001. Information Theory of Deep Learning - Naftali Tishby 

this is an awesome talk. Professor Tishby explains his idea so clearly and so many inspirations I have.

I am happy to hear the reference to work from the early 90's. That was really when the foundational ideas for deep learning where laid. However, it seems that lecturers are often willing to repackage and sell these fairly old ideas as something new, and this can be quite frustrating. So, thank you!

This seems absolutely revolutionary, great upload, thanks!

What an awakening talk! I wonder how the curves look like for different optimization settings, for example, with Adam chosen as optimizer.

So many experiments for so many grad students still to do!

Can you please upload the 2nd lecture?

Complex of a singular god....... run deep ! Jewish question in Dr Strangelove of abstraction.

Informathematico abire p

Blue Beep Learning Circuit Cosmic.0

Js nism kelnar.