🛡️ AI Security: Adversarial Attacks on a Classifier Decision Boundary

Watch a classifier's decision boundary in embedding space, then push a point across it with a bounded adversarial perturbation. Tune the attack budget, attack steps and adversarial-training defense, and see clean vs robust accuracy respond live.

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install