Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: Drag and drop neural nets

getalkanet.com

Post reply on HN
Print view View original

1–9 of 9 posts

Re: Show HN: Drag and drop neural nets

#2

Post by drwl »

As someone who's taken introductory course on neural networks I have a few questions as to what your app/product does. Though I don't quite understand the need for this. Neural networks are actually fairly easy to write, and for the most part code bases exist for different architectures. What is hard, in my opinion, is picking the right architecture and parameters, cost function, etc.

Re: Show HN: Drag and drop neural nets

#3

Post by noobie »

drwl wrote: post #2

As someone who's taken introductory course on neural networks I have a few questions as to what your app/product does. Though I don't quite understand the need for this. Neural networks are actually fairly easy to write, and for the most part code bases exist for different architectures. What is hard, in my opinion, is picking the right architecture and parameters, cost function, etc.

Any recommendations for someone who knows nothing about neural networks?

Re: Show HN: Drag and drop neural nets

#5

Post by drwl »

noobie wrote: post #3
drwl wrote: post #2

As someone who's taken introductory course on neural networks I have a few questions as to what your app/product does. Though I don't quite understand the need for this. Neural networks are actually fairly easy to write, and for the most part code bases exist for different architectures. What is hard, in my opinion, is picking the right architecture and parameters, cost function, etc.

Any recommendations for someone who knows nothing about neural networks?

I think there are a bunch of great resources that have been posted to HN in the past few weeks. Stanford has a deep learning wiki that you can google. Also, my school had a summer month long graduate course that has recorded videos with lectures from the top minds in machine learning/neural networks (google "UCLA ipam neural networks") it's from 2012 and is still very very relevant.

I apologize for not being able to link directly as I'm replying from mobile.

Re: Show HN: Drag and drop neural nets

#6

Post by noobie »

drwl wrote: post #5
noobie wrote: post #3

Earlier quoted context omitted.

Any recommendations for someone who knows nothing about neural networks?

I think there are a bunch of great resources that have been posted to HN in the past few weeks. Stanford has a deep learning wiki that you can google. Also, my school had a summer month long graduate course that has recorded videos with lectures from the top minds in machine learning/neural networks (google "UCLA ipam neural networks") it's from 2012 and is still very very relevant. I apologize for not being able to…

Oh dont, thank you!

Re: Show HN: Drag and drop neural nets

#7

Post by jakebian »

drwl wrote: post #2

As someone who's taken introductory course on neural networks I have a few questions as to what your app/product does. Though I don't quite understand the need for this. Neural networks are actually fairly easy to write, and for the most part code bases exist for different architectures. What is hard, in my opinion, is picking the right architecture and parameters, cost function, etc.

You're absolutely correct - the intention of this certainly isn't to somehow simplify the process of building neural nets, which, if you're just running things locally at least, is fairly straightforward. The intention of the drag and drop interface is to give a clear and consistent way to visualize your neural net architecture. It also lends itself to a nice way to display per layer outputs.
Post reply on HN