compact convolutional backbone to extract local texture patterns and a multi-head self-attention block to capture long-range tissue context. To further improve robustness, we adopt a performance-aware ensemble strategy that selects the top-performing models from 10-fold cross-validation and aggregates their predictions via mean probability and majority voting. Explainability is provided through Grad-CAM applied at both the single-model and ensemble levels, yielding class-specific heatmaps that highlight clinically meaningful regions. On the held-out test set, the best single HistoAttn-Net achieves 99.6% accuracy, 99.5% macro Fl-score, and Cohen’s k of 0.996, while the top 3 ensemble attains 100% accuracy, macro Fl-score, and k, with macro ROC-AUC effectively equal to 1.0. These results indicate that the proposed framework offers an attractive balance of accuracy, efficiency, and interpretability for histopathology-based cancer diagnosis.

The scalability, flexibility, and cost-efficiency of cloud computing have made it an essential platform in today's digital age. But, with the rise of cloud-based applications, there have been great concerns about data security, unauthorized access, and cyber threats. Distributed cloud infrastructures can pose significant privacy leakage risks and vulnerabilities to traditional security mechanisms, especially against sophisticated attacks and insider threats. In this paper, we propose a Privacy-Aware Data Protection Model for Cloud Environments that combines encryption techniques, access control policies, anonymization strategies, and intelligent threat detection mechanisms to guarantee secure data storage and transmission. The proposed framework aims to improve the security aspects of sensitive cloud data including confidentiality, integrity, and availability using hybrid cryptographic algorithms and machine learning based anomaly detection. In addition, the model includes privacy-preserving authentication and dynamic trust evaluation to alleviate risks of multi-tenant cloud systems. The experimental study shows that the proposed approach increases the accuracy of data protection, reduces the number of attempts to access the data unauthorizedly and increases the reliability of the data protection system while keeping the computational performance efficient. The results show that the proposed privacy-preserving approach can be a viable solution for cloud data management in the modern enterprise to achieve secure and scalable management.