News for hackers. Stuff that matters.
Re:Deep Learning: What, Why and Applications by nedsma on Friday December 02, 2016 @05:33AM (#13087543)
Re:Deep Learning: What, Why and Applications by aibottle on Friday December 02, 2016 @03:32AM (#13087053)
If you already know a bit about Machine Learning don't read this, you will gain nothing. If you don't already know about Machine Learning don't read this, you will not learn anything here. I have no idea how this can be trending. Over-simplified, generalized bs.
Re:Deep Learning: What, Why and Applications by signa11 on Friday December 02, 2016 @04:07AM (#13087189)
> If you already know a bit about Machine Learning don't read this, you will gain nothing. If you don't already know about Machine Learning don't read this, you will not learn anything here. I have no idea how this can be trending. Over-simplified, generalized bs.

indeed. much better off reading this instead: http://karpathy.github.io/neuralnets/

Re:Deep Learning: What, Why and Applications by k__ on Friday December 02, 2016 @08:26AM (#13088333)
For me the dificult is rising too fast after "Strategy 1"

First they come with

    f(x,y) = xy
Well, sounds easy.

suddendly

    df(x,y)/dx = (f(x+h,y)-f(x,y))/h
wait what?!
Re:Deep Learning: What, Why and Applications by Aeolos on Friday December 02, 2016 @08:53AM (#13088512)
This is essential knowledge you can get from a number of different places: first semester of university, any sort of 3d or image/signal processing, or just plain old wikipedia.

You could still use a deep learning framework as a black box without knowing how to differentiate a function. However, you'd have trouble following any relevant blog post / scientific paper, understanding how it works or developing any relevant new algorithm.

Re:Deep Learning: What, Why and Applications by k__ on Friday December 02, 2016 @10:35AM (#13089395)
Okay, I'm sorry.

The article is well written, my brain just turns off if I read higher math.

Re:Deep Learning: What, Why and Applications by BinRoo on Friday December 02, 2016 @03:25AM (#13087029)
>> Figure-2 shows that performance of deep learning is much better than non-deep learning algorithm.

This is a nonsensical over-generalization.

>> In addition, it is automatically do the feature extraction.

At the cost of interpretability. Let's not even mention the dreadful nights of tweaking parameters (such as dropout probability, activation function, network architecture, learning rate, optimization function, various pre-processing tricks, pre-training to warm-start, convolution parameters, maxpool parameters, and so much more).

Re:Deep Learning: What, Why and Applications by jonathanstrange on Friday December 02, 2016 @07:17AM (#13087976)
Whenever I read about deep learning, I immediately think about Microsoft's Tay AI nazi chatbot disaster.

Anyway, as a non-expert I've always thought that inferential statistical methods only work adequately when correct assumptions about the underlying distribution, hence also about the underlying analytical model and/or causal relationships are made, and I wonder how deep learning approaches deal with that issue.

Re:Deep Learning: What, Why and Applications by taneq on Friday December 02, 2016 @01:19PM (#13090916)
The real answer here, as always, is "more data".

How do humans deal with the same question? Context. Understanding. Background knowledge. The go-to rebuke of the smug liberal cognoscenti: 'educate yourself'.

To be able to interact appropriately with humans, AI needs to understand not only its own subject matter but the preconceptions and prejudices of the humans with whom it interacts.

Re:Deep Learning: What, Why and Applications by karmacondon on Friday December 02, 2016 @05:56AM (#13087637)
Many algorithms have parameters that have be tweaked, and dreadful nights are common with any machine learning concept.

Deep learning is not "better" than other forms of algorithmic prediction. There's a best tool for every job

Re:Deep Learning: What, Why and Applications by rawnlq on Friday December 02, 2016 @04:01AM (#13087171)
I know nothing about deep learning but I'm curious. Why can't tweaking parameters be automated? Why is human intuition necessary here?
Re:Deep Learning: What, Why and Applications by DavidSJ on Friday December 02, 2016 @04:17AM (#13087228)
What BinRoo called "parameters" are usually known as "hyperparameters" (parameters are the settings that are learned through gradient descent). There are methods to learn hyperparameters automatically [1], but they're extremely computationally intensive. Even with a single set of hyperparameters, it can take hours to days for training to converge. Now imagine having to run hundreds of experiments to find the best setting.

[1] https://papers.nips.cc/paper/4443-algorithms-for-hyper-param...

Re:Deep Learning: What, Why and Applications by visarga on Friday December 02, 2016 @04:52AM (#13087381)
> Why can't tweaking parameters be automated?

Not only that it can be done, but it works even by random trial. Sophisticated optimization techniques work out just 2x faster, so if you have cheap GPU, you can run 2x more trials and get your hyperparameters fine tuned.

If you want to apply optimization, you can use one of the popular libraries like hyperopt and MOE.

Re:Deep Learning: What, Why and Applications by kdelok on Friday December 02, 2016 @07:14AM (#13087961)
This really depends on what the hyperparameter landscape looks like. In general, global optimisation is a very hard problem and doing it by random sampling might take longer than the lifetime of the universe (cf Levinthal's paradox for proteins).
Re:Deep Learning: What, Why and Applications by BinRoo on Friday December 02, 2016 @04:14AM (#13087223)
Good question! Brute-force searching to find values of a parameter is a start. But, you see the issue here, right? It'll take a longass™ time. Lucky for us, learning the weight parameters of a neural network can be accomplished using back-propagation (or other methods) to speed up this search, and we do not need to brute force all possibilities. That same technique is not always applicable for hyper-parameters.
Re:Deep Learning: What, Why and Applications by signa11 on Friday December 02, 2016 @04:43AM (#13087336)
> I know nothing about deep learning but I'm curious. Why can't tweaking parameters be automated? Why is human intuition necessary here?

ok, well, imagine (for simplicity) that you are standing on a 3d-surface with mountains and valleys and such. you can only see so far. you need to walk/run etc. to get to a pot of gold which is there at the global minima of this landscape.

what do you do ? one option is choose gradient descent :) start walking somewhere, and see if the your height, from your previous position, is decreasing. if it is, then maybe you are on the right track...

Re:Deep Learning: What, Why and Applications by samet on Friday December 02, 2016 @03:16AM (#13087001)
Slashdot/HN effect:

Resource Limit Is Reached

The website is temporarily unable to service your request as it exceeded resource limit. Please try again later.

Re:Deep Learning: What, Why and Applications by Bombthecat on Friday December 02, 2016 @03:52AM (#13087133)
Well, at least he set one. Imagine this poor guy would have needed to oay over $1000 :)
Re:Deep Learning: What, Why and Applications by m00dy on Friday December 02, 2016 @04:24AM (#13087260)
Can someone build a 3rd party service that can mirror all the websites pushed here ? :)