Hi, I am still quite new to data science, this is the first time I see someone using information theory to measure whether a neural network has suitable amount of parameters.

Do you mind sharing more? Like the reference, or some examples. I would love to know more about this. Thank you!