Scholarlink

<p>Intrusion Detection Systems (IDS) serve a vital role in protecting modern networks against increasingly complex and evolving cyberattacks. While existing IDS solutions rely on Machine Learning (ML) and Deep Learning (DL) techniques, they often struggle to capture highly complex feature dependencies and are vulnerable to privacy and scalability challenges. In response to these concerns, this work presents an Electric Eel Brown-Bear Algorithm-based Quantum Dilated Convolutional Neural Network (EEBA_QDCNN) for intrusion detection. The framework comprises of quantum devices and Quantum Machine Learning (QML) service provider. Each quantum device collects log files and processes them using a local model. The trained local models are sent to the QML cloud, where a Privacy-Quantum Aware Model evaluates the data with quantum noise optimally selected by the Electric Eel Brown-Bear Algorithm (EEBA), which integrates Electric Eel Foraging Optimization (EEFO) and Brown-Bear Optimization Algorithm (BOA). A global updater modifies the global model and redistributes it to local models for further training. Locally, log data is normalized using Peldschus Normalization, features are selected utilizing EEBA, and data is augmented employing Synthetic Minority Over-sampling Technique with Edited Nearest Neighbors (SMOTE-ENN). Finally, intrusion is detected utilizing a Quantum Dilated Convolutional Neural Network (QDCNN), which is trained utilizing EEBA. The proposed EEBA_QDCNN framework is evaluated using the network intrusion detection dataset and intrusion detection dataset comprising a wide variety of intrusions simulated in a military network environment. Moreover, EEBA_QDCNN has achieved maximal accuracy and precision of 97.468%, and 96.577%, as well as minimal Mean Squared Error (MSE) and Root Mean Squared Error (RMSE) of 0.021, and 0.147.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Privacy-Preserving Quantum Learning Framework for Intrusion Detection using Electric Eel Brown-Bear Algorithm based Deep Learning

  • 期刊论文

  • 发表日期：

    2026年6月19日

查看全文
  • P. Mano Paul,
  • I. Diana Jeba Jingle,
  • P. Balashanmuga Vadivu,
  • Daniya Thavasilingam

摘要

Intrusion Detection Systems (IDS) serve a vital role in protecting modern networks against increasingly complex and evolving cyberattacks. While existing IDS solutions rely on Machine Learning (ML) and Deep Learning (DL) techniques, they often struggle to capture highly complex feature dependencies and are vulnerable to privacy and scalability challenges. In response to these concerns, this work presents an Electric Eel Brown-Bear Algorithm-based Quantum Dilated Convolutional Neural Network (EEBA_QDCNN) for intrusion detection. The framework comprises of quantum devices and Quantum Machine Learning (QML) service provider. Each quantum device collects log files and processes them using a local model. The trained local models are sent to the QML cloud, where a Privacy-Quantum Aware Model evaluates the data with quantum noise optimally selected by the Electric Eel Brown-Bear Algorithm (EEBA), which integrates Electric Eel Foraging Optimization (EEFO) and Brown-Bear Optimization Algorithm (BOA). A global updater modifies the global model and redistributes it to local models for further training. Locally, log data is normalized using Peldschus Normalization, features are selected utilizing EEBA, and data is augmented employing Synthetic Minority Over-sampling Technique with Edited Nearest Neighbors (SMOTE-ENN). Finally, intrusion is detected utilizing a Quantum Dilated Convolutional Neural Network (QDCNN), which is trained utilizing EEBA. The proposed EEBA_QDCNN framework is evaluated using the network intrusion detection dataset and intrusion detection dataset comprising a wide variety of intrusions simulated in a military network environment. Moreover, EEBA_QDCNN has achieved maximal accuracy and precision of 97.468%, and 96.577%, as well as minimal Mean Squared Error (MSE) and Root Mean Squared Error (RMSE) of 0.021, and 0.147.