Deep Learning Book Book Club

Part 1 (2017)
1

Hey guys,

We are starting a Deep Learning Book Book Club right here at Data Institute. Book is available online for free at http://www.deeplearningbook.org. This group is a time set aside to go over book’s materials and together work through problems. We’ll pick a one or a few chapters to read and discuss.

First meeting is on May 1. Chapter 1 will be presented by the author Ian Goodfellow. If you are interested RSVP here: https://www.meetup.com/Deep-Learning-Book-Club/events/239111791/

We will meet every Monday from May 1 through June 26 (with exception of Memorial Day May 29) at USF Data Institute at 101 Howard St.

Any and all suggestions on how to make this more useful and/or more awesome are greatly appreciated! Also please let me know if anyone wants to help out and become an administrator for the meetup :slight_smile:

22 Likes
2

Any chance of this being streamed for those of us who are coastally challenged?

11 Likes
3

David I think this is a great idea. I’ll investigate prior to class

3 Likes
4

Awesome! Look forward to it

5

Many thanks for setting this up @alenavkruchkova!

1 Like
6

I agree with Dave. I’d love to be able to stream it.

2 Likes
7

Is there a plan to create a separate forum tag/group (like part1, part2, etc) to discuss theoretical ai subjects?

The reason I ask is fastai is very much applied and I imagine question related to this book will be the polar opposite.

For example, @jeremy mentioned in class that one can understand cross-entropy loss in relation to KL divergence, which gives you the insight of neural net models in relation to the information theory. These types of discussion enrich one’s understanding of ideas/concepts, but it doesn’t directly cause an improvement in implementation outcome or skills.

2 Likes
8

That’s a good idea. I do plan to create some new categories, when part 2 goes online.

2 Likes
9

Is there a plan to make the book club event online? Or if not, any plan to record it and release later?

4 Likes
10

This sounds valuable; is it open to non-students? Maybe I can contribute something, given that I’m a mathematician turned data scientist (and I’ve taught linear algebra, calculus etc. at the undergraduate level). E.g. I can answer questions or give a math “lightning talk”.

“We will need people who can explain things well(especially ch 2-5)”

3 Likes
11

100% open to none students. Here is a link on meetup.com: https://www.meetup.com/Deep-Learning-Book-Club/ join and RSVP for event!

And a thousand times yes to your offer to contribute! Let’s connect and see what sounds appealing!

We will be breaking into groups to solve problems and will need people who can answer questions and help people along! What would you like to take on? Let’s chat!

12

Hi Kent!
We have a slack team already. Let me know what is your email address and I’ll add you.
I think streaming live/ posting videos online would be great since not everyone is in SF.
I will bring it up with folks at USF.
-A

13

We have started a slack team with a separate channel for each chapter. Let me know what is your email and I will add you!

Forum may be a great idea as well! What do you envision? Forum instead of slack or in addition to? Is this something you would like to take on?

1 Like
14

[quote]Any chance of this being streamed for those of us who are coastally challenged?[quote]
Great idea will bring this up with USF

15

Hey guys and gals!
Getting excited about upcoming first meeting for our book club! Hope you are excited as well.
Wanted to sum things up:

  1. There is a slack team deeplearningclubsf.slack.com - you can dm me or email me at alenavkruchkova@gmail.com and I’ll add you.

  2. People were suggesting to create a separate forum and Jeremy is open to idea :slight_smile:

  3. We will stream presentation and Q&A. I’ll work with USF to find a solution, but at the very list we’ll be able to do facebook life video! Same goes for recording event

  4. meet up is open to public, so bring a friend :slight_smile:

  5. here is a meetup page https://www.meetup.com/Deep-Learning-Book-Club/ become a
    member and RSVP to first event. We are already at capacity and have folks on waiting list. I’ll see if we could spill over into right side of our room.

  6. Who here is comfortable with linear algebra and relativity theory? We’ll need volunteers to help people in those sessions. Let me know if you’d like to get involved!

2 Likes
16

All I remember from relativity is einstein summation notation, if that helps. :slight_smile:

Something geeky but kind of awesome, that’s actually supported in numpy as einsum

1 Like
17

An incomplete version also available in tensorflow! https://www.tensorflow.org/api_docs/python/tf/einsum

1 Like
18

OK I’ve created a new category for this book club, and for discussing more theoretical stuff in general: http://forums.fast.ai/c/theory

2 Likes
19

Hi
I am sure you have resources to get you airborne but in case you don’t the Manchester Geek Nights meet up usually provides a video stream of their meet ups. Perhaps you may get some tips from them. (ThoughtWorks)

2 Likes
20

Hello all!

A few things I wanted to mention prior to today’s Book Club Meeting!

Upon popular demand we will be doing a live stream of tonights session. Here is a link: https://www.youtube.com/channel/UCF9O8Vj-FEbRDA5DcDGz-Pg/live We will go live at 6:15 to test and troubleshoot, so if you can tune in a bit early and let me know if everything looks and sounds good I would greatelly appreciate it!
Quiet a few folks are on the waitlist, so if at this point you know you can’t attend could you please change your RSVP. Thank you!
If you are not part of our slack channel yet and you’d like to be please message me directly with your email address and I will add you
Very excited to see you tonite!

Alena

3 Likes