f



Backpropagation for input=[0 0;0 1; 1000 0; 1 1],t=[0;1000;1;0]

I'm working backpropagation for xor training. the following code works for xor function but it cannot work  input=[0 0;0 1; 1000 0; 1 1],t=[0;1000;1;0]
Here is my code:
 clear all;
clc;
tic,

%2-in-and
%x=[0 0 1 1; 0 1 0 1];
%t=[0 0 0 1];

%2-in-or
%x=[0 0 1 1; 0 1 0 1];
%t=[0 1 1 1];

%f=A+BC
%x = [0 0 0 0 1 1 1 1; 0 0 1 1 0 0 1 1;0 1 0 1 0 1 0 1];  %input
%t = [0 0 0 1 1 1 1 1];           %target

%f=A(B+C)
%x = [0 0 0 0 1 1 1 1; 0 0 1 1 0 0 1 1;0 1 0 1 0 1 0 1];  %input
%t = [0 0 0 0 0 1 1 1];           %target

%3-in-XOR(sum of 3-in 1 bit adder)
% x = [0 0 0 0 1 1 1 1; 0 0 1 1 0 0 1 1;0 1 0 1 0 1 0 1];  %input
% t = [0 1 1 0 1 0 0 1];           %target

%2-in-XOR
% x=[0 0 1 1; 0 1 0 1];
% t=[0 1 1 0];

x=[0 0 1000 1; 0 1 0 1];
t=[0 1000 1 0];

% [x,meanx,stdx] = prestd(x1);
% [t,meant,stdt] = prestd(t1);
% x=x';
% t=t';


[ni N] = size(x);
[no N] = size(t);

%nh = 2;  % no of hidden unit 
errlimit=0.05;      %maximum error limit
Ntrials = 10;       % no. of trials for choosing the best network
out=zeros(no,N); % output after the backpropagation
for j=1:10
    nh=j;               %no.of hidden unit
    for trial=1:Ntrials
    %initialization
    wih =randn(nh,ni+1);
    who =randn(no,nh+1);
    neth=zeros(no,nh);
    outh=zeros(no,nh);
    delh=zeros(no,nh);
    netk=0;
    outk=0;
    delk=0;
%out=zeros(no,N);

    iteration = 0; %c=number of iteration

     while(iteration < 1000)
        iteration = iteration+1;
        for i = 1:N    
            for j = 1:nh
                neth(j) = wih(j,1:end-1)*x(:,i)+wih(j,end);
                %outh(j) = 1./(1+exp(-neth(j)));
                outh(j) = tansig(neth(j));
            end
        % hidden to output layer
            for k = 1:no            
                netk(k) = who(k,1:end-1)*outh' + who(k,end);
                outk(k) = 1./(1+exp(-netk(k))); 
                delk(k) = outk(k)*(1-outk(k))*(t(k,i)-outk(k)); 
            end
         % back propagation 
         
            for j = 1:nh
                s=0; 
                 for k = 1:no 
                     s = s + who(k,j)*delk(k); 
                end
                delh(j) = outh(j)*(1-outh(j))*s; %for each hidden unit 
            end 
           
            for k = 1:no
                for l = 1:nh
                    who(k,l) = who(k,l)+.5*delk(k)*outh(l);
                end
                who(k,l+1) = who(k,l+1)+1*delk(k)*1;
            end  
            for j = 1:nh
                for p = 1:ni
                    wih(j,p) = wih(j,p)+.5*delh(j)*x(p,i);
                end
                    wih(j,p+1) = wih(j,p+1)+1*delh(j)*1;
            end    
        end   
     end
h = tansig(wih*[x;ones(1,N)]);
%h=1./(1+exp(wih*[x;ones(1,N)]))
y = logsig(who*[h;ones(1,N)]);
error=abs(t-y);
errmax=max(max(error));
        if errmax<errlimit
            out=y;
        end  
    end
end
out=out
%out=(out)
%a= poststd(out,meant,stdt)
toc
 I try to make the input to zero mean and unit std. But its not working. Whats the problem?
Can anyone help me??


 
0
files1 (71)
10/15/2012 10:15:08 AM
comp.soft-sys.matlab 211265 articles. 25 followers. lunamoonmoon (257) is leader. Post Follow

1 Replies
2004 Views

Similar Articles

[PageSpeed] 35

"shamima" wrote in message <k5gnnc$qp7$1@newscl01ah.mathworks.com>...
> I'm working backpropagation for xor training. the following code works for xor function but it cannot work  input=[0 0;0 1; 1000 0; 1 1],t=[0;1000;1;0]
> Here is my code:

C1: After writing a long code with loops use
      SELECT ALL
      SMART INDENT

>  clear all;
> clc;
> tic,
> 
> %2-in-and
> %x=[0 0 1 1; 0 1 0 1];
> %t=[0 0 0 1];
> 
> %2-in-or
> %x=[0 0 1 1; 0 1 0 1];
> %t=[0 1 1 1];
> 
> %f=A+BC
> %x = [0 0 0 0 1 1 1 1; 0 0 1 1 0 0 1 1;0 1 0 1 0 1 0 1];  %input
> %t = [0 0 0 1 1 1 1 1];           %target
> 
> %f=A(B+C)
> %x = [0 0 0 0 1 1 1 1; 0 0 1 1 0 0 1 1;0 1 0 1 0 1 0 1];  %input
> %t = [0 0 0 0 0 1 1 1];           %target
> 
> %3-in-XOR(sum of 3-in 1 bit adder)
> % x = [0 0 0 0 1 1 1 1; 0 0 1 1 0 0 1 1;0 1 0 1 0 1 0 1];  %input
> % t = [0 1 1 0 1 0 0 1];           %target
> 
> %2-in-XOR
> % x=[0 0 1 1; 0 1 0 1];
> % t=[0 1 1 0];
> 
> x=[0 0 1000 1; 0 1 0 1];
> t=[0 1000 1 0];
> 
> % [x,meanx,stdx] = prestd(x1);
> % [t,meant,stdt] = prestd(t1);
> % x=x';
> % t=t';
> 
> [ni N] = size(x);
> [no N] = size(t);
> 
> %nh = 2;  % no of hidden unit 
> errlimit=0.05;      %maximum error limit
> Ntrials = 10;       % no. of trials for choosing the best network
> out=zeros(no,N); % output after the backpropagation
> for j=1:10
>     nh=j;               %no.of hidden unit
>     for trial=1:Ntrials
>     %initialization
>     wih =randn(nh,ni+1);
>     who =randn(no,nh+1);

C2. Initial weights should be smaller to prevent sigmoid saturation

>     neth=zeros(no,nh);
>     outh=zeros(no,nh);
>     delh=zeros(no,nh);
>     netk=0;
>     outk=0;
>     delk=0;
> %out=zeros(no,N);
> 
>     iteration = 0; %c=number of iteration
> 
>      while(iteration < 1000)
>         iteration = iteration+1;

C3: Remove semicolon to help keep track. Also add

       elapsed time = toc

>         for i = 1:N    
>             for j = 1:nh

C4: UHOH! You are already in a "j" loop

>                 neth(j) = wih(j,1:end-1)*x(:,i)+wih(j,end);
>                 %outh(j) = 1./(1+exp(-neth(j)));
>                 outh(j) = tansig(neth(j));
>             end
>         % hidden to output layer
>             for k = 1:no            
>                 netk(k) = who(k,1:end-1)*outh' + who(k,end);
>                 outk(k) = 1./(1+exp(-netk(k))); 
>                 delk(k) = outk(k)*(1-outk(k))*(t(k,i)-outk(k)); 
>             end
>          % back propagation 
>          
>             for j = 1:nh

See C4

>                 s=0; 
>                  for k = 1:no 
>                      s = s + who(k,j)*delk(k); 
>                 end
>                 delh(j) = outh(j)*(1-outh(j))*s; %for each hidden unit 
>             end 
>            
>             for k = 1:no
>                 for l = 1:nh
>                     who(k,l) = who(k,l)+.5*delk(k)*outh(l);
>                 end
>                 who(k,l+1) = who(k,l+1)+1*delk(k)*1;
>             end  
>             for j = 1:nh

See C4

>                 for p = 1:ni
>                     wih(j,p) = wih(j,p)+.5*delh(j)*x(p,i);
>                 end
>                     wih(j,p+1) = wih(j,p+1)+1*delh(j)*1;
>             end    
>         end   
>      end

C5: Learning Rates are too high. Will saturate the sigmoids.

> h = tansig(wih*[x;ones(1,N)]);
> %h=1./(1+exp(wih*[x;ones(1,N)]))
> y = logsig(who*[h;ones(1,N)]);
> error=abs(t-y);
> errmax=max(max(error));
>         if errmax<errlimit
>             out=y;
>         end  
>     end
> end
> out=out

C6: You don't keep the best of the 10 trials 

> %out=(out)
> %a= poststd(out,meant,stdt)
> toc
>  I try to make the input to zero mean and unit std. But its not working. Whats the problem?
> Can anyone help me??

I only ran your 2-input XOR and made the 6 above comments. 

Search the archives for old posts of mine re the probability of convergence for the XOR design. It is surprisingly low. Therefore, I suggest you run at least 10 weight initialization trials for each value of nhi. 

As for your initial problem with 1000. The solution is obvious:

When you are dealing with sigmoids,

1. Normalize your inputs (help/doc zscore)
2. Use nonsaturating initial weights
Search comp.ai.neural-nets using the keyword nonsaturating

Hope this helps.

Greg

 
0
heath (3983)
10/15/2012 4:57:08 PM
Reply:

Similar Artilces:

[0, 0, 0, 1, 1, 1, 0] ... remove all 0 values
Hi python - hackers, just one question. How can i remove all 0 values in a list? Sure - i can loop over it, but that s not a neat style. list.remove() will only remove the first occurence. Doing that while no exception is raised is also uncool, right? Some suggestions? Best, Dan Daniel Austria a �crit : > Hi python - hackers, > > just one question. How can i remove all 0 values in a list? Sure - i > can loop over it, but that s not a neat style. list.remove() will > only remove the first occurence. Doing that while no exception is > raised is also uncool, right? > ...

0.1 + 0.1 + 0.1
This is gcc 4.8.1 with no compiler option used: #include <stdio.h> #include <stdlib.h> int main( void ) { double x = rand() < 0 ? 0 : 0.1; double y = rand() < 0 ? 0 : 0.3; printf( "%g\n", x + x + x - y ); printf( "%g\n", 0.1 + 0.1 + 0.1 - 0.3 ); } 2.77556e-017 5.55112e-017 The second number written becomes �0� as soon as I add the compiler option �-std=c11�, why? On 10/23/14 9:41 PM, Stefan Ram wrote: > This is gcc 4.8.1 with no compiler option used: > > #include <stdio.h> > #include <stdlib.h...

Why (eql #c(1 0) 1) but (not (eql #c(1.0 0.0) 1.0))?
Is there a deep reason--mathematical, implementational or otherwise--that numbers of type (complex float) can have a zero imaginary part while (complex rationals) can not? Or is it just one of those things? -Peter -- Peter Seibel peter@javamonkey.com Lisp is the red pill. -- John Fraser, comp.lang.lisp Peter Seibel <peter@javamonkey.com> writes: > Is there a deep reason--mathematical, implementational or > otherwise--that numbers of type (complex float) can have a zero > imaginary part while (complex rationals) can not? Or is i...

*.so , *.la , *.a , *.so.0.0.0 , *.so.1.0.0 ...
I noticed so many extensions in the /usr/lib directory of my Linux system. .so stands for shared object but also there are other extensions can anyone please explain or atleast give pointers to where I can find information about these extensions in details. regards, Onkar onkar <onkar.n.m@gmail.com> wrote: > I noticed so many extensions in the /usr/lib directory of my Linux > system. .so stands for shared object but also there are other > extensions can anyone please explain or atleast give pointers to where > I can find information about these extensions in details. Sorry, but you're in the wrong group with this. This group is about the C programming language, but what you want to know is something concerning the naming conventions for some types of files under Linux which is completely unrelated to C. You probably will find much more help in a newsgroup dedicated to programming under Linux, e.g. comp.os.linux.development.apps, or a group for programming under Unix, comp.unix.programmer (since your question is actually not Linux specific). Regards, Jens -- \ Jens Thoms Toerring ___ jt@toerring.de \__________________________ http://toerring.de onkar wrote: > I noticed so many extensions in the /usr/lib directory of my Linux > system. .so stands for shared object but also there are other > extensions can anyone please explain or atleast give pointers to where > I can find informati...

ANN: fathom 0.4.0, fathom-tools 0.1.0, qfathom 0.1.0
Hello everyone! I would like to announce new version of fathom as well as two more packages built on top fathom. All this code is still very young and experimental, so it is not supposed to be production ready. However, I would greatly appreciate any advice or ideas on what could be useful and what would you expect of those packages. fathom 0.4.0 Fathom is python 3 library for inspecting database schema. It produces objective model of all database objects and thus allows easy database inspection. Changes: * Indices are top level objects and provide information about index uniquen= ess * Proc...

if time = 3 seconds, how do i set the values of the time steps 0,0.1,0.2,0.3,0.5,1.0,1.5,2.0,3.0,5.0 to -9999 from time =3 seconds
Hi &nbsp; I have 18 inputs (arrays) which contain certain statistics and if time&nbsp;&nbsp;= 3 seconds, set the values of time step 5.0 to -9999.0. &nbsp; I was wondering if there is an efficient way to do this without using a lot of case statements? &nbsp; Thank you. Hello AdrianT, it's not really clear to me what you want to do... I suspect: you have an array containing values and you want to output them with a fixed timing of 3 seconds: you should use a for loop autoindexing trough your array and sending the value to your indicator/output/whatever. Put a "wait for next multiple" into your for loop. Or do you want to have a minimum time delay of 3 seconds? Again use a for loop, get the time value and use a InRange&amp;Coerce with minimum set to 3 seconds... Or you should explain more precisely what you need... Hi Thanks for that. Let me explain. &nbsp; The time may vary. Let me give an example. if time&nbsp; = 2.1, I want to set the time step's &nbsp;3.0 and 5.0's values to -9999.0. if the time =0.1 seconds then I want to set 0.2,0.3, 0.5 etc values to -9999.0 etc .................................................................... &nbsp; Can you please help? Thank you. Hello TUDS, do you use different accounts/nicknames ? Well, that's easy, at least for a sorted array: Search for your time-value in the array (Threshold 1D array). Fill all elements from that index till the end with your "error...

#if 0?0?0:0:0
The following 4-lines source fragment test.c (fourth line empty) #if 0?0?0:0:0 #endif int main(void){return 0?0?0:0:0;} cause: test.c(1) : fatal error C1017: invalid integer constant expression when compiled by cl.exe aka "Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 15.00.30729.01 for 80x86". On the other hand this compile and runs fine. #if 0?(0?0:0):0 #endif int main(void){return 0?0?0:0:0;} Did I hit a but it this compiler's preprocessor? Francois Grieu On 3/30/2010 12:47 PM, Francois Grieu wrote: > The following 4-lines source fr...

Why does python break IEEE 754 for 1.0/0.0 and 0.0/0.0?
I've read over and over that Python leaves floating point issues up to the underlying platform. This seems to be largely true, but not always. My underlying platform (IA32 Linux) correctly handles 1.0/0.0 and 0.0/0.0 according to the IEEE 754 standard, but Python goes out of its way to do the wrong thing. 1/0 is defined by the standard as +Inf and 0/0 is NaN. That's what my platform does for programs written in C. Python apparently checks for division by zero and throws and exception rather than returning the correct value calculated by the underlying platform. Is there any way...

changing compatible parameter from 8.1.0.0.0 to 8.1.7.0.0
Hello, I need to increase the compatibility level of a database instance running on a solaris box from 8.1.0.0.0 to 8.1.7.0.0. but I never did it before. What are the main issues that this change could bring about? Are there any specific caveats? Thanks in advance, Flavio flavio@tin.it (FC) wrote in news:2bd78ddf.0410150651.bbceccb@posting.google.com: > Hello, > I need to increase the compatibility level of a database instance > running on a solaris box from 8.1.0.0.0 to 8.1.7.0.0. but I never did > it before. > > What are the main issues that this change could bring ab...

[ANN] pikka-bird collector 0.2.0, server 0.1.0, puppet 0.1.0 released
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Dear People, I'm pleased to announce 3 related releases of Pikka Bird, a new ops monitoring tool aiming for ease of use and configuration. pikka-bird-collector 0.2.0: Pikka Bird ops monitoring tool Collector component. (Python) https://github.com/tiredpixel/pikka-bird-collector-py pikka-bird-server 0.1.0: Pikka Bird ops monitoring tool Server component. (Python) https://github.com/tiredpixel/pikka-bird-server-py pikka-bird-puppet 0.1.0: Pikka Bird ops monitoring tool Puppet module. (Puppet) https://github.com/tiredpixel...

[ANN] Graffiti 1.0, SynCache 1.0, Whitewash 1.0
Three components of the Samizdat open publishing system are finally released as stand-alone modules: Graffiti relational RDF store, SynCache thread-safe object cache, Whitewash whitelist-based HTML filter. Graffiti 1.0: Graffiti is an RDF store based on dynamic translation of RDF queries into SQL. Graffiti allows to map any relational database schema into RDF semantics and vice versa, to store any RDF data in a relational database. This RDF store was used as the primary means of data access in Samizdat from the start. Release Notes: http://samizdat.nongnu.org/release-notes/graffit...

Availability of Firefox 1.5.0.9, Firefox 2.0.0.1 and Seamonkey 1.0.7
Builds of Firefox and Seamonkey for AIX 4.3.3 are available now. Firefox: "http://www.linkitup.de/ForAIX/Firefox" Seamonkey: "http://www.linkitup.de/ForAIX/Mozilla" the POWER optimized build of Firefox 1.5.0.9 will follow tomorrow. The machine is a little slow for building all three on one day ;-) Feedback appreciated (as always...) -- Uli Link Hi! > Firefox: "http://www.linkitup.de/ForAIX/Firefox" That's going to require some diddling around, I see the files are split between multiple sites. > Seamonkey: "http://www.linkitup.de/For...

0,0 not 0,0
hi all, i have 2 drawings - ground floor and first floor. when i do an ID on a point on the grid i get the same readings for both drawings however, when i xref one into the other or both into a new drawing they come in in different locations. i've checked they're both using world UCS and the same units. any ideas? i thought with autocad 0,0 was always 0,0? cheers rob "Coro, Rob" <RE-Coro@bdp.co.uk> schrieb: >hi all, > >i have 2 drawings - ground floor and first floor. when i do an ID on a >point on the grid i get the same readings for both drawings however, >when i xref one into the other or both into a new drawing they come in >in different locations. i've checked they're both using world UCS and >the same units. any ideas? i thought with autocad 0,0 was always 0,0? UCS? INSBASE? Tom Berger -- ArchTools: Architektur-Werkzeuge f�r AutoCAD (TM) ArchDIM - architekturgerechte Bema�ung und H�henkoten ArchAREA - Fl�chenermittlung und Raumbuch nach DIN 277 Info und Demo unter http://www.archtools.de insbase, that's the bugger. thanks tom rob -----Original Message----- From: Tom Berger [mailto:berger@archtools.de] Posted At: 13 July 2004 12:24 Posted To: autocad Conversation: 0,0 not 0,0 Subject: Re: 0,0 not 0,0 "Coro, Rob" <RE-Coro@bdp.co.uk> schrieb: >hi all, > >i have 2 drawings - ground floor and first floor. wh...

(= 0/0 0/0)
Hi, shouldn't > (= 0/0 0/0) => NIL or > (= 0/0 0/0) => T instead of giving out division by zero? 0/0 is different from 1/0, which is the non-existent number n such that 1 x n = 0. The actual result of that expression depends of what you think 0/0 is. I personally feel that 0/0 is any number n such that 0 x n = 0, that is all numbers. hal9@cyberspace.org (Hal Niner) writes: > Hi, > > shouldn't > > > (= 0/0 0/0) => NIL > > or > > > (= 0/0 0/0) => T > > instead of giving out division by zero? 0/0 is different fro...

[ANN] Net::SSH 1.0.1, Net::SFTP 1.0.0
Net::SSH 1.0.1 is a maintenance release that fixes a bug that caused data to be sent twice in some circumstances. It also adds a :timeout option to Net::SSH.start, and allows the Putty Agent to be used under Windows (reportedly). Project page: http://rubyforge.org/projects/net-ssh Users Manual: http://net-ssh.rubyforge.org API Docs: http://net-ssh.rubyforge.org/api Net::SFTP 1.0.0 is the more-or-less stable, more-or-less final release of Net::SFTP. (More-or-less, meaning, there may be subsequent releases as bugs are found, patches submitted, and time allocated to work on...

Installing Plucker Desktop 1.6.0.1 over 1.6.0.0
Is it safe to install Plucker Desktop 1.6.0.1 over 1.6.0.0? I'm using the Windows version. Chuck -- <http://www.drchuck.net/> Home of the Dr. Chuck Network "Charles A. Kunz" <drchuck@i-2000.com> wrote in message news:6546qv07ajb7nbr9vo7tl53u71cmc5dod3@4ax.com... > Is it safe to install Plucker Desktop 1.6.0.1 over 1.6.0.0? I'm using > the Windows version. > > Chuck > -- Hi Chuck, I guess it safe; if not, email the author of the software of any problems. Byron --- Byron's Emporium - Cool eBooks for PDA's ENJOYOU...

Why timestamp can not show time between 0:0:0 JAN 1, 1970 to 8:0:0 Jan 1, 1970?
Why timestamp can not show time between 0:0:0 JAN 1, 1970 to 8:0:0 Jan 1, 1970? Hi Qizhen, what do you mean? <img src="http://forums.ni.com/attachments/ni/170/318945/2/Unbenannt.PNG"> MikeMessage Edited by MikeS81 on 04-25-2008 09:46 AM Unbenannt1.PNG: http://forums.ni.com/attachments/ni/170/318945/2/Unbenannt1.PNG <img src="http://forums.ni.com/attachments/ni/170/318948/1/aa.png"> ???Qizhen?04-25-2008 04:53 PM????? aa.png: http://forums.ni.com/attachments/ni/170/318948/1/aa.png Create a new timestamp constant on block diagram.&nbsp; Then try to in...

[ANN] Ruby/Watchcat 1.0.1 and Ruby/Watchcat-pure 1.0.0
Hello I'm pleased to announce a new release of Ruby/Watchcat. This release includes a fix to a possible descriptor leak in the C extension and also a pure-Ruby version of the library, which makes it usable in other Ruby implementations. RDoc documentation is now available too. Ruby/Watchcat is an extension for Ruby for the development of watchcatd-aware applications. Watchcatd is a watchdog-like daemon in the sense that it takes actions in situations where a machine is under heavy load and/or unresponsive. However, watchcatd isn't as drastic as the usual watchdog systems, which rebo...

[ANN] Unicorn 2.0.0, 1.1.5, 1.0.2 released!
Unicorn is an HTTP server for Rack applications designed to only serve fast clients on low-latency, high-bandwidth connections and take advantage of features in Unix/Unix-like kernels. Slow clients should only be served by placing a reverse proxy capable of fully buffering both the the request and response in between Unicorn and slow clients. * http://unicorn.bogomips.org/ * mongrel-unicorn@rubyforge.org * git://git.bogomips.org/unicorn.git tldr change summaries: 2.0.0 - cleanups for Rainbows!, but should be ready for general use barring portability issues 1.1.5 - bug f...

IP 0.0.0.0/0
Hi Any useful link that explains this IP address range in details, please? Thanks in advance! The Dude In article <Jp6Hg.458762$IK3.24918@pd7tw1no>, The Dude <The Dude@thedu.de> wrote: >Any useful link that explains this IP address range in details, please? Urrr -- 0.0.0.0/0 is the *entire* IPv4 address range, and 0.0.0.0/32 is just the single IPv4 address 0.0.0.0. For any given network, the lowest address in the network is reserved. Historically, the lowest address was one of the two allowed choices for the broadcast address; later, the broadcast a...

Schools subnet = 10.0.0.1 Classes subnet 10.0.0.1 can't connect to internet
Schools subnet = 10.0.0.1 Class subnet 10.0.0.1 -- can't connect to internet Don't know if this is exactly a firewall issue. My school allows us to do a little lan gaming in one computer lab. I like to set up my dlink to pass out 10.0.0.x addresses because it's easy to type in. Games and map servers are under 10.0.0.100. My gateway is 10.0.0.1. I connect that to the schools system (lets call it a linksys) which is also a 10.0.0.x subnet. I use the DLINK because there's a bunch of windows shares I don't want the gamers getting to and it seems to isolate them. But when I...

ClearInbox 1.0.0.0
Is your Outlook inbox out of control ? With ClearInbox you can: 1) Move email group faster to suggested folder; 2) will never miss important email from VIPs; 3) search attachment & email faster; 4) auto-classify & manage new emails; 5) replay email as chat history for easy track More Info: http://www.clearinboxsoft.com Download URL: http://www.clearinboxsoft.com/downloads/ClearInbox.exe Screenshot URL: http://www.clearinboxsoft.com/images/ScreenShot.jpg Order Page: http://secure.plimus.com/jsp/dev_store1.jsp?developerId=824322 ...

SuperOrca 1.0.0.0
SuperOrca from "Pantaray Research Ltd." is a direct replacement to the "Orca" utility from Microsoft. SuperOrca may be used to examine and modify an MSI database. SuperOrca - can be freely downloaded from the internet, no need to download and install the entire Microsoft SDK. SuperOrca - will allow you to scan and manipulate all MSI database table in an intuitive and user friendly manner. SuperOrca - will not lock the MSI file you are working on, so you can test it with other tools or even install it while the file is still opened in SuperOrca. SuperOrca - features ...

Why would Webrick use 0.0.0.0:3000 instead of 127.0.0.1:3000 ?
------=_Part_3258_21100517.1134839838299 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Someone working through my Rolling with Ruby on Rails tutorial ran into this. Does anyone know what could cause this? I pasted the text of his post below. Curt I am using Instant Rails which comes with Rails 1.0, WEBrick 1.3.1 and Ruby 1.8.2 and I ran the WEBrick server and it starts on http://0.0.0.0:3000 and not http://127.0.0.1:3000. However, when I tried to connect to http://0.0.0.0:3000 using my browser, a dialog pops up telling m...

Web resources about - Backpropagation for input=[0 0;0 1; 1000 0; 1 1],t=[0;1000;1;0] - comp.soft-sys.matlab

Backpropagation - Wikipedia, the free encyclopedia
Backpropagation , an abbreviation for "backward propagation of errors", is a common method of training artificial neural networks . From a desired ...

NeuroGen - Genetic Algorithms for Neural Network Training on Transputers
The use of both, genetic algorithms and artificial neural networks, was originally motivated by the astonishing success of these conceptsin their ...

ESANN - Author index
Electronic preceedings author index ESANN - European Symposium on Artificial Neural Networks [ Electronic proceedings home page ] A B C D E F ...

Peter Jansen - PhD Candidate - Cognitive Science Laboratory
stuff about me my name is peter, and i'm a bit strange, often blue-haired graduate student in the cognitive science laboratory at mcmaster university.my ...

Teaching Portfolio - Giorgos Mountrakis
Welcome to the webpage of my classes. My courses are designed for a variety of student backgrounds, so feel free to contact me for further information. ...

Applying Genetic Algorithms to Recurrent Neural Networks for Learning Network Parameters and Architecture ...
Recurrent neural networks such as the fully interconnected Hopfieldnetwork have the potential of being applied to many input output mappingproblems, ...

You Don’t Need Coursera to Get Started with Machine Learning
Since I currently work at a Machine Learning company, it may surprise some to find out that I am currently enrolled in Andrew Ng’s Machine Learning ...

Complex Systems
Complex Systems is a journal devoted to the science, mathematics, and engineering of systems with simple components but complex overall behavior. ...

Big Data Predictions for 2013 - Architects Zone
If you just invested a lot of money in a Big Data solution from any of the traditional BI vendors (Teradata, IBM, Oracle, SAS, EMC, HP, etc.) ...

j schmidhuber - Google Scholar
... , J Schmidhuber - Neural computation, 1997 - ieeexplore.ieee.org Learning to store information over extended time intervals by recurrent backpropagation ...

Resources last updated: 3/10/2016 9:06:48 PM