adversarial examples

Read universal and transferable as bounded research findings. Separate optimization proxies from policy failures, and local model access from hosted application authority.

Every model makes mistakes. An adversarial example is a mistake someone chose for it, and that changes how you measure and defend.

The two attacks every robustness paper starts from. The scored answer explains the gradient step, the budget as a threat model, and why the iterative attack is the standard for evaluation.

An attacker who cannot see your model can attack one they can see, and much of the damage carries over. That single property turns 'the weights are secret' into a weak defence.

Two different attacks hide in 'malicious images'. One writes instructions a model can read; the other optimises pixels to push the model's internal representation. They need different tests and different defences.