Pages: [1]   Go Down

Alternatives to classic computer-cognitive approaches?  (Read 1233 times)

ivanv

  • Trusty Member
  • Jr Member
  • *****
  • Posts: 89
  • automatic mistake machine

Alternatives to classic computer-cognitive approaches?
« on: August 24, 2011, 10:59:13 PM »
hi there spaceboys and spacegirls  :o

has anyone been involved with computer cognitive systems other than predicate system and neural networks? or experimented with the same?

mendicott

  • Newbie
  • *
  • Posts: 22

Re: Alternatives to classic computer-cognitive approaches?
« Reply #1 on: August 24, 2011, 11:41:07 PM »
Can you provide links to 3 examples of "predicate systems" ?

ivanv

  • Trusty Member
  • Jr Member
  • *****
  • Posts: 89
  • automatic mistake machine

Re: Alternatives to classic computer-cognitive approaches?
« Reply #2 on: August 25, 2011, 02:15:13 AM »
higher order logic:             http://en.wikipedia.org/wiki/Higher_order_logic
calculus of constructions: http://en.wikipedia.org/wiki/Calculus_of_constructions
type theory:                     http://en.wikipedia.org/wiki/Type_theory

i cannot think of anything else

mendicott

  • Newbie
  • *
  • Posts: 22

Re: Alternatives to classic computer-cognitive approaches?
« Reply #3 on: August 25, 2011, 02:46:11 AM »
That's just abstract theory.  I want to see examples of concrete implementations.

infurl

  • Jr Member
  • **
  • Posts: 70
    • Home Page

Re: Alternatives to classic computer-cognitive approaches?
« Reply #4 on: August 25, 2011, 03:50:54 AM »
That's just abstract theory.  I want to see examples of concrete implementations.

Wise posters leave concrete implementations as exercises for the reader.

You don't really expect us to do your homework for you, do you Marcus? :)

ivanv

  • Trusty Member
  • Jr Member
  • *****
  • Posts: 89
  • automatic mistake machine

Re: Alternatives to classic computer-cognitive approaches?
« Reply #5 on: August 25, 2011, 11:56:38 AM »
 :'(
actually i've done some experiments with implementing functions in pure propositional logic using resolution rule. in these examples i've selected boolean functions for a sake of simplicity. i assume that other functions can be implemented too. some background knowledge about propositional logic (http://en.wikipedia.org/wiki/Propositional_logic) and resolution rule (http://en.wikipedia.org/wiki/Resolution_%28logic%29) is required.

function types in propositional logic can look like this:


types could be deduced by applying resolution rule from following set of formulas. of course, some converting of resolution results from conjunctive normal form to above form (which express complex consequences) is required. these folrmulas look like this:


so, how to use these functions?
first, convert them to conjunctive normal form, as with standard procedure for deducing in propositional logic. next, if we want ie. to get result of function "T & F", we have to give following sentence (which also has to be converted to conjunctive normal form) to system:
And & (p1 -> T & p2 -> F)
then, when we apply resolution rule to this and above sentences, we get as result:
F
of course, other results deduced from the same set of sentences also might be important, although i didn't found their purpose. maybe in some complex type system?

that's about it. thought someone might be interested. greetings from zombie room...
« Last Edit: August 25, 2011, 06:00:33 PM by ivanv »

GeekCaveCreations

  • Trusty Member
  • Senior Member
  • *****
  • Posts: 630
  • Safe, Reliable Insanity, Since 1961
    • Geek Cave Creations

Re: Alternatives to classic computer-cognitive approaches?
« Reply #6 on: August 25, 2011, 04:34:50 PM »
The only problem I have with all that, Ivan (aside from the fact that it's beyond my knowledge level) is that the images aren't showing up. :( I suspect that it may be a DNS resolution problem, which certain members have experienced here before, and is therefor not your doing. :)
Comforting the Disturbed, Disturbing the Comfortable
Chat with Morti!
LinkedIn Profile
CAPTCHA4us

ivanv

  • Trusty Member
  • Jr Member
  • *****
  • Posts: 89
  • automatic mistake machine

Re: Alternatives to classic computer-cognitive approaches?
« Reply #7 on: August 25, 2011, 06:03:29 PM »
The only problem I have with all that, Ivan (aside from the fact that it's beyond my knowledge level) is that the images aren't showing up. :( I suspect that it may be a DNS resolution problem, which certain members have experienced here before, and is therefor not your doing. :)

i've updated original post with attachments reflecting first and second image.
thank you very much for your interest. :D

GeekCaveCreations

  • Trusty Member
  • Senior Member
  • *****
  • Posts: 630
  • Safe, Reliable Insanity, Since 1961
    • Geek Cave Creations

Re: Alternatives to classic computer-cognitive approaches?
« Reply #8 on: August 25, 2011, 06:08:13 PM »
Much improved, thank you. :)
Comforting the Disturbed, Disturbing the Comfortable
Chat with Morti!
LinkedIn Profile
CAPTCHA4us

mendicott

  • Newbie
  • *
  • Posts: 22

Re: Alternatives to classic computer-cognitive approaches?
« Reply #9 on: August 25, 2011, 06:31:41 PM »
Andrew, as my niece is fond of saying, "if its not on Facebook, then it didn't happen" .. ;^)

ivanv

  • Trusty Member
  • Jr Member
  • *****
  • Posts: 89
  • automatic mistake machine

Re: Alternatives to classic computer-cognitive approaches?
« Reply #10 on: August 25, 2011, 07:24:58 PM »
Hello GeekCaveCreations :)

I'm really curious, what inference rules is Morty O0 using? isn't it a predicate system?

GeekCaveCreations

  • Trusty Member
  • Senior Member
  • *****
  • Posts: 630
  • Safe, Reliable Insanity, Since 1961
    • Geek Cave Creations

Re: Alternatives to classic computer-cognitive approaches?
« Reply #11 on: August 25, 2011, 09:10:06 PM »
Morti is a 'multi-tiered educational platform', in that he's the platform I'm using to learn the basics of AI, beginning with AIML (which isn't really AI at all, but a simple pattern matching response system) and moving forward from there. His next 'tier' will be a hybrid platform, mixing some rudimentary NLP/grammar parsing with a more advanced type of pattern matching system (e.g. ChatScript) which will be designed to 'build' responses to questions and statements, rather than simply 'parrot' what's stored in his database. I've still got a lot to learn about various aspects of this next part, so it will be a while before I'll have anything to show. Till then, Morti will still be a whimsical little smart-alec, sitting not-so-quietly in his own little corner of the internet.
Comforting the Disturbed, Disturbing the Comfortable
Chat with Morti!
LinkedIn Profile
CAPTCHA4us

ivanv

  • Trusty Member
  • Jr Member
  • *****
  • Posts: 89
  • automatic mistake machine

Re: Alternatives to classic computer-cognitive approaches?
« Reply #12 on: August 25, 2011, 10:16:53 PM »
Morti is a 'multi-tiered educational platform', ...

gee, this is entirely new area to me  :D
thanks!

GeekCaveCreations

  • Trusty Member
  • Senior Member
  • *****
  • Posts: 630
  • Safe, Reliable Insanity, Since 1961
    • Geek Cave Creations

Re: Alternatives to classic computer-cognitive approaches?
« Reply #13 on: August 26, 2011, 03:30:43 AM »
Well, it sounds good "on paper", but in truth, Morti came about as a way to learn both my first steps in AI, but also to try out new concepts and ideas in PHP. He's my second "testing framework" for learning something without actually taking courses or classes in my chosen area of study. My first "testing framework" is a chatroom script called pChat, that began in 2002-3 as a "language port" of a PERL script called "Blue Chat". That particular script has evolved from a simple, flat-file based script, with no features beyond a choice of four text colors and only a single 'area' to chat in, to a database driven, feature rich 'house', with multiple rooms, private messaging, emoticons, a huge number of text style choices, and even a virtual telescope (when you're in the Observatory, at least), and has served to help me learn a great deal of new ideas in PHP and MySQL. I hope that Morti will help me do the same with AI, over the next few years.
Comforting the Disturbed, Disturbing the Comfortable
Chat with Morti!
LinkedIn Profile
CAPTCHA4us
Pages: [1]   Go Up
 

Welcome

Please login or register.

Login with username, password and session length

Users Online

19 Guests, 2 Users
Users active in past 15 minutes:
ivan.moony, Data
[Global Moderator]
[Trusty Member]

Most Online Today: 53. Most Online Ever: 208 (August 27, 2008, 08:36:30 AM)

Articles

Latest in Media

 
Zoom Details 4
Red Haired Girl - Daz Studio
Been a while since I did some rendering, Daz is looking good.
 Views: 20705
 Comments: 4
Posted by Freddy
Mar 02, 2014
in Artwork

Friends

Chatterbot Collection
Chatterbot Collection
CallMom
Chatbots.org AI Zone
Datahopa
free counters