The Mila Math Bootcamp

A physicist, just waking up from a daze, finds himself in a strange place. In front of him sits an old friend, and a copy of a single book, one with pretty flowers on the cover. He opens and starts reading; “Deep… Learning?”, Lenny questions, “Where are we George?”

“All in due time, Lenny, all in due time.”

As deep learning becomes more ubiquitous, we find ourselves, as a community, in somewhat of a predicament. By training larger and larger neural networks, ones with fancier connections and activations, it’s easy to forget that deep learning itself is not alchemy, but rather a science. While the analysis problems are more difficult, they are also equally more interesting - and as a result, more demanding.

The Math Bootcamp is an attempt to ease the jump from alchemy to science for students, researchers, and practitioners. We focus on covering mathematical concepts that are grounded in machine and deep learning, helping those in need of math abilities acquire them. Internally, we hope to use this as a resource at Mila to get new students up to speed with the tools they need to do the mathematics they want.

Thanks to the following:

• John Otander and the Github repository for the theme.
• BASSCSS
• Jekyll
• Refills
• Type Scale

Alphabetic list of contributors:

• TODO