Artemy Kolchinsky
Artemy Kolchinsky
The cross-entropy loss acts as a variational lower bound on the I(M;Y) term. The compression term I(M;X) is added in as a regularizer.
You should be able to have 'buildmodel.py' construct everything for you. Or, even more simply, try `python run.py` (do `python run.py --help` for instructions) If you would like to construct...
Thank you for your careful reading. Eq (18) only applies in the more general case where the distribution over the input random variable U is itself a mixture that may...