Department of Computer Science, Louisiana State University Shreveport, USA.
Research Article
World Journal of Advanced Research and Reviews, 2024, 24(03), 1116-1132
Received on 03 November 2024; revised on 11 December 2024; accepted on 13 December 2024
Deep learning (DL) has changed the cybersecurity domain by providing sophisticated tools for detecting and mitigating an evolving landscape of cyber threats. This study explores the application of deep learning techniques, including convolutional neural networks (CNNs) and recurrent neural networks (RNNs), in real-time threat detection and response. These models excel in identifying patterns and anomalies within vast and complex datasets, enabling accurate detection of malware, phishing attempts, and insider threats. Their ability to autonomously learn from diverse sources, such as network traffic, user behaviour, and system logs, enhances the efficacy of cybersecurity systems. Despite these advancements, the field faces significant challenges, including adversarial attacks designed to exploit vulnerabilities in deep learning algorithms. These attacks manipulate input data to deceive models, potentially bypassing security mechanisms and compromising critical systems. Addressing this issue requires a multi-faceted approach, integrating robust training methods, data augmentation, and defensive mechanisms such as adversarial training and gradient masking. Furthermore, explainability and interpretability of deep learning models remain crucial for building trust and improving decision-making in security operations. The paper also emphasizes the importance of a proactive, layered defense strategy to counteract sophisticated cyber threats. This includes combining deep learning with traditional cybersecurity measures and incorporating threat intelligence to enhance system resilience. By bridging the gap between state-of-the-art DL methodologies and practical applications in cybersecurity, this research provides a roadmap for improving threat detection and response capabilities, ultimately contributing to the development of secure, adaptive, and resilient cyber infrastructures.
Deep Learning; Cybersecurity; Adversarial Attacks; Threat Detection; Neural Networks; Resilience Strategies
Preview Article PDF
Maureen Oluchukwuamaka Okafor. Deep learning in cybersecurity: Enhancing threat detection and response. World Journal of Advanced Research and Reviews, 2024, 24(3), 1116-1132. Article DOI: https://doi.org/10.30574/wjarr.2024.24.3.3819