Interaction with Robot AI Minds

Ian Field i.field at virgin.net
Sat Mar 23 08:33:50 EST 2002


"Arthur T. Murray" <uj797 at victoria.tc.ca> wrote in message
news:3c9a6315 at news.victoria.tc.ca...
> mats_trash at hotmail.com (mat) wrote on 21 Mar 2002:
> > mentifex at scn.org (Arthur T. Murray) wrote in message
> > news:<662dfbc2.0203201050.4c1a262b at posting.google.com>...
> >> The HCI module of the Robot AI Mind is documented on-line at
> >> http://www.scn.org/~mentifex/hci.html -- free AI source code.
> >>
> >> http://www.scn.org/~mentifex/jsaimind.html is in JavaScript.
> >>
> >> http://www.scn.org/~mentifex/mind4th.html is in Win32Forth.
> >>
> >> A.T. Murray
> >
> > Can you explain what I am missing about your programme as to
> > me it seems totally unlike interacting with another 'mind'.
> ATM/Mentifex:
> Right now the program is an extremely simple mind with a
> single-digit IQ probably in the range of 2 or 3.
>
> Because it takes about thirty software modules just to get
> the AI Mind up and running, the coding work slows down
> as the software gets progressively more complicated and
> needs constant rewriting of the on-line documentation, e.g.,
> http://www.scn.org/~mentifex/audrecog.html -- today's work.
>
> http://www.scn.org/~mentifex/userman.html -- the User Manual --
> describes how the AI starts out as nothing more than a
> knowledge base (KB), queryable only with three-word
> subject-verb-object (SVO) sentences like this exchange:
> Human:  cats eat fish
> Robot:  CATS EAT FISH
> Human:  boys play war
> Robot:  BOYS PLAY WAR
> Human:  cats
> Robot:  CATS EAT FISH
> > Simple questions get answers completely devoid of sense or
> > understanding of what I asked.  I fail to see any merit
> > in your 'Robot' at all.
> For neuroscience, the AI serves to illustrate an original
> http://www.scn.org/~mentifex/theory5.html Theory of Mind.
>
> Anyway, please stand by for improvements to the basic AI,
> which went up as the release of Mind-1.0 in July 2001 at
> http://sourceforge.net/projects/mind (Open Source AI)
> and is now almost ready for release as Mind-1.1 -- (but
> all the intermediate versions are also being released).
>
> Meanwhile, I am preparing to gather up all the code and
> documentation for ODP (on-demand-publication) as a book
> of which the first edition will cover Mind-1.1,
> the second edition will cover Mind-1.2, and so on.
>
> Thank you for taking a look at the JavaScript AI.
>
> Sincerely,
> Arthur T. Murray

What your AI robot needs is to get some "street-cred" - how about tracking
down the Muppet (troll) who uploaded the JS/Loop virus, and once it's
identified the culprit - throw down the gauntlet. The AI Mind could learn a
lot running a hacker round in circles!

Although from what I saw of it - some people have smarter "out of office
autoresponder message generators" one of the UK (opposition party)
politicians has an auto responder that you have to study the message very
carefully to figure out whether it was typed by a person, or constructed
from context & keywords by a machine!


---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.338 / Virus Database: 189 - Release Date: 3/14/02





More information about the Neur-sci mailing list