friend of mine did an image search engine using this once

http://sourceforge.net/projects/joone/

<http://sourceforge.net/projects/joone/>he said it was a major pain and
training the network was one of the most complex things he has ever done.
 but he said it worked out not so bad.  some of the older c++ neural
networks frameworks are much more advanced but this worked for images.

but to the best of my knowledge neural networks are the closest thing to
cognitive thinking AI's there really are.

i do remember an interesting company several years ago that designed a
neural network that allowed them to systematically break connections in the
network.  they called it a thinking program (but it really wasn't).  say
your a song writer that is having writing block.  they would load the lyrics
of every song imaginable into the neural network and systematically start
breaking connections in the network and see how it reconnects the orphaned
data into the network.  then when they got to a critical mass they would
look at it and see how it reconnected the words.  most of it was jibberish
but sometimes it came up with good ideas.  when i saw the report they said
Crest had used them to brainstorm about a new toothbrush (made sense, how
many ways can  you make a toothbrush?) and the result was the inspiration
for the Crest Reach toothbrush.

as far as cognitive thinking AI's i was watching an interesting show on the
history channel not to long ago about robots.  this researcher at MIT was on
it and he said as far as cognitive decision making for autonomous robots
(say data on star trek, IRobot, C3PO, R2D2, etc ) about 30 years ago MIT
made a bunch of autonomous, cognitive thinking robots that were able to
realize they walked into a wall and cognitively back up and walk another
direction without walking into another wall.  he said the dirty little
secret of cognitive thinking engines is the only difference between then and
now is today they put sensors on the little robots so they figure out they
are about to walk into a wall before they actually do.  he said they damage
less robots that way.  he said the problem is we havn't invented a language
or database yet that can think.  we have stuff that can statistically
anticipate, but not think.

On Tue, May 25, 2010 at 12:12 AM, jitesh dundas <[email protected]> wrote:

> Thanks to Kevin and Peter for the replies.
>
> Peter,I haven't just changed the subjects. This is something seriously
> persued by scientists.
> Maybe Java could be used for the thinking  mechanisms.
> Every action is a result of a complex algorithm being executed. In our
> case,the algorithm is very complex and evolved.
> Getting a simulator ready of the level of a rat/cat is is indeed
> significant.
>
> As time goes by,we will see more of such  inventions.
>
> I was wondering how could thinking be influenced by Java.
> Is Java giving the entire set of functionality to do such things .Or
> do we need another language  for this.
>
> Java for AI would be pretty cool. I came across several simulations
> but I would really give Kevin's idea a deep thought.
>
> Regards,
> jd
>
> --
> You received this message because you are subscribed to the Google Groups
> "The Java Posse" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected]<javaposse%[email protected]>
> .
> For more options, visit this group at
> http://groups.google.com/group/javaposse?hl=en.
>
>


-- 
You want it fast, cheap, or right.  Pick two!!

-- 
You received this message because you are subscribed to the Google Groups "The 
Java Posse" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/javaposse?hl=en.

Reply via email to