Home▸Cybersecurity▸Adversarial ML Attacks: Evasion & Poisoning

Adversarial ML Attacks: Evasion & Poisoning

Interactive 3D visualization of evasion and data-poisoning attacks on a machine-learning classifier: watch an adversarial perturbation push a feature vector across the decision boundary, or watch poisoned training points drag the boundary itself.

Cybersecurity3DModerate60 FPS

A minimal nearest-centroid classifier splits a 3D feature space into two classes with a flat decision boundary. Switch between an evasion attack, which perturbs a single input vector until it crosses the boundary, and a data-poisoning attack, which injects mislabeled training points that pull the boundary itself out of place.

⚙ Under the hood

3D nearest-centroid classifier over a feature-space cube: run an evasion attack that perturbs a single input vector across the decision boundary, or a data-poisoning attack that injects mislabeled training points and drags the boundary itself.

Three.jsAdversarial ExamplesData PoisoningMachine LearningCybersecurity

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