TY - GEN
T1 - Harnessing the Power of Quantum Computing for URL Classification
T2 - 20th EAI International Conference on Security and Privacy in Communication Networks, SecureComm 2024
AU - Qayyum, Tariq
AU - Tariq, Asadullah
AU - Haseeb, Muhammad Waqas
AU - Alrabae, Saed
AU - Trabelsi, Zouheir
AU - Sallabi, Farag
AU - Serhani, Mohamed
N1 - Publisher Copyright:
© ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2026.
PY - 2026
Y1 - 2026
N2 - Classifying URLs is crucial for enhancing network security by identifying and blocking harmful websites. This research contrasts the effectiveness of traditional machine learning models, a basic convolutional neural network (CNN), and a quantum convolutional neural network (QCNN) in URL categorization. QCNN leverages quantum mechanics to enhance computational efficiency. QCNN transforms classical data into quantum states using amplitude encoding and employs quantum-specific convolutional and pooling layers to extract features. The traditional machine learning (ML) models examined are Decision Tree, Random Forest, AdaBoost, K-Nearest Neighbors, Stochastic Gradient Descent, Extra Trees, and Gaussian Naïve Bayes. We assessed the performance of these models using accuracy, precision, recall, and F1 score measures. Findings suggest that the QCNN surpasses other models in all these metrics. While the basic CNN and grouped methods like Random Forest and Extra Trees also exhibit high performance, the QCNN’s exceptional results stem from its capacity to recognize intricate data patterns through quantum mechanics, enhancing its classification capabilities. Hence, the QCNN emerges as a leading choice for practical uses in flagging harmful URLs, with the basic CNN and grouped methods as effective alternatives.
AB - Classifying URLs is crucial for enhancing network security by identifying and blocking harmful websites. This research contrasts the effectiveness of traditional machine learning models, a basic convolutional neural network (CNN), and a quantum convolutional neural network (QCNN) in URL categorization. QCNN leverages quantum mechanics to enhance computational efficiency. QCNN transforms classical data into quantum states using amplitude encoding and employs quantum-specific convolutional and pooling layers to extract features. The traditional machine learning (ML) models examined are Decision Tree, Random Forest, AdaBoost, K-Nearest Neighbors, Stochastic Gradient Descent, Extra Trees, and Gaussian Naïve Bayes. We assessed the performance of these models using accuracy, precision, recall, and F1 score measures. Findings suggest that the QCNN surpasses other models in all these metrics. While the basic CNN and grouped methods like Random Forest and Extra Trees also exhibit high performance, the QCNN’s exceptional results stem from its capacity to recognize intricate data patterns through quantum mechanics, enhancing its classification capabilities. Hence, the QCNN emerges as a leading choice for practical uses in flagging harmful URLs, with the basic CNN and grouped methods as effective alternatives.
KW - MAchine Learning
KW - Malicious URLs
KW - QCNN
KW - Quantum Computing
UR - https://www.scopus.com/pages/publications/105023329175
UR - https://www.scopus.com/pages/publications/105023329175#tab=citedBy
U2 - 10.1007/978-3-031-94445-1_19
DO - 10.1007/978-3-031-94445-1_19
M3 - Conference contribution
AN - SCOPUS:105023329175
SN - 9783031944444
T3 - Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST
SP - 357
EP - 375
BT - Security and Privacy in Communication Networks - 20th EAI International Conference, SecureComm 2024, Proceedings
A2 - Alrabaee, Saed
A2 - Choo, Kim-Kwang Raymond
A2 - Damiani, Ernesto
A2 - Deng, Robert H.
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 28 October 2024 through 30 October 2024
ER -