🔥 A Deep Learning Approach to (WAF) -
Web Application Firewall
Web applications receives a variety of input parameters from users and attackers which generally inject malicious payloads trying to whether steal data or manipulate the application on their favor.
An approach widely used to protect the application is the WAF (web application firewall), which generally is rule based. In the present work, deep learning and Transformer, which is a state-of-the-art model in natural language processing, are applied in the scenario of hard-coded rules, successfully evading anti-WAF techniques. The proposed method achieves 96.5% accuracy in the proposed dataset.
Blog post
Code