Cross-site scripting adversarial attacks based on deep reinforcement learning: Evaluation and extension study

📰 ArXiv cs.AI

Deep reinforcement learning is used to evaluate and extend cross-site scripting adversarial attacks

advanced Published 23 Mar 2026
Action Steps
  1. Identify the vulnerabilities of deep learning models in detecting XSS attacks
  2. Employ deep reinforcement learning to generate adversarial attacks
  3. Evaluate the effectiveness of these attacks on various models
  4. Extend the study to improve the robustness of models against such attacks
Who Needs to Know This

Security teams and AI engineers can benefit from this research to improve web application security and develop more robust models to detect XSS attacks

Key Insight

💡 Adversarial attacks can exploit the discontinuous nature of DL models, making them vulnerable to XSS attacks

Share This
💡 Deep reinforcement learning generates powerful XSS adversarial attacks

Key Takeaways

Deep reinforcement learning is used to evaluate and extend cross-site scripting adversarial attacks

Full Article

Title: Cross-site scripting adversarial attacks based on deep reinforcement learning: Evaluation and extension study

Abstract:
arXiv:2502.19095v2 Announce Type: replace-cross Abstract: Cross-site scripting (XSS) poses a significant threat to web application security. While Deep Learning (DL) has shown remarkable success in detecting XSS attacks, it remains vulnerable to adversarial attacks due to the discontinuous nature of the mapping between the input (i.e., the attack) and the output (i.e., the prediction of the model whether an input is classified as XSS or benign). These adversarial attacks employ mutation-based st
Read full paper → ← Back to Reads