Understanding the usage
60 views
Skip to first unread message
Fabiano P Soriani
Apr 20, 2010, 8:16:31 PM4/20/10
to ruby_fann
Hello Community World!
I ran the example in the website and it worked great, even though
there seems to be no meaning in the data, I still liked the thing!
My goal with this ANN is to crack a captcha, just for fun. I'm a
computer scientist, but don't work in the field right now.
Intro done! My question is on this example code:
fann = RubyFann::Standard.new(
:num_inputs=>3,
:hidden_neurons=>[2, 8, 4, 3, 4],
:num_outputs=>1)
What is :hidden_neurons ? The ruby-doc is misterious... so I'm in
doubt if this is the initialization of neurons, or the topology..
As topology, would this mean: 2 neurons in the first layer that
connect to 8 in the second (8*2 conn.), that connect to 4 in the third
(3*8) and so on ??
Or would that be just the total of 5 neurons (the array length)?
--
You received this message because you are subscribed to the Google Groups "ruby_fann" group.
To post to this group, send email to ruby...@googlegroups.com.
To unsubscribe from this group, send email to ruby_fann+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ruby_fann?hl=en.
I ran the example in the website and it worked great, even though
there seems to be no meaning in the data, I still liked the thing!
My goal with this ANN is to crack a captcha, just for fun. I'm a
computer scientist, but don't work in the field right now.
Intro done! My question is on this example code:
fann = RubyFann::Standard.new(
:num_inputs=>3,
:hidden_neurons=>[2, 8, 4, 3, 4],
:num_outputs=>1)
What is :hidden_neurons ? The ruby-doc is misterious... so I'm in
doubt if this is the initialization of neurons, or the topology..
As topology, would this mean: 2 neurons in the first layer that
connect to 8 in the second (8*2 conn.), that connect to 4 in the third
(3*8) and so on ??
Or would that be just the total of 5 neurons (the array length)?
--
You received this message because you are subscribed to the Google Groups "ruby_fann" group.
To post to this group, send email to ruby...@googlegroups.com.
To unsubscribe from this group, send email to ruby_fann+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ruby_fann?hl=en.
sunder
Apr 20, 2010, 8:41:14 PM4/20/10
to ruby_fann
As topology, would this mean: 2 neurons in the first layer that
> connect to 8 in the second (8*2 conn.), that connect to 4 in the third
> (3*8) and so on ??
Yes.
> connect to 8 in the second (8*2 conn.), that connect to 4 in the third
> (3*8) and so on ??
> Or would that be just the total of 5 neurons (the array length)?
Best Regards,
Steven Miers
Fabiano P Soriani
Apr 21, 2010, 12:12:24 PM4/21/10
to ruby_fann
Ok, fast answer, thanks!
What would be the advantage of using multiple layers? Discovering more
patterns or something?
From what I've learned, as N.N. are complex to map though the
combination we make is really important for the success in a task.
But, it is not easy to foresee the results of the topology and
activation functions we set.
In order to break a captcha, and yet I don't see how... should I use
lots of layers? I'm positive about lot of neurons because the input is
gonna be large, even being filtrated before. Is there an activation
function recommended?
Great job on this bind, works great!
What would be the advantage of using multiple layers? Discovering more
patterns or something?
From what I've learned, as N.N. are complex to map though the
combination we make is really important for the success in a task.
But, it is not easy to foresee the results of the topology and
activation functions we set.
In order to break a captcha, and yet I don't see how... should I use
lots of layers? I'm positive about lot of neurons because the input is
gonna be large, even being filtrated before. Is there an activation
function recommended?
Great job on this bind, works great!
Reply all
Reply to author
Forward
0 new messages