Adversarial AI Comparison Hub
3 papers - avg viability 5.7
Top Papers
- NetDiffuser: Deceiving DNN-Based Network Attack Detection Systems with Diffusion-Generated Adversarial Traffic(7.0)
NetDiffuser generates natural adversarial examples to enhance the security of deep learning-based network intrusion detection systems.
- Learning to Inject: Automated Prompt Injection via Reinforcement Learning(6.0)
Automated prompt injection framework using reinforcement learning to enhance LLM vulnerability exploitation.
- Not-in-Perspective: Towards Shielding Google's Perspective API Against Adversarial Negation Attacks(6.0)
Develop a reasoning-based wrapper to enhance Google's Perspective API by mitigating adversarial negation attacks, improving toxicity detection accuracy.
- LLM-Based Adversarial Persuasion Attacks on Fact-Checking Systems(5.0)
Develop robust AFC systems resilient to persuasion-based adversarial attacks using LLMs.