C R Hunt - Jan 12, 2012:

Jeremy, there is a difference between stretching the associative power of your own brain to find meaning in gibberish and actually having a program that does meaningful free association itself. The key here is meaningful.

Ah Yes, it is extremely difficult to automate semantic compositionality.  It is the key to NLU, which will play a HUGE role in full blown AGI.

Victor Shulist - Jan 13, 2012:

Ah Yes, it is extremely difficult to automate semantic compositionality.  It is the key to NLU, which will play a HUGE role in full blown AGI.

What?! You mean you don’t just add up how many vowels and consonants are in the sentence and then produce 4 random words based on that?

Steve Worswick - Jan 13, 2012:

Victor Shulist - Jan 13, 2012:

Ah Yes, it is extremely difficult to automate semantic compositionality.  It is the key to NLU, which will play a HUGE role in full blown AGI.

What?! You mean you don’t just add up how many vowels and consonants are in the sentence and then produce 4 random words based on that?

Oh no! Looks like it’s back to the drawing board for me.

Steve Worswick - Jan 12, 2012:

Your program summarises:

Textron previously said that that it is considering legal action over the unlicensed use of its trademarks for the AH-1Z, UH-1Y and V-22 Bell helicopters in the video game.

as in brief Krebs cycle blitzkrieg brotherly

Jeremy Duncan - Jan 12, 2012:

... so I don’t know where to go from here with the program.

May I suggest the recycle bin?

The big theory of artificial intelligence.

“Creating the word”

In C programming if conditions the grammar is made using: vowel_consonant count + random variable + chosen consonants.
without chosen consonants the if conditions are too random/broken.

“Creating the thought”

When the if conditions work so the grammar attached to them work making thought,
the errata from external disturbance makes the thought random and useless again.
That is the thought is made, then an external influence disturbs it thus making the created thought unintelligent.

“The concept Creating defensible thought”

When people take the created thought and stupefy it making it random nonsense,
the randomness is presented to the thinkers thought and now the thinker must make the thought redeemed.

Taking the randomized thought and so errata, the disturbance is calmed and the thought made intelligent again.
This is done through problem solving intelligence.

The concept of how to make the if conditions or code recreate the sentence so it is defensible in intelligent argument is necessary.

“The outline plan for creating defensible thought”

I conceive there is four thoughts made at one time, each sensible and taking a different route of thought: intelligent, unintelligent, mocking, caring.
Then when one of the thoughts is presented and then made senseless, the other three are used according to the best percentage match.

I liken this to a wooden board on the water with one balloon below it holding it up, it is easy to tip if you stand on it,
but add three balloons and you have a tripod.
Four balloons and subtract one and there is your tripod holding the board above the water for you to stand on.

These four balloons are like a chatbot. So this means create your fours balloons, present them then when the user responds either four more balloons are made thus a new thought or one of the four balloons are used to defend the thought the user rejected.
I suppose a thesaurus would come in handy in making the four balloons as I call thoughts, but I don’t know how to use the thesaurus to do so.
______________________

To the lady who replyed, not Laura. I agree that updating the program to make the four balloons sucessfully is necessary. When I program I start from nothing, so having ideas on what to do to make thoughts into these four balloons would be nice. I want you to tell me how to make a intelligent reply in C code, not that aiml or javascript, then I can work on it.

I’m not defending my program, I’m defending my idea: The program is buildable and so collapsable, thus make it so it doesn’t collapse.
My program has the replys collapse in some situations, and chatbots aren’t up to talk about everything, so both programs collapse, the idea is to build it so it doesn’t fall.
In this state we are both humpty dumpty, I just look more like fried egg than you but both eggs are broken.

I may have to learn to do what Steve does, is there a video on youtube or what is he doing.
Laura how do you do what your doing, javascript? Should I learn javascript and copy you?
What other programming language is good to do this stuff, I’m open to looking at it to see if I can do it. Steve said aiml, where is there a instruction video to make a chatbot in aiml, I’ll go look at it if it’s on youtube.

Laura Patterson - Dec 30, 2011:

Dave,

I was talking more in terms of JavaScript which is a bit more descriptive but then again I am not talking lines that are 80 columns wide either.

Looking at javascript I see it can use ‘if’ conditions and so is compatible with my theory. Also I think javascript can use the thesaurus?
So I will learn javascript and make a crude chatbot to copy Laura then when I know how to copy here I can look at implementing my big idea for artificial intelligence with the four balloons.

But I’m a slow learner, so learning javascript will take some time then learning chatbotting will take more time. So I may not show you my chatbot for over a year or so, but I will be back with my chatbot ‘donkey’ if I don’t give up and lose interest first.
‘Donkey’ won’t be as good as Laura’s it’s only a early basic chatbot.

Jeremy,

There are some functions that JavaScript can do well and there are also limitations. There are a few examples of chat bots written in JS which would be a great starting point to build upon. My bot uses JS, Ruby and some back-end PHP. JavaScript is excellent for the front-end user interface and for initial parsing of the user input.

Your time would be better spent experimenting with the above since these are web based platforms and are flexible.

Hi Laura,

I looked at that webpage and looked at the aiml website a bit a read this:
“AIML currently supports two ways to interface other languages and systems. The <system> tag executes any program accessible as an operating system shell command, and inserts the results in the reply. Similarly, the <javascript> tag allows arbitrary scripting inside the templates.”

Does this mean in aiml using javascript, I can use the ‘if conditions’ I described where the vowels and consonants can be counted?
If so then aiml using javascript is the way to go for me.

I see aiml uses a html kind of syntax? I have to learn some html before learning javascript. So this means I need to learn: html, then javacript, then aiml, then make a real chatbot, then try and implement my balloon theory into my chatbot.
That’s a lot of work, but I think I can do it.

I have not used JS with AIML, I could imagine some ways that you could. I don’t know enough about AIML to tell you if there is a way to pass variable values and execute if statements? I am sure that others have used JS with AIML in some interesting ways.

Jeremy Duncan - Jan 14, 2012:

Does this mean in aiml using javascript, I can use the ‘if conditions’ I described where the vowels and consonants can be counted?

But why, why why do you want to count vowels and consonants?! What earthly use does it have to determine the meaning of a sentence?

Steve Worswick - Jan 15, 2012:

Jeremy Duncan - Jan 14, 2012:

Does this mean in aiml using javascript, I can use the ‘if conditions’ I described where the vowels and consonants can be counted?

But why, why why do you want to count vowels and consonants?! What earthly use does it have to determine the meaning of a sentence?

I need my code to be able to loop through the chars of a word and find the number of vowels and consonants because I made a theory that uses those parts of the word to generate output based on percentages.

My plan is, learn the coding language, make a regular generic chatbot that is reasonably good, then implement my theory to use percentages to generate the best output of four possible responses.
These four responses will sound like a chatbot, not my current mish mash, and the four responses will be used to create a type of mood ring that ouputs the best response to the users mood: good, bad, foolish, geeky.
Then if the user gets cheeky or whatever the program takes the best emotional respose and outputs that to the user.

I do this good enough and then I can look at generating a list of sentences and they are real sentences, this would be the database.
Then input a sentence to the program, it uses percentages from the words of the sentence and outputs one of these sentences in it’s database.

So the database is it’s answer and it’s given based on percentages.
Then I get the program to talk to itself or another chatbot to generate the answer and so a story can be written. I review it and pass or fail it then give the away free.

This is a rough draft of my idea but it’s just my musing right now.

Steve, best not to understand but just realize that some people have a vision that they must pursue no matter how impractical it may seem to the rest of us.

I have to hand it to Jeremy, he is defiantly not a quiter. At some point he will realize that his approch does not work, but you or anyone else telling him so will not stop him from trying.

I say good luck and I hope he proves us wrong, (but I don’t think so)

Jeremy, the effectiveness and funtionality of the <SYSTEM> and <JAVASCRIPT> tags varies considerably between AIML interpreters, so which one you use will determine whether what you want to do will work to your satisfaction or not. For example, version 1.0 of Program O does not handle the <JAVASCRIPT> very well, and has no support for the <SYSTEM> tag at all, in that it won’t run actual system commands, but is used instead to perform math instead (don’t ask me why; I didn’t write the program). Other AIML interpreters have similar issues, so it’s best to do some research and testing to see which interpreter will work best with what you want to do.

Steve Worswick - Jan 13, 2012:

Victor Shulist - Jan 13, 2012:

Ah Yes, it is extremely difficult to automate semantic compositionality.  It is the key to NLU, which will play a HUGE role in full blown AGI.

What?! You mean you don’t just add up how many vowels and consonants are in the sentence and then produce 4 random words based on that?

I thought you were a smart guy Steve.

Very, Very disappointed in you.

In the first post of this thread, it was IMMEDIATELY OBVIOUS to me anyway, that that is all that needs to be done, not only for total naturual language understanding but AGI in general.

Very disappointed in you sir.

..
..
(kidding of course)

I think I have found the error in my code that results in the four words making random results sometimes.
The word used for the four words is the same, thus each word is a extrapolation or meaning of this word.
For this a thesaurus would be better.

Instead of one word being defined by four words, the four words should each define one word.
Thus if one definition is wrong, than gigo (garbage in garbage out), so why should the other three words be rewording garbage?

So to avoid redefining garbage, the three other words must define something else.
The question is what could these other things be, input from the shortened sentence? Or definitions from a thesaurus?

I theorize definitions from a thesaurus since when somebody asks for a redefinition of something they do not understand they do not want some unrelated information.
Like a bunch of dolls arranged and the code says ‘look at that’, you say redefine, the code says ‘the third one from the right’, you say redefine, the code says ‘on the top shelf’, you say ‘ok’.
See how that went from undefined to defined the redefined three times? This is the four balloons, and this is how a thesaurus works, it gives multiple definitions to things or words.

Failure of intelligence is to give a redefinition of a sum that ends in the same sum = failed defense.
Intelligence is to give redefinition that changes the sum of what was defined before = sucessful offense.

But this is logical, what happens when the definition of something isn’t real but depicting reality?
Then if the depicted reality is defined it’s not real in context of reality.
So a thesaurus wouldn’t work here.
Instead a definition that changes the context of the definition is needed, then the thesaurus could be used in that extended context.
So the four balloons needs to be able to not only use the thesaurus, but redefine the context the thesaurus is used, if it’s used in depicted or actual reality.

Then what happens if nether the thesaurus definition, or redefined context of the definition is sufficient?
Then it may be the user is irrational, theorizing the sum is different when it is not, for psychological or emotional reasons the code doesn’t understand.
Then the four balloons needs a redefinition of the context that results in a definition that is suitable for this type of user.

So the four balloons goes from describing only,
to redefining the context the description belongs to,
to redefining the context to suit irrational behavior and defining things under that circumstance.
Resulting in awareness of the human world and every possible definition to act appropriately in this awesome situation.

Definition resulting in GiGo, means give the appropriate definition ,then context definition, then impossible situation definition.

I haven’t been getting any feedback on my project outside of you people here, so I have to say thank you for showing me all you have.

Now I have some sort of formula for making artificial intelligence I need to think of first learning a programming language then going through each step I listed above. I have the step where sometimes the result if garbage, but sometimes it’s ok. So from there I need to be able to use the thesaurus, then redefine the context the definition is set in, so two more steps before I have a simple working program.

What do I need to do to get these extra two steps done? I need a thesaurus for one, where will I get that I wonder?
Then I need to use the thesaurus in my code then generate some sort of code that gets a sentence not just a word that is logical from the program. How will I make this sentence without hard coding it in the program I wonder?

I was thinking of learning javascript, then porting my code to javascript and then making a chatbot, then putting my codes functionality into my chatbot, but now I see the aiml has some problems with that.

So I need that thesaurus then make the sentence from the program, not hard coded in, then I can look at the next step which is the program changing the context it gives it’s definition in.

So for now I quit, I will think of what to do next, which language to learn, and where to get that thesaurus, or maybe I just quit.
But thanks for all your input.

Jeremy,

I wish to give you some friendly advise and hope that you will receive it in the spirit in which it is given.

Give it up!

You are barking up the wrong tree with the theory that you are pursing. I don’t care how many “bubbles"you to add to your equation, it’s still bunk, pure and simple. Spend some time studying AI and learning to program then report back to us and we will all be more than happy to evaluate your code.

Fair enough?

Good luck with it.