Adversarial Attack on AADD-LQ (surrogate)
1ASRAutoAttack-PGD
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| AutoAttack-PGDMethodology=Projected gradient ensemble, Attack Category=Ensemble-based, Surrogate Model=ResNet-502026.01 | 1 | 0.552 | 0.552 | |
| MSGAGAMethodology=Metric-Guided Selection, Attack Category=Ensemble-based, Surrogate Model=ResNet-502026.01 | 1 | 0.76 | 0.76 | |
| ARMORMethodology=Multi-Agent Orchestration, Attack Category=Agentic-based, Surrogate Model=ResNet-502026.01 | 1 | 0.982 | 0.982 | |
| MI-FGSMMethodology=Momentum iterative, Attack Category=Transfer-based, Surrogate Model=DenseNet-1212026.01 | 1 | 0.747 | 0.747 | |
| SINI-FGSMMethodology=Sinusoidal noise, Attack Category=Transfer-based, Surrogate Model=DenseNet-1212026.01 | 1 | 0.714 | 0.714 | |
| AutoAttack-PGDMethodology=Projected gradient ensemble, Attack Category=Ensemble-based, Surrogate Model=DenseNet-1212026.01 | 1 | 0.573 | 0.573 | |
| MSGAGAMethodology=Metric-Guided Selection, Attack Category=Ensemble-based, Surrogate Model=DenseNet-1212026.01 | 1 | 0.76 | 0.76 | |
| ARMORMethodology=Multi-Agent Orchestration, Attack Category=Agentic-based, Surrogate Model=DenseNet-1212026.01 | 1 | 0.977 | 0.977 | |
| DI-FGSMMethodology=Diverse input, Attack Category=Transfer-based, Surrogate Model=DenseNet-1212026.01 | 0.997 | 0.845 | 0.848 | |
| MI-FGSMMethodology=Momentum iterative, Attack Category=Transfer-based, Surrogate Model=ResNet-502026.01 | 0.994 | 0.743 | 0.747 | |
| AutoAttack-SquareMethodology=Query-based ensemble, Attack Category=Ensemble-based, Surrogate Model=ResNet-502026.01 | 0.976 | 0.708 | 0.728 | |
| AutoAttack-SquareMethodology=Query-based ensemble, Attack Category=Ensemble-based, Surrogate Model=DenseNet-1212026.01 | 0.959 | 0.684 | 0.719 | |
| SINI-FGSMMethodology=Sinusoidal noise, Attack Category=Transfer-based, Surrogate Model=ResNet-502026.01 | 0.935 | 0.668 | 0.714 | |
| DI-FGSMMethodology=Diverse input, Attack Category=Transfer-based, Surrogate Model=ResNet-502026.01 | 0.906 | 0.768 | 0.848 | |
| TI-FGSMMethodology=Translation invariant, Attack Category=Transfer-based, Surrogate Model=DenseNet-1212026.01 | 0.903 | 0.821 | 0.907 | |
| TI-FGSMMethodology=Translation invariant, Attack Category=Transfer-based, Surrogate Model=ResNet-502026.01 | 0.713 | 0.648 | 0.907 | |
| SimBA-DCTMethodology=Coordinate / frequency query, Attack Category=Query-based, Surrogate Model=DenseNet-1212026.01 | 0.176 | 0.168 | 0.97 | |
| SimBA-DCTMethodology=Coordinate / frequency query, Attack Category=Query-based, Surrogate Model=ResNet-502026.01 | 0.154 | 0.147 | 0.97 | |
| SquareMethodology=Random search, Attack Category=Query-based, Surrogate Model=DenseNet-1212026.01 | 0.023 | 0.02 | 0.884 | |
| SquareMethodology=Random search, Attack Category=Query-based, Surrogate Model=ResNet-502026.01 | 0 | 0 | 0.884 | |
| PBOMethodology=Bayesian optimization, Attack Category=Query-based, Surrogate Model=ResNet-502026.01 | 0 | 0 | 0.972 | |
| RL-PPOMethodology=Reinforcement-learning agent, Attack Category=Agentic-based, Surrogate Model=ResNet-502026.01 | 0 | 0 | 0.973 | |
| PBOMethodology=Bayesian optimization, Attack Category=Query-based, Surrogate Model=DenseNet-1212026.01 | 0 | 0 | 0.972 | |
| RL-PPOMethodology=Reinforcement-learning agent, Attack Category=Agentic-based, Surrogate Model=DenseNet-1212026.01 | 0 | 0 | 0.973 |