The utilization of deep convolutional neural networks for the purpose of diagnosing diseases in the skin area has proven to yield similar accuracy levels to those obtained by dermatologists in different studies. Nevertheless, many challenges are still present, including underperformance and poor generalization in some cases, as well as low interpretability related to the use of black box models. This creates major obstacles for practical implementation since it requires explainability in addition to accurate diagnostics, making it necessary to find a solution. To overcome the mentioned difficulties, an explainable deep learning framework for skin lesion classification (EDLF-SLC) is developed within this study. The framework makes use of various approaches in machine learning and explainable artificial intelligence (XAI) to ensure improvements in terms of both accuracy and interpretability in a three-stage manner. In the initial stage, deep representations extracted from multiple pretrained CNN architectures are fused to preserve complementary discriminative information learned by different network architectures before classification using an SVM with a radial basis function kernel. Next, Support Vector Machine (SVM) classifiers with a radial basis function kernel are used to classify the obtained features. Finally, the predictions made by the model are interpreted through local interpretable model-agnostic explanations (LIME). Experimental results show that EDLF-SLC reaches 88.0% in accuracy, 89.0% in precision, 87.0% in recall, and 88.0% in F1 score, demonstrating competitive performance compared with several recently reported methods under the experimental conditions considered in this study.
Hasan Hashim, M. Rokaya, Mohammed Farsi et al.· Journal of Visualized Experi...· 0 citations
In the face of increasingly sophisticated cyber threats in an interconnected world, the need for scalable, intelligent, real-time cybersecurity solutions is intensifying. While traditional deep learning, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), is effective for detecting anomalies in network traffic, their current capabilities do not address challenges with high-dimensionality networks, computational performance, or adaptation to new attacks. Whereas traditional methods suffered constraints, quantum computing offers a complementary advantage in both parallel computing and superior optimization. This study provides a hybrid deep learning and quantum computing framework that employs (i) convolutional neural networks (CNNs) for effectively extracting spatial features, (ii) recurrent neural networks (RNNs) for leveraging temporal patterns, and (iii) Variational Quantum Circuits (VQCs) for quantum-enhanced classification. The framework is built using the CICIDS-2017 dataset to extract features in an extensive preprocessing phase but is a rather laborious process of normalization, feature selection, feature time structuring, and a feature balancing procedure using techniques such as on-over sampling and under-sampling. The hybrid model was evaluated across multiple metrics; for accuracy, recall, F1 score, and ROC-AUC area showed high accuracy (93%) and low false-positive rates across different categories of cyberattacks within an experimental and virtual cyber environment. The suggested framework proved to be generalizable and scalable, suggesting its efficiency in cybersecurity applications that necessitate flexibility and intelligence. This contribution demonstrates a foundational step towards real-time, high performance hybrid quantum - deep learning models for cyber threat detection and prevention.
Abdel-Haleem Abdel-Aty, Mohammed Farsi, M. Hafez et al.· Scientific Reports· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
