A web attack detection method, device and equipment based on double encoders
By employing a dual-encoder web attack detection method, which performs deep semantic encoding and fusion on HTTP requests and responses respectively, the method solves the problem of insufficient accuracy caused by isolated analysis of HTTP requests and responses. It achieves end-to-end collaborative detection, improving the accuracy of attack determination and the adaptability of the model.
Patent Information
- Application Number
- CN202610540287.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-22
- Publication Date
- 2026-07-17
AI Technical Summary
Existing web attack detection methods suffer from insufficient accuracy in determining attack success due to isolated analysis of HTTP requests and responses, and struggle to adapt to new attack patterns in dynamic threat environments.
A web attack detection method based on dual encoders is adopted. By using two pre-trained language models to perform deep semantic encoding on HTTP requests and responses, the semantic representations of requests and responses are extracted and fused. Combined with incremental training and feature fusion mechanisms, end-to-end collaborative detection is achieved.
It improves the accuracy and reliability of attack success determination, enhances the ability to identify complex and covert attacks, and improves the model's adaptability in dynamic threat environments.
Smart Images
Figure CN122419863A_ABST