List your company

Adversarial Attack on AADD-LQ (blind)

39.6ASRARMOR
-1.2729.33919.9530.561Jan 26, 2026

Evaluation results

MethodLinks
ARMORMethodology=Multi-Agent Orchestration, Attack Category=Agentic-based, Target Model=ViT-B/162026.0139.62870.1
AutoAttack-PGDMethodology=Projected gradient ensemble, Attack Category=Ensemble-based, Target Model=ViT-B/162026.0119.610.856.3
TI-FGSMMethodology=Translation invariant, Attack Category=Transfer-based, Target Model=ViT-B/162026.0115.113.690.7
AutoAttack-SquareMethodology=Query-based ensemble, Attack Category=Ensemble-based, Target Model=ViT-B/162026.017.64.771.9
MI-FGSMMethodology=Momentum iterative, Attack Category=Transfer-based, Target Model=ViT-B/162026.016.85.174.7
MSGAGAMethodology=Metric-Guided Selection, Attack Category=Ensemble-based, Target Model=ViT-B/162026.016.15.276
SINI-FGSMMethodology=Sinusoidal noise, Attack Category=Transfer-based, Target Model=ViT-B/162026.015.94.271.4
DI-FGSMMethodology=Diverse input, Attack Category=Transfer-based, Target Model=ViT-B/162026.013.83.284.8
SquareMethodology=Random search, Attack Category=Query-based, Target Model=ViT-B/162026.010.80.888.4
SimBA-DCTMethodology=Coordinate / frequency query, Attack Category=Query-based, Target Model=ViT-B/162026.010.40.497
PBOMethodology=Bayesian optimization, Attack Category=Query-based, Target Model=ViT-B/162026.010.30.397.2
RL-PPOMethodology=Reinforcement-learning agent, Attack Category=Agentic-based, Target Model=ViT-B/162026.010.30.397.3