Re: GPU, neural networks and APU

On Thu, Aug 04, 2016 at 11:22:55PM +0200, Ivan Sergio Borgonovo wrote:
> On 08/03/2016 07:22 PM, Karen Shaeffer wrote:
> 
> > For those interested, here is an informative paper:
> >
> > http://ai.stanford.edu/~acoates/papers/CoatesHuvalWangWuNgCatanzaro_icml2013.pdf
> 
> If anyone want to have a more procedural and general idea:
> 
> http://neuralnetworksanddeeplearning.com/chap2.html
> 
> (you've to enable js to see the formulas properly)

Hi Ivan,
thanks for your comments and suggestions.

Coursera is offering a free course taught by Geoffrey Hinton on Deep Learning
that is open to all and starting very soon. Geoffrey Hinton's research and
his graduate students at University of Toronto played a leading role in the
emergence of the Deep Learning training algorithms that have disrupted the
trajectory of ML in recent years.

https://www.cs.toronto.edu/~hinton/

and Stanford is playing a major role in Deep Learning:

http://www.bayareadlschool.org

Everyone is focused on the Billion dollar problems in their deep learning
research. But deep learning will eventually disrupt all the smaller problem
domains in the coming years as well. It is a big deal.

I like Richard Socher's (2014) Phd thesis:

www.socher.org

One of the key concepts of deep learning is that it automatically learns the
optimal features, eliminating the need for costly, fragile, feature engineering
that defined the industry for quite a while. Deep learning is going to disrupt
every aspect of machine learning in the years to come. Its very early in the
maturation of this nascent technology.

Very busy. Must not become distracted.
Karen
-- 
Karen Shaeffer                 Be aware: If you see an obstacle in your path,
Neuralscape Services           that obstacle is your path.        Zen proverb