New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

the derivation of BALD #1

Open

ShellingFord221 opened this issue Aug 15, 2019 · 0 comments

ShellingFord221 commented Aug 15, 2019

I have read the note of implementation of Deep Bayesian Active Learning with Image Data, it is very very helpful!!

I have a little question about the derivation of conditional mutual information in BALD. I'd be very appreciate if we can have a discussion about it.

My question is, in the paper, BALD is given as:

But according to the definition of conditional mutual information, the derivation should be:

which means

It seems very simple, but I just don't know why they are equivalent.

If we can discuss about it, I'd be very very thankful!!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment