Skip to main content

Reminder that image recognition models are largely not robust against adversarial perturbation:
https://github.com/anishathalye/obfuscated-gradients

But I'm sure that #ChatControl thing is going to be completely immune to this problem. Certainly.

GitHub - anishathalye/obfuscated-gradients: Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples

Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples - GitHub - anishathalye/obfuscated-gradients: Obfuscated Gradients Give a False Sense of Security...
GitHub
#chatcontrol
This entry was edited (1 year ago)

Lo, thar be cookies on this site to keep track of your login. By clicking 'okay', you are CONSENTING to this.

⇧