Question about the paper formula

[wetliu]

Hello. Thank you for the great work.

Should the equation defining the corruption robustness be Ec∼C[P(x,y)∼D(f(c(x)) = y)] instead of Ec∼C[P(x,y)∼D(f(c(x) = y))]?

If it should not, would you mind explaining what it means by c(x) = y?

Thank you so much!

[hendrycks]

f(c(x)) = y f is the classifier; c is the corruption function; x is the original image before corruption; if f(c(x)) = y, then that means the classifier f correctly classified the corrupted image c(x)

…

On Sun, May 29, 2022 at 11:28 PM wetliu ***@***.***> wrote: Hello. Thank you for the great work. Should the equation defining the corruption robustness be Ec∼C[P(x,y)∼D(f(c(x)) = y)] instead of Ec∼C[P(x,y)∼D(f(c(x) = y))]? If it is not, would you mind explaining what it means by c(x) = y? Thank you so much! — Reply to this email directly, view it on GitHub <#56>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACZBITVCNIIRM2LV35YWTBLVMRNW5ANCNFSM5XJOSPJA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[wetliu]

Thank you so much for your clarification!