Neural Networks and Deep Learning(neuralnetworksanddeeplearning.com)
neuralnetworksanddeeplearning.com
Neural Networks and Deep Learning
http://neuralnetworksanddeeplearning.com/index.html
16 comments
I find this stuff fascinating. I would love to study this formally but I only have a finance degree. What do people study to pursue this at a graduate level? Math/Comp Science?
From Hinton's course on coursera:
Recommended Background
Programming proficiency in Matlab, Octave or Python. Enough knowledge of calculus to be able to differentiate simple functions. Enough knowledge of linear algebra to understand simple equations involving vectors and matrices. Enough knowledge of probability theory to understand what a probability density is.
https://www.coursera.org/course/neuralnets
Recommended Background
Programming proficiency in Matlab, Octave or Python. Enough knowledge of calculus to be able to differentiate simple functions. Enough knowledge of linear algebra to understand simple equations involving vectors and matrices. Enough knowledge of probability theory to understand what a probability density is.
https://www.coursera.org/course/neuralnets
http://www.metacademy.org/ is a great tool that shows you all the prerequisites of a concept. As well as those concepts prerequisites and so on, and where you can go to learn them.
Normally Computer Science, certainly not Math. My university had a separate graduate program for Artificial Intelligence.
Maybe also Neuroscience or Psychology, but they usually use other kinds of neural network models, which are more biologically plausible.
Maybe also Neuroscience or Psychology, but they usually use other kinds of neural network models, which are more biologically plausible.
Still long way to deep learning stuff. I have seen this book in almost same state in last posts. But the content so far are great, especially for understanding neural network and hands on experience.
Again?
I love this, and bought the (unfinished) book since it was 1st announced, but it's the 3rd or 4th time this is posted here.
The duplicity algorithm clearly needs improvement. Maybe this book could help.
I love this, and bought the (unfinished) book since it was 1st announced, but it's the 3rd or 4th time this is posted here.
The duplicity algorithm clearly needs improvement. Maybe this book could help.
I'm the author. Thanks for the support! And I'm glad you're enjoying the book.
The HN search results suggest that this is the second time the book has made it to the HN front page.
However, you may have picked up a different impression because the second chapter -- a largely self-contained introduction to backpropagation -- also made it once onto the front page. Given that I'm rolling out chapters as I finish them, this seems to me like multiple posts from the same blog appearing on the front page. Certainly, the individual chapters are (I hope) far more substantive than most blog posts.
(For the record, I had no idea this would be posted.)
The HN search results suggest that this is the second time the book has made it to the HN front page.
However, you may have picked up a different impression because the second chapter -- a largely self-contained introduction to backpropagation -- also made it once onto the front page. Given that I'm rolling out chapters as I finish them, this seems to me like multiple posts from the same blog appearing on the front page. Certainly, the individual chapters are (I hope) far more substantive than most blog posts.
(For the record, I had no idea this would be posted.)
I've found 8 so far [1][2][3]. Some are exact duplicates, some point to / while others to /index.html, /about.html, etc.
Note: most submissions were done by real and reputable users, so I'm not implying any foul play here. Quite the opposite; I backed the original project on Indie Gogo, and am pleased with the progress so far. Would strongly recommend anyone interested in the topic to donate to Michael (BTC address below).
My gripe is with HN duplicity detector, that let people submit equal or very similar links, over and over again.
[1] Book: (some are exact duplicates)
https://news.ycombinator.com/item?id=7920183
https://news.ycombinator.com/item?id=6796703
https://news.ycombinator.com/item?id=7555191
https://news.ycombinator.com/item?id=7143192
https://news.ycombinator.com/item?id=6801036
[2] Indie Gogo:
https://news.ycombinator.com/item?id=6795285
https://news.ycombinator.com/item?id=6794584
[3] Original Blog post:
https://news.ycombinator.com/item?id=6796917
Note: most submissions were done by real and reputable users, so I'm not implying any foul play here. Quite the opposite; I backed the original project on Indie Gogo, and am pleased with the progress so far. Would strongly recommend anyone interested in the topic to donate to Michael (BTC address below).
My gripe is with HN duplicity detector, that let people submit equal or very similar links, over and over again.
[1] Book: (some are exact duplicates)
https://news.ycombinator.com/item?id=7920183
https://news.ycombinator.com/item?id=6796703
https://news.ycombinator.com/item?id=7555191
https://news.ycombinator.com/item?id=7143192
https://news.ycombinator.com/item?id=6801036
[2] Indie Gogo:
https://news.ycombinator.com/item?id=6795285
https://news.ycombinator.com/item?id=6794584
[3] Original Blog post:
https://news.ycombinator.com/item?id=6796917
I believe the dupe detector is left purposely "leaky."
I don't understand what's so wrong with dupes, especially the amount of effort you've put into this one. I don't see every headline on HN, neither do you, nor does anyone (I hope). What is your methodology for trawling through the archives to find interesting things you might have missed? Some people like having a chance for them to pop up again, which usually indicates that they're high quality. And inevitably someone will show up in the comments and link all the old discussions and then people can review those and also start a new discussion! And if you've already seen it, you can skip it! What's the issue?
Obviously, a massive flood of dupes is not desirable, so there is balance to be found, but I think they hit it pretty well. They also have to factor in the wide range of reading habits.
I don't understand what's so wrong with dupes, especially the amount of effort you've put into this one. I don't see every headline on HN, neither do you, nor does anyone (I hope). What is your methodology for trawling through the archives to find interesting things you might have missed? Some people like having a chance for them to pop up again, which usually indicates that they're high quality. And inevitably someone will show up in the comments and link all the old discussions and then people can review those and also start a new discussion! And if you've already seen it, you can skip it! What's the issue?
Obviously, a massive flood of dupes is not desirable, so there is balance to be found, but I think they hit it pretty well. They also have to factor in the wide range of reading habits.
The policy, which has been stable for a long time and which we intend to put in the FAQ, is that reposts are ok if an item hasn't had significant attention in the last year.
In this case, both available chapters of the book have had significant attention [1,2], so I think a post of its home page has to count as a dupe. If and when Chapter 3 appears, however, that will make a nice non-dupey post.
(Good luck with the book, Michael!)
1. https://news.ycombinator.com/item?id=6794308
2. https://news.ycombinator.com/item?id=7588158
In this case, both available chapters of the book have had significant attention [1,2], so I think a post of its home page has to count as a dupe. If and when Chapter 3 appears, however, that will make a nice non-dupey post.
(Good luck with the book, Michael!)
1. https://news.ycombinator.com/item?id=6794308
2. https://news.ycombinator.com/item?id=7588158
Thanks for the clarification. Yes, it would be great to have that policy in the FAQ.
The parent mentioned that he bought the book, but I can't seem to find anywhere on the site to contribute. I've followed this book since it was first posted on Hacker News, and I would love to be able to give back. Is there a place to donate or "buy" the book?
I ran an IndieGoGo, which is long since over. Other than the IndieGoGo I haven't given much thought to setting things up so people can contribute, that's very kind of you to offer. I'll investigate adding PayPal or some alternative when I have a little time. In the meantime, you could send me Bitcoin at: 17ukkKt1bNLAqdJ1QQv8v9Askr6vy3MzTZ It'd be much appreciated!
I've submitted to http://hackershelf.com/book/525/neural-networks-and-deep-lea...
Great work Michael. Sorry about my errant email to you.
Great work Michael. Sorry about my errant email to you.
hey - im pretty exciting about this also - when is it slated to be finished ?
Chapter 3 (of six) will be released mid-July. Apart from that, I haven't set a hard finish date. Some authors are helped by such deadlines, but my experience writing my previous two books is that (for me) such deadlines hinder more than they help.