Deep Learning School, 2016

Over the weekend of September 24th and 25th, some of the top (human) experts in Deep Learning got together at Stanford and presented “Deep Learning School”. The content was a mix of introductions to areas of the subject, some tutorials, and also some in-depth discussions with some of the very latest findings.

Don’t worry if you couldn’t make it. Stanford did an excellent job of providing great quality videos of the event online.

Some of my favourite parts included Ruslan showing a slide of machine and human drawn sanskrit characters. It was totally impossible to tell which was which. Richard gave a great introduction to natural language processing, cumulating in some amazing work, showing machine responses to questions about an image, and Andrej, who included a discussion about hardware options with a slide showing a hard disk drive with the caption; “Moving parts, lol”. I’m totally stealing that slide, next time I have to justify hardware costs!

Moving parts, lol

There’s far too much to go over in detail, and any summary wouldn’t do the subject justice. If you’re still looking for motivation, Andrew Ng has some advice for building a career in machine learning. I’d recommend, going through the schedule and diving right in to the videos.

Schedule

Descriptions of the talks can be found in the full schedule. Here are links to the right points in time on the videos:

Day 1

Day 2

Who’s Who

Here are the speakers and committee members.

Adam Coates

Alex Wiltschko

Andrew Ng

Andrej Karpathy

Hugo Larochelle

John Schulman

Pascal Lamblin

Pieter Abbeel

Quoc Le

Richard Socher

Ruslan Salakhutdinov

Samy Bengio

Sherry Moore

Shubho Senguta

Yoshua Bengio