Literature Review: The Application of Deep Learning in Web Attack Detection

This study is motivated by the increasing need for web security systems, the rapid development of web-based applications, and the rise in cyber threats, such as cross-site scripting (XSS), SQL injection (SQLi), and HTTP manipulation. This study reviews the application of deep learning methods in web attack detection with a focus on frequently used algorithms, model performance, and challenges faced. This study uses a systematic literature review approach by analyzing various relevant scientific publications. Data were collected from reliable literature sources and analyzed qualitatively to identify patterns, methods, and research gaps. The results indicate that convolutional neural networks (CNNs), recurrent neural networks (RNNs), and long short-term memory (LSTM) algorithms are the most widely used methods and exhibit good performance in detecting web attacks. However, several obstacles remain, such as limited representative datasets, data imbalance, and high computational requirements. This study contributes in the form of a comprehensive synthesis of the application of deep learning in web security and identifies research gaps and development opportunities to improve model effectiveness and efficiency, particularly in the implementation of real-time attack detection.