Ai Dreams Forum

Chatbots => General Chatbots and Software => Topic started by: korrelan on April 13, 2018, 12:00:16 pm

Title: KorrBot
Post by: korrelan on April 13, 2018, 12:00:16 pm
I've not had much experience with NLP, Chatbots, Json, etc so there is obviously a gap in my experience/ knowledge.

As side project I've took a few hours to write a simple chatbot engine/ parser.

The bot uses sentence templates, a dictionary and mostly simple substitution and searches.

The bot can learn simple Json triples and also apply simple inference rules, so it knows birds can fly because birds have wings and wings can fly, etc

I've got an idea how to add conversation goals, moods and a personality etc...

Example conversation... so far lol

Code: [Select]
U: what is a dog
B: i'm not sure what is a dog

U: what is a cat
B: i'm not sure what is a cat

U: can birds fly
B: i'm not sure

U: can a bird fly
B: no a bird can not fly you dumb ass.

U: can a cat fly
B: no a cat can not fly you dumb ass.

U: does a dog have legs
B: i'm not sure

U: does a cat have feet
B: i don't know

U: does korrelan own a car
B: i'm not sure

U: my name is korrelan **Begin Learning**
B: good to know.

U: a bird is a animal
B: good to know... tell me more.

U: a bird has wings
B: ok... what else should i know?

U: korrelan is a human
B: you don't say.

U: korrelan is a male
B: you don't say.

U: a male is a human
B: good to know... tell me more.

U: male is a sex
B: you don't say.

U: wings can fly
B: ok I understand

U: wings can fly
B: ok I understand

U: a human has legs
B: ok... what else should i know?

U: feet can walk
B: ok I understand... tell me more.

U: legs have feet
B: cheers for the information.

U: a dog has legs
B: ok... what else should i know?

U: a cat has legs
B: ok... what else should i know?

U: a dog is a animal
B: good to know... tell me more.

U: a cat is a animal
B: good to know... tell me more.

U: do you know korrelan is a mammal
U: korrelan owns a car
B: sweet... tell me more.

U: a plane has wings **Stop Learning**
B: ok... what else should i know?

U: what is a dog
B: a dog is a animal

U: what is a cat
B: a cat is a animal

U: can birds fly
B: yes birds can fly

U: can a bird fly
B: yes a bird can fly

U: can a cat fly
B: no a cat can not fly you dumb ass.

U: does a dog have legs
B: yes a dog does have legs

U: does a cat have feet
B: yes a cat does have feet

U: does korrelan own a car
B: of course korrelan has a car

U: what colour is the car
B: i'm not sure. what colour is the car?

U: the car is green
B: green is a nice colour.

U: what colour is the car
B: the car is green

U: what colour is the sky
B: i'm not sure. what colour is the sky?

U: the sky is blue
B: blue is a nice colour.

U: what colour is the sky
B: the sky is blue

U: can korrelan walk
B: yes korrelan can walk

U: can a dog walk
B: yes a dog can walk

U: can a dog fly
B: no a dog can not fly you dumb ass.

U: do humans have legs
B: yes a human does have legs

U: where is the tv
B: i'm not sure. where is the tv?

U: the tv is in the lounge
B: the lounge is the best place for it.

U: where is the computer
B: i'm not sure. where is the computer?

U: the computer is in the office
B: the office is the best place for it.

U: where is the tv
B: the tv is in the lounge

U: where is the computer
B: the computer is in the office

U: tell me about albert
B: i don't know who albert is

U: albert is a scientist
B: you don't say.

U: tell me about albert
B: albert is a scientist

U: tell me about korrelan
B: korrelan is a human


https://www.youtube.com/watch?v=VfpUOgPZ21A

Haha... loads of fun...

Ed: What idiot invented the English language?... it's so contrived lol.

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on April 13, 2018, 02:36:26 pm
Korrelan suddenly realized he was missing something is all this is about haha :D

You're closer than ever now "korr".
Title: Re: KorrBot
Post by: korrelan on April 13, 2018, 03:08:46 pm
Nah!... just a fun distraction from my main project... to help keep me sane.

 :)
Title: Re: KorrBot
Post by: ivan.moony on April 13, 2018, 03:10:23 pm
Quote
What idiot invented the English language?... it's so contrived lol.

It was invented a million years ago, and still we are using, more or less the same version of it. Is it that hard to think of something better?
Title: Re: KorrBot
Post by: Freddy on April 13, 2018, 03:21:05 pm
That's really good Korrelan  O0

I have to point out that the coloured words are just the code prettifier picking up common programming words. I'll have to look at that again sometime.

Chatbot making is a lot of fun I agree.  I look forward to seeing what more you can do :)
Title: Re: KorrBot
Post by: korrelan on April 13, 2018, 03:44:47 pm
@Ivan

I hear Elvish is pretty hard to learn lol.

@Freddy

Perhaps I shouldn't have used the 'code' identifier? I didn't want a huge list to be shown.

Quote
I look forward to seeing what more you can do

Cheers Freddy, who would have thought there were so many ways of asking the same questions lol.

Hmmm... voice/ phoneme/ face/ object recognition perhaps lol.

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on April 13, 2018, 03:54:58 pm
Make it talk on its own, and what it wants most atm. And make it ask/talk to itself.

I thought the colored words were the predicate/verb. No keep it?? XD I liked that.

I could say ivan's sentence "Is it that hard to think of something better?" as "Imagine something better. Hard?" but even that leaves out things you wanted to say. English is perfect, so is other languages out there. They are modernized I think. Just look at my sentences. Hard to minimize that.

:D I bet either this will become something or his main project will utilize this.
Title: Re: KorrBot
Post by: WriterOfMinds on April 13, 2018, 04:21:25 pm
U: the car is green
B: green is a nice colour.

U: what colour is the car
B: the car is green

Does KorrBot know that "the car" here is korrelan's car that was previously referred to, or will it think that any car is green now?
(I actually can't remember at the moment what I've got Acuitas doing with stuff like this, but I think he just balks at anything marked with "the" and doesn't learn facts about it.)

The English language is only about 2000 years old, by the way (and that includes Old English, which is so different from modern English that you have to translate it for today's readers).  Attempts have been made to invent an idealized language (see: Esperanto, Lojban), but none of them have been accepted for common use.
Title: Re: KorrBot
Post by: Don Patrick on April 13, 2018, 04:58:39 pm
You make it look so easy, Korrelan. :)
Yes, English is horrible. I chose to work in English instead of my native language because I was taught English had very strict and consistent grammar. They lied.
Title: Re: KorrBot
Post by: LOCKSUIT on April 13, 2018, 05:20:23 pm
I only speak English but I'd say English is strict and consistent.

Actually English has big capability to go deep and complex......you don't want simple wimple. But if you do then you can using English. Use Simple English or your own wikidata logic.
Title: Re: KorrBot
Post by: korrelan on April 13, 2018, 06:10:30 pm
@WOM

Hehe… no the bot as no idea at this point, it just knows a car is green.  I’m writing the code to recognise ‘ownership’ etc now. So…

korrelans car is green… should create…

1   korrelan-subject-car
2   1-colour-green
3   1-model-tesla
4   korrlean-wishes-he owned a tesla

Where the second triple relates to the index of the first… hopefully, still considering the possibilities. 

I’m really not used to this type of coding… enjoying it though,

@Don

I can see some complicated coding arising…

Korrelan owns a imaginary green tesla with a red roof and three wheels… hmmm lol

 :)
Title: Re: KorrBot
Post by: ivan.moony on April 13, 2018, 06:25:59 pm
We can try talking in assembler :)

Latin is older than English, yet it has almost the same structure (although Latin has a noun declination, which I never liked). Croatian also doesn't differ from English much. German is in the same wine with English, so it is more similar to English. I suspect that almost all natural languages have same types of words, which are only pronounced and sequenced in a different way. There are always subjects, predicates, objects, adjectives, prepositions, adverbs... Like all the language creators was in the collar between themselves. It is a real mystery, I have to admit. And I bet it wasn't easy to pick a grammar that is enough expressive to describe things we want to articulate. Turning centuries probably did the most of the job, but still, there is amazing similarity between spoken languages.

It is possible that when one language introduced a cool expression advance, other languages copied it too. Starting from subject-predicate combination of our ancestors, it may be that all the languages introduced whatever new cool combination was invented in each of them.
Title: Re: KorrBot
Post by: korrelan on April 13, 2018, 07:59:52 pm
@ivan

Or... That all languages are somehow an expression/ linked to the universal 'logic' functioning of the human brain. The language commonalities were forced to arise, like all programming languages have similarities because of the computers binary logic.

 :)
Title: Re: KorrBot
Post by: ivan.moony on April 13, 2018, 08:16:44 pm
Well, an instinct has to have some means to be passed over generations. Maybe you are right that metalanguage of instinct is what influences spoken language creation. And that is the thing that I find as an important part of an artificial neural network: a language it operates on.
Title: Re: KorrBot
Post by: korrelan on April 13, 2018, 08:46:03 pm
Just as all machine programming languages can be traced back through several levels of abstraction to the base binary code, ultimately all human languages can be traced back to our neural code.

This is why we struggle to code AI's using high level languages, they lack the resolution/ tools/ finesse they are high level human designed constructs.

This is one of the reasons why I chose my neuromorphic approach for creating my AGI, I get to program using base human neural code.

 :)
Title: Re: KorrBot
Post by: Art on April 13, 2018, 09:17:15 pm
And to think, we haven't even touched on Context or Intent or Inflection, all of which can dramatically affect the meaning or even output/response.
Then there's the double entendre, puns, homonyms, etc... Are you sure you want to do this? Heh!!! ;)
Title: Re: KorrBot
Post by: korrelan on April 13, 2018, 09:23:36 pm
haha... I did say it was a side project... something different to think about when I need a rest from my AGI.

I'll leave the full chatbots to the chatbot gurus.

 :)
Title: Re: KorrBot
Post by: infurl on April 13, 2018, 10:13:12 pm
English is one of the few modern languages which has been completely analyzed and documented. It was a massive undertaking which took scores of linguists several decades but the results have all been collated and published in one magnificent volume, The Cambridge Grammar of the English Language. It could well be the thickest, most expensive, and most interesting book you ever own.

https://www.amazon.com/Cambridge-Grammar-English-Language/dp/0521431468

If you live outside the USA you can get it more cheaply from Book Depository in the UK.

https://www.bookdepository.com/Cambridge-Grammar-English-Language-Geoffrey-K-Pullum/9780521431460

Note that this is a descriptive grammar. It is all about how people actually write and speak the English language in all its dialects. It is not a prescriptive grammar or style guide written by some snotty academic who thinks they're better than you.

If $400 is too much to spend on a book then there is the student version which is much abridged and much cheaper. I have both and highly recommend them to anyone who is remotely interested in how natural language really works.
Title: Re: KorrBot
Post by: korrelan on April 13, 2018, 10:28:04 pm
Cambridge-Grammar of the English-Language…

I’ve read that book… terrible story, but at least they explain each word as you read along… hehe… the old puns are the best puns.

No but seriously… it’s an excellent resource.

I regularly use this, though you wouldn’t tell by my spelling lol.

https://dictionary.cambridge.org/

 :)
Title: binary
Post by: 8pla.net on April 14, 2018, 02:00:50 am
Yet "binary code" is another
of the "levels of abstraction".

SYMBOL TABLE:
A one.               A zero ...
A vertical line.  An oval ...

Symbols have no (single) meaning.


________________________________________
    What you've done is very neat!
    Thanks for  the Korrbot  video!
________________________________________
Title: Re: KorrBot
Post by: korrelan on April 14, 2018, 11:42:06 am
I've not had long to work on it but now it understands ownership, etc...

Code: [Select]

U: does infurl own a car
B: i'm not sure if infurl owns a car

U: what colour is arts car
B: i don't know what colour arts car is

U: what colour is korrelans car
B: i don't know what colour korrelans car is

U: art owns a car
B: sweet... what else does art own?

U: the car is silver
B: silver is a nice colour.

U: what colour is arts car
B: i think arts car is silver

U: the blue car belongs to infurl
B: so infurls car is blue

U: does infurl own a car
B: of course infurl has a car

U: what colour is infurls car
B: i think infurls car is blue

U: korrelans car is green
B: ok... so korrelans car is green

U: what colour is korrelans car
B: i think korrelans car is green

U: what colour is arts car
B: i think arts car is silver

U: what colour is infurls car
B: i think infurls car is blue

Title: Re: KorrBot
Post by: spydaz on April 14, 2018, 04:17:13 pm
English is one of the few modern languages which has been completely analyzed and documented. It was a massive undertaking which took scores of linguists several decades but the results have all been collated and published in one magnificent volume, The Cambridge Grammar of the English Language. It could well be the thickest, most expensive, and most interesting book you ever own.

https://www.amazon.com/Cambridge-Grammar-English-Language/dp/0521431468

If you live outside the USA you can get it more cheaply from Book Depository in the UK.

https://www.bookdepository.com/Cambridge-Grammar-English-Language-Geoffrey-K-Pullum/9780521431460

Note that this is a descriptive grammar. It is all about how people actually write and speak the English language in all its dialects. It is not a prescriptive grammar or style guide written by some snotty academic who thinks they're better than you.

If $400 is too much to spend on a book then there is the student version which is much abridged and much cheaper. I have both and highly recommend them to anyone who is remotely interested in how natural language really works.

While its true that there are a lot of tools to analyse the it has not been completely solved ..... During my research and travel i believe that many other languages would be easier to decipher for an artificial intelligence chat bot. the problem is that NLP is primarily focused in the English domain.   the grammar and meaning of the English language has so many nuances. where as other languages maybe Arabic  grammatically are simpler.
Norm Chomsky actually talks about these issues of the growth of the language, the colloquial usages and meanings. newly created words. in reality there is no secret answer to NLP. the techniques we currently have merely scratch the surface of data extraction.Yet in grammatical analysis only theoretical and traditional paper approaches exist. (linguistics) . 

korrelan... I like the progress;

what else does art own?

It seems as though the conversation was one sided and the question posed by the AI was not responded to... how did it handle that event? how was it sure that the next response was not an answer to the question... obviously it was handled well. Lol. I Like the way it can build the Ownership for each person. !Great!

Title: Re: KorrBot
Post by: 8pla.net on April 14, 2018, 07:12:57 pm
Korrelan in part said, "AI's using high level languages, they lack the resolution/ tools/ finesse".

I feel that A.I.'s using high level languages benefits the AI at least until computing power catches up one day.

Let's run a test to see how far a high level language can get by itself without A.I.

Test Source code:
Code: [Select]
<?php

$I
=array(
"Does person own a thing?",
"What color is persons thing?",
"Person owns a thing.",
"Persons thing is colored.",
"The thing is colored.",
"The colored thing belongs to person.",
);

$O=array(
"I'm not sure if person owns a thing.",
"I don't know what color persons thing is.",
"What else does person own?",
"I think persons thing is colored.",
"Colored is a nice color.",
"Of course person has a colored thing.",
);

foreach(
$ID as $i){
  
$input=$I[$i];
  
$output=$O[$i];
  
$response AI($input,$output);
  echo 
$response;
}


?>


Test Program Output:
Quote
I: The spacesuit is green.
O: green is a nice color.

I: The red computer belongs to Adam.
O: Of course Adam has a red computer.

I: Does Adam own a computer?
O: I'm not sure if Adam owns a computer.

I: Donna owns a spacecraft.
O: What else does Donna own?

I: What color is Bettys spacesuit?
O: I don't know what color Bettys spacesuit is.

I: Adams computer is green.
O: I think Adams computer is green.

Then the A.I. takes over, and it is less work for the A.I.   In short, high level languages help neural networks look normal.

Live Demo Site:  http://aihax.com/topic/?id=22

The use of any content from this forum post is strictly limited to discussion purposes only.
Title: Re: KorrBot
Post by: korrelan on April 15, 2018, 10:47:43 am
Cool.

Yes basic substitution is definitely part of my chat bot, it’s just a bit of fun, I’m not a chatbot guru by any stretch of the imagination…

it’s the AI($input,$output) I’m ultimately interested in.

A high level language can be used to create the illusion of intelligence, but that’s not the same thing as actual intelligence (IMO).  I just want to see how far I can go creating the illusion of intelligence with high level constructs.

It’s just another string I’m adding to my bow, so I can better understand you guys when you’re talking about chatbots and high level language parsing, etc.

Writers block occurs when a person concentrates too much/ long on a particular subject, the GTP networks/ pathways that cover the topic become over used/ they become too sensitive to the internal/ sensory stimulus of that particular problem space.  This then puts the brain at a disadvantage; it stops new ideas forming from other pattern combinations because the subject’s main over used GTP is swamping all other patterns out.  I use side projects like this to stop this happening with my brain.  I like to add to my pattern combinations rather than burn one particular subject into my cortex lol.

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on April 15, 2018, 10:54:48 am
In reply to korrelan; With my moderation of burn-in:
I work on all of AGI and what will happen soon....I'm not just studying ex. trees.....I think bout everything from self-fires to generation to search to performance to computers to connections to databases to [put anything here].....ok u get it.

I don't implement yet. But we ALL should already be um, scrambled like eggs everywhere, there's so much.......
Title: Re: KorrBot
Post by: 8pla.net on April 15, 2018, 02:24:12 pm
Korrelan,

Suggestion:

Build an original chatbot as a puppet for your AGI to pull the strings.  Korrelan said, "It’s the AI($input,$output) I’m ultimately interested in.".  Let's make it more interesting by introducing a beginner level rudimentary A.I. technique:

Test Source Code:
Code: [Select]
$decisions = array(
0=>array(0,2),
1=>array(1,3),
2=>array(2),
3=>array(3),
4=>array(4),
5=>array(5),
);

Test Run 1 Program Output:
Quote
I: Person owns a thing.
O: What else does person own?

I: Persons thing is colored.
O: I think persons thing is colored.

I: Person owns a thing.
O: What else does person own?

I: Persons thing is colored.
O: I think persons thing is colored.

I: The thing is colored.
O: Colored is a nice color.

I: The colored thing belongs to person.
O: Of course person has a colored thing.


Test Run 2 Program Output:
Quote
I: Does person own a thing?
O: I'm not sure if person owns a thing.

I: What color is persons thing?
O: I don't know what color persons thing is.

I: Person owns a thing.
O: What else does person own?

I: Persons thing is colored.
O: I think persons thing is colored.

I: The thing is colored.
O: Colored is a nice color.

I: The colored thing belongs to person.
O: Of course person has a colored thing.

Or... Ultimately your AGI may be interested in such an entry point to your chatbot A.I. being under its control.

In short, an AGI may pull the strings on a chatbot puppet and report its confidence in the chatbot response.

Otherwise:

Yes.  What you have suggested about building a new chatbot is very cool.

It was fun, my chatbot airing internationally on cable news in prime time!

For fame and glory, often AGI developers casually take part in chatbot A.I.




_________________________________________________________________________________________________________________________________________________

The use of any content from this forum post is strictly limited to discussion purposes only.  Any test source code snippets or test  program output, etc. are just test runs for the sake of a conversation about merging different A.I. techniques.
Title: Re: KorrBot
Post by: ivan.moony on April 15, 2018, 03:40:00 pm
Korr, may I ask is it possible to get NN to parse a natural language sentence? It has to covert a string of characters into a parse tree. My knowledge on NN reaches extracting a single response from a flat list of possible responses to some input. Simply, upon some stimulus, a single neuron at the top gets lit up. Is it even possible to extract a structured response?

My best shot is to recognize something as a sentence (single neuron). Then we can look back down the tree to see what is the sentence consisted of.
Title: Re: KorrBot
Post by: LOCKSUIT on April 15, 2018, 04:54:29 pm
LSTM does that no? It considers past words+order. 4 nodes in layer1 predict layer2 node.
Title: Re: KorrBot
Post by: 8pla.net on April 15, 2018, 04:58:15 pm
Ivan,

Oh, that's a good question for Korrelan.  It may be of interest that an MLP (Multi Layer Perceptron) has some related Natural Language Processing (NLP) applications, such as speech recognition and machine translation.
Title: Re: KorrBot
Post by: LOCKSUIT on April 15, 2018, 04:59:17 pm
see attachement for what I mean

it only uses 2 layers to like store many words and sentences
Title: Re: KorrBot
Post by: LOCKSUIT on April 15, 2018, 05:06:07 pm
oh and my friend's LSTM generated this:
Title: Re: KorrBot
Post by: LOCKSUIT on April 15, 2018, 05:08:37 pm
Amazing right? He trained it on this:
Title: Re: KorrBot
Post by: LOCKSUIT on April 15, 2018, 05:33:06 pm
On top of that amazing stuff, I should note he admitted his isn't the best LOL because 1) it trained on letters not words and 2) obviously he didn't invent this, and he's only 24 and not a ML ANN wiz yet (but is darn smart on so much).
Title: Re: KorrBot
Post by: 8pla.net on April 15, 2018, 05:35:36 pm
We may need an A.I. to read those 109053 lines of generated text... LOCKSUIT.   :D

I used to have an ANN on the web... I wonder where it went?
Title: Re: KorrBot
Post by: LOCKSUIT on April 15, 2018, 05:40:37 pm
I don't know what to say...

I'm flabbergasted.
Title: Re: KorrBot
Post by: LOCKSUIT on April 15, 2018, 06:21:20 pm
Well his AI did read them :D
Title: Re: KorrBot
Post by: 8pla.net on April 15, 2018, 07:29:07 pm
I just setup a Multi Layer Perceptron (MLP).

But it has only initial training at this point.

Curious? Click the MLP button on: aihax.com (http://www.aihax.com/MLP/)

A  new thread for this DNN maybe started eventually.

Did you know, DNN, stands for Deep Neural Network?

_________________________________________________________________________________________________________________________________________________
Title: Re: KorrBot
Post by: LOCKSUIT on April 15, 2018, 07:57:42 pm
Yes I knew that DNN stands for a Deep Neural Network.

Do you need help on getting a XOR net to work or what its weights look like if visualized? I got those.

https://www.desmos.com/calculator/7j1mysqtjk

and attachement:
Title: Re: KorrBot
Post by: 8pla.net on April 16, 2018, 01:02:53 am
Yes.

That image attachment is a big help.

How should the interface be updated?



Title: Re: KorrBot
Post by: Art on April 16, 2018, 03:44:03 am
English is one of the few modern languages which has been completely analyzed and documented. It was a massive undertaking which took scores of linguists several decades but the results have all been collated and published in one magnificent volume, The Cambridge Grammar of the English Language. It could well be the thickest, most expensive, and most interesting book you ever own.

https://www.amazon.com/Cambridge-Grammar-English-Language/dp/0521431468

If you live outside the USA you can get it more cheaply from Book Depository in the UK.

https://www.bookdepository.com/Cambridge-Grammar-English-Language-Geoffrey-K-Pullum/9780521431460

Note that this is a descriptive grammar. It is all about how people actually write and speak the English language in all its dialects. It is not a prescriptive grammar or style guide written by some snotty academic who thinks they're better than you.

If $400 is too much to spend on a book then there is the student version which is much abridged and much cheaper. I have both and highly recommend them to anyone who is remotely interested in how natural language really works.


I couldn't help but recall that bit with Black Adder and Dr. Johnson, compiler of the entire book of the English Language, according to himself. Black Adder is quick to have a jab...how could he resist?  ;D
https://www.youtube.com/watch?v=hOSYiT2iG08 (https://www.youtube.com/watch?v=hOSYiT2iG08)

-----
All in good fun infurl... O0
Title: Re: KorrBot
Post by: LOCKSUIT on April 16, 2018, 08:56:01 pm
My friend said those 109K lines of text came from metal bands.....Slayer, Armor Maiden, etc. The other note is the generation btw.
Title: Re: KorrBot
Post by: korrelan on April 17, 2018, 10:12:34 am
Added a few more types/ functions, etc.

https://www.youtube.com/watch?v=T3s4X3XPdTY

 :)
Title: Re: KorrBot
Post by: Art on April 17, 2018, 01:24:48 pm
Very nicely done but I'm still trying to figure out just how it knew I have a silver car?!  :knuppel2: ;)
Title: Re: KorrBot
Post by: LOCKSUIT on April 17, 2018, 01:34:24 pm
Moore's Law.
Title: Re: KorrBot
Post by: LOCKSUIT on April 17, 2018, 01:39:07 pm
Soon it'll be time to give it a test here!:
http://vhmats.iwarp.com/rich_text_2.html


LOL

Test ID: 1     Coverage: Remember, recall, classify, interpret, translate.
Given: [Remember] Tom *is a* human.
Q: Describe Tom.
A: Tom is a human.

Test ID: 2     Coverage: Remember, recall, classify, interpret, translate.
Given: [Remember] Humans *are* mammals.
Q: Describe humans.
A: Humans are mammals.

Test ID: 3     Coverage: Remember, recall, classify, verify, interpret, translate.
Given: [Remember] Humans *are* mammals. [Remember] Tom *is a* human.
Q: Is Tom a mammal?
A: Yes, [Tom is a mammal].


Harry > CIA test > "How do we know when "he" is ready?":
http://harryworkman.iwarp.com/index.html
lol
Title: Re: KorrBot
Post by: korrelan on April 17, 2018, 02:40:37 pm
@Art

I guess you just seem like a silver car… kind of person lol.  ;D

@Lock

Quote
Soon it'll be time to give it a test here!

Nice resource. 

Using NAME is a human, humans are mammals, etc must be common testing solution, strange how close my test sentences where… and I’d not see this site.

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on April 17, 2018, 02:45:25 pm
LOL "Harry's native language is "Commando""

That site is good resourse, it has CIA test, Harry, mind overview.
http://mindmap.iwarp.com/

lol one of his things says "Transform sentence."
Title: Re: KorrBot
Post by: LOCKSUIT on April 17, 2018, 05:16:27 pm
This is korrelan's expression when he saw that website:
https://www.youtube.com/watch?v=WPUJQow1QlY
Title: Re: KorrBot
Post by: 8pla.net on April 18, 2018, 10:45:23 pm
Hey Korrelan,

I am enjoying your videos.

Thank you.
Title: Re: KorrBot
Post by: Art on April 19, 2018, 04:10:56 am
Agreed! Keep it up, you're not that far away from entering those, "My Chatbot is smarter than your Chatbot" contests!!  ;)

You can make it happen... O0
Title: Re: KorrBot
Post by: 8pla.net on April 20, 2018, 12:03:37 am
INTERFACE UPDATE:

http://aihax.com/MLP/

While it's not connected to the Deep Neural Network yet.
It does write an array of the training inputs and outputs
from the brand new training interface module.

Title: Re: KorrBot
Post by: Art on April 20, 2018, 03:26:00 am
Your interface does not seem to be working properly. If one touches the colored area with their cursor and scrolls the mouse wheel, the colored area shifts upward as shown in the screen cap.

Probably just a slight fix but I thought you should be made aware of the snafu.
Title: Re: KorrBot
Post by: 8pla.net on April 21, 2018, 04:48:54 am
Thanks Art!

I can't explain that. It's like it has a mind of its own.
Maybe the Deep Neural Network was hungry
for those training inputs, and tried to escape.  ;)

Apologies Korrelan for my deep neural network
running amuck on your thread with your readers. 

It's back in the cage now ( I think. )
I tightened the chains on the interface.




Title: Re: KorrBot
Post by: korrelan on April 21, 2018, 11:36:47 am
Quote
Apologies Korrelan for my deep neural network running amuck on your thread with your readers.

Not a problem… It’s always good to see other projects, no matter how unruly they are lol.

As I improve the Korrbot and it’s functions become more complex, I’m finding it difficult to track what it’s learned.

So I’ve wrote a little app to display it’s knowledge trees.  I can now easily visualise the data and add to it visually/ dynamically.

https://www.youtube.com/watch?v=-zRmTWF5d3g

This is the first incarnation of the code; It will be improved along with Korrbot. The link colours will represent the predicate and the nodes the objects and subjects. It runs in 3D as well and has been designed for ease of use on my large touch screens.

If anyone has any ideas to improve the visualisation or vids of similar systems I can nick ideas from... speak up.

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on April 21, 2018, 01:10:38 pm
I really like where this is going.

Can you make it form i.e. generate discoveries by first searching through the branches ex.
Hammers canBreak crystals, glass isA crystal, therefore hammers canBreak glass!
Hammers canBreak glass, windows areMadeOf glass, therefore hammers CanBreak windows!
Hammers canBreak windows, windows cost money, therefore hammers canBreak money!)

Is it possible to combine the 3D tree/net you have there with Parse Trees to make it stronger?

And what about the attachment to this post (see attachment).

Keep it simple so you can make new milestones. Do something new here.
Title: Re: KorrBot
Post by: korrelan on April 21, 2018, 04:29:34 pm
https://www.youtube.com/watch?v=dgbGAKjdsC0

Haha... highlighted a problem with double 's'... glass, I'm sure many peeps have had that problem before... lol

Hmmm... time to explain the transference/ conversion of energy types and material properties to the bot... lol

I'm glad this chatbot is just a bit of fun... and my AGI will solve all these problems.

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on April 21, 2018, 06:27:40 pm
Hmm, it made a big connection loop I see there. So basically you said "Can hammers break windows?", then it says hmm maybe it's true, then it search-verifies the Transfer Learned (Acquisition) generation from you and calculates that it is true and says yes.
Title: Re: KorrBot
Post by: LOCKSUIT on April 22, 2018, 10:54:28 pm
Also, I realized something. It could learn "Iron-man loves crystals." (instead of "Windows areMadeOf glass."), and this presents a problem to which I have a solution. See, your algorithm found a in-direct connection loop that A=B B=C etc etc but....it can do that with anYthiNg, and easily become ultra pointless. Solution? You need confidences (weights!) of how much ex. B=C. Because if glass literally IS or "=" crystals, then hammers can break crystals and likely glass too, whereas if ironman loves crystals then it won't conclude hammers can break his iron-solid body if it really isn't connected or "=" crystals. Confidences (connection weights) also for how many facts support something (3 things weight-in one 1 node). There's also positive/neutral/negative integer ranks. Also another type of confidence for the probability of how true a fact is, ex. "Hammers canBreak iron man." has a probability of being 4% true.
Title: Re: KorrBot
Post by: korrelan on April 22, 2018, 11:39:34 pm
Ed: Yes the chain event with using Iron man would have caused a problem... and yes I solved the problem by storing numerical values against the objects. So the tensile strength, brittleness and thermal properties for example can be applied to a set of functions.

I've just started writing the library of functions; so the bot can apply numerical judgements, formulas, rules, etc.

Still a bit rough around the edges but I'm very limited on time just lately.

https://www.youtube.com/watch?v=CLWpcB4fF38

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on April 23, 2018, 11:53:50 pm
Hmm, if HUGE marshmallows (their tiny core) areMadeOf crystals, glass isMadeOf crystals, windows AreMadeOf glass, and windows cost money, then, if you ask "Can hugeMarshmallows break windows?", well, it can't begin the loop/connection around BY "predicate weight" because in the case it had learned that hammers canBreak crystals - the link predicate "canBreak" is not as strong as MadeOf (Madeof is stronger yet no proof hammers can break anything). And here with the hugeMarshmallows we have a predicate link "MadeOf" to crystals which is strong but doesn't say anything about being able to break anything. Hence this example loop would loop around and wrongly conclude hugeMarshmallows canBreak windows if asked so. SO, weight forwards a connection? No... So then is it predicate NAME? I guess so.
Title: Re: KorrBot
Post by: LOCKSUIT on April 25, 2018, 12:49:56 pm
fixed typos
Title: Re: KorrBot
Post by: korrelan on April 25, 2018, 06:06:29 pm
At this stage of play it can involve several variables, so the hammer has a mass, velocity and material type which also has its own properties (so foam hammer can be defined) like brittleness, tensile, etc… and the impact surface has its own parameters.

I added access to a vision module so the bot can recognise objects/ users etc… I was just messing around with the data formats.  The vision module is based on my AGI design and simply outputs an id/ index of a recognised face/ object/ scene into the chat bot.  The vision module is self driving though, so it picks salient objects… the chatbot has no control over it… yet.

Also added logical assumptions… so ‘art owns a silver car’ adds the facts that art is probably a human, he owns a car and its colour… just from the one sentence.

This project is still in its very early stages; and I’ve only spent around seven hour’s development time so far.  I’m still getting my head around the required data structures and their visualisation, etc.

Keep in mind this is just a bit of fun, a side project… there is no way that databases and tree search algorithms will ever produce anything remotely close to a ‘true artificial general intelligence’… they might produce a decent conversation though.

And yes… I know I’m re-inventing the wheel, there are a lot of very capable chatbots out there… this is mainly for my benefit… and you never know I might find a better way… or not lol.

I'll keep posting vids of my progress as I meet milestones.

 :)
Title: Re: KorrBot
Post by: unreality on April 25, 2018, 06:12:06 pm
there is no way that databases and tree search algorithms will ever produce anything remotely close to a ‘true artificial general intelligence

And hence the reason you fail.
Title: Re: KorrBot
Post by: korrelan on April 25, 2018, 06:17:08 pm
You do it your way... and I'll do it mine.

We are all entitled to our ideas/ opinions... aren't we?

 :)
Title: Re: KorrBot
Post by: Art on April 25, 2018, 06:54:15 pm
The distinction is that one is drawing from a database and the other one is Building such a database of knowledge acquired from experience it learned.

Then again who knows what lurks in the mind of A.I.s? ;)
Title: Re: KorrBot
Post by: unreality on April 25, 2018, 07:22:59 pm
What are you referring to, Art? What is drawing? What is building?
Title: Re: KorrBot
Post by: LOCKSUIT on April 25, 2018, 07:32:03 pm
Oh this is getting really good.



So you made vision object recognition select IDs ex. if it sees a hammer then the output into the chatbot is "hammer"?



WHAT? Dude. Listen carefully.

First of all, those triples you are using like "Iron isVery heavy." are not just triples, they can be 2s or 10s watch "Sam died." "You eatHealthy food, you canWin theOlympics."

Secondly, your database/tree can actually be a complex network hierarchy with any node able to link to any node and also that shares features like properties/facts.



@Art's comments: those should be part of the real AGI system. Drawing-from network database, generating/tweaking desires. Building knowledge database from experiences and doing KDD in it.
Title: Re: KorrBot
Post by: korrelan on April 25, 2018, 11:54:45 pm
@Lock

Yes… triples can be nested to any depth… hammer-madeof-(iron-mass->100), etc.

Yes… the ‘triples tree’ does have cross links, loads of them lol.

Making ‘discoveries’ is not actually my main goal with KorrBot, learning and conversation are key… and the data structures.  The bot is using standard rules and searches to link simple words/ concepts (collections of words/ variables) together.

As far as my main project goes… I think your definition of ‘discoveries’ is a little grand for what the system is achieving at the moment. 

Remember our conversation on my project thread.

http://aidreams.co.uk/forum/index.php?topic=10804.msg51381#msg51381

All discoveries are made by recognising commonalities between diverse concepts.

Understanding is achieved by recognising commonalities between similar concepts.


One of the learning mechanisms we use is recognising commonalities.  This happens at many different levels/ resolutions within the connectome.  Even down to learning to see… your visual cortex is basically recognising the commonalities between the frequency patterns being generated by the rod/ cone/ ganglion cells behind (Ed: In-front) of your retina. 

So in my AGI videos when you see a pattern being injected into the connectome, and the bar graph rises to signal confidence in recognising that single pattern, even that is a type of ‘discovery’.  I’ve not shown or taught the AGI how to learn/ categorise or even distinguish between the patterns, it’s ‘discovered’ how to do it all on its own.  It’s this mechanism/ ability once compounded over many layers of complexity that will enable the AGI to make the kind of ‘discoveries’ I think you have in mind.

KorrBot requires a totally different type/ style of programming/ thinking to my main project.  By running two projects with similar goals but totally different schemas I’m leveraging my brains natural ability to cross link/ find commonalities/ make ‘discoveries’ that could hopefully help me with both projects.

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on April 26, 2018, 02:15:18 am
fixed more typos
Title: Re: KorrBot
Post by: Art on April 26, 2018, 04:32:36 am
What are you referring to, Art? What is drawing? What is building?

I was comparing bot types.
There are those with static databases, holding thousands of words and meanings of various types.
There are those that learn through repetition and building their databases based on the interpretation of items and words as they relate to each other and what it believes it has learned on its own.

There was an older chatbot, that started out with only a basic set of instructions. Over time, it learned how to craft sentences and assign priorities and proper or best case usage of certain words. It was learning through positive reinforcement and self-trial.

Its counterpart had the similar construct of a framework but it had a database at its disposal from which to draw upon. While it seemed pretty adept at producing more logical responses, that was as good as it would ever get. The other bot, in time, surpassed it.

I think the flexibility of allowing the bot to learn more self-directed or even from human reinforcement results in a far more capable bot than one using a fairly static database or one in constant need of upgrading from it's "master".

Just my thoughts...A bot that can learn on its own would be a great thing.

Title: Re: KorrBot
Post by: unreality on April 26, 2018, 04:41:50 am
Art, exactly! Flexibility is the secret. I'm betting that I'm not the only one who's had the realization, "Gee, chess engines annihilate humans. Make an engine that's more flexible that plays the game of life."  Confirmed! :)
Title: Re: KorrBot
Post by: korrelan on April 26, 2018, 09:57:54 am
KorrBot is a bit crude atm but it uses a few methods for learning, these will expand as I get more adept lol.

If you look at any of the last few videos you will notice a column on the triples grid labelled ‘Cnt’. This tracks the number of times a triple has been correctly referenced; it’s a measure of validity for the item of knowledge.  This is used when the system is resting to search for new relationships between the triples.  It will only try to derive fresh insights from the triples that have been reinforced/ proven through experience to be correct.  It also does this with a limited branch search during conversation; this is how it can connect concepts… sparrow is a bird, birds have wings, wings can fly… therefore sparrows can fly, etc.

There is direct user input of course; each user has a truth/ validity score which sets the ‘Cnt’ variable for any new knowledge.  This is so new users can’t mess with the system, but trusted users can.

It creates a set of triples/ record for each topic/ object/ user that can referenced by the index of the root node for that concept.  This enables the system to create triples between complete/ complex concepts… The indexes can be seen in the leftmost column of the triple grid.  Conversations/ triples are user/ date stamped so the bot can recall who said what and when.

If the system can’t parse a sentence correctly it asks you enter the same sentence using a different syntax.  When it does understand the miss-understood sentence is reformatted and added to the sentence templates.

I’m testing moods, so the bot can both recognise the users mood and have its own moods.  This data is collected by asking the user what mood they are in before and during a learning session; this ‘mood index’ is then added to the sentence templates and triplestore to be referenced by the bots moods.

I’ve started writing a 3D world/ object definition module, so the bot can understand/ relate to spatial relationships of/ between objects, etc. So as well as knowing the ‘word’ for the location of an object/ person it has an idea of its actual 3D location/ distances involved/ etc.

Various import routines for Db/ Json and delimited files.  I’ve wrote the code so several bots on the same network can easily exchange triples, etc.  The system does not require the ‘cloud’ lol, it runs standalone.

Fixed coded functions for certain types of triple comparisons, distance, time, speed, weight, mass, logic, Math, etc…

A user can enter information through the graph interface… nice and quick, easy.

Once I’ve built the sentence template store to a sufficient complexity it will be able to learn/ glean information from the internet.

If the bot is trying to form a connection/ commonality and requires certain information to make the decision, it asks for it, and obviously stores it.

The bot has a short term memory regarding the conversation, this is how once someone/ something has been mentioned by name, the bot can understand who ‘he/ she/ it/ that’ is/ means.

I’m testing conversation trees, so the bot can learn the common order/ format for asking/ answering questions… the flow of conversations, etc.

No knowledge is instantly fixed; it designed so experience reinforces knowledge.  If the bot finds a commonality in the triplestore that’s not been explicitly expressed, it asks if it’s correct before reinforcing it.

I’m also trying to keep the system as flexible/ versatile as possible for future upgrades, and the data sets as compact/ fast as possible.  The triple table for example once parsed comprises of only dictionary/ data indexes integer/ long variables. Each root node has link index to all other relevant nodes, so it doesn’t need to search millions of triples, recall is instant.

Nothing magical lol…early days… just straight forward coding…

@Lock or anyone else…

If you have the spare time you could compile a triplestore import file, just a simple list using “\” as a delimiter on any subject you like... I can use for testing.  Stick to single words for the object/ subject at the moment, keeps testing easier. Keep the predicates short and simple, is a, has a, made of, etc.

1\car\is a\vehicle
2\vehicle\can\move
3\car\has a\door
4\car\has a\boot
5\car\has a\trunk
6\car\has a\wheel
7\wheel\has a\tyre
8\tyre\has a\tread
9\wheel\has a\hub
10\hub\has a\nut
11\wheel\shape\round
12\car\has a\colour
13\10\count\six   <<< referencing the triple index as the object = ((hub\has a\nut) \count\six)
14\3\count\four
15\6\count\four
16\car\is a\object

etc…

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on April 26, 2018, 03:19:16 pm
Wow korr nice.
Title: Re: KorrBot
Post by: korrelan on April 26, 2018, 09:26:13 pm
I’m not entirely sure which question I’ve not answered lol.

Also… you need to calm down… Korrbot is just a simulation/ imitation of intelligence for the sake of a good conversation.  It’s an exercise in logical programming, just to see how far I can take it, it will never be sentient or self aware… even though hopefully it might appear so.

That’s what my main project is for…

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on April 26, 2018, 10:51:54 pm
Ok I'll calm down.
Title: Re: KorrBot
Post by: LOCKSUIT on May 03, 2018, 04:42:57 am
nvm
Title: Re: KorrBot
Post by: spydaz on May 03, 2018, 02:37:41 pm
I’m not entirely sure which question I’ve not answered lol.

Also… you need to calm down… Korrbot is just a simulation/ imitation of intelligence for the sake of a good conversation.  It’s an exercise in logical programming, just to see how far I can take it, it will never be sentient or self aware… even though hopefully it might appear so.

That’s what my main project is for…

 :)


But 'the techniques your using do work and give a good conversation to the user as well as learn data. The replacement technique works well; just the data collection is not desirable. For pure conversation its only limited by imagination... plus even this technique has a positive evolution pathway if followed and developed. (obviously it would not be your sole technique) I think it produces suprising conversation from the users perspective as it Learns / Remembers and Recalls. 
Title: Re: KorrBot
Post by: 8pla.net on May 05, 2018, 07:47:38 pm

1\car\is a\vehicle
2\vehicle\can\move
3\car\has a\door
4\car\has a\boot
5\car\has a\trunk
6\car\has a\wheel
7\wheel\has a\tyre
8\tyre\has a\tread
9\wheel\has a\hub
10\hub\has a\nut
11\wheel\shape\round
12\car\has a\colour
13\10\count\six   <<< referencing the triple index as the object = ((hub\has a\nut) \count\six)
14\3\count\four
15\6\count\four
16\car\is a\object

What happens to 13 if we move 10 to 17?
Title: Re: KorrBot
Post by: infurl on May 05, 2018, 11:44:11 pm
What happens to 13 if we move 10 to 17?

The leading numbers are not indexes, they are primary keys and as such don't change. If you were to scramble the order of the items, the numbering wouldn't need to change.
Title: Re: KorrBot
Post by: LOCKSUIT on May 06, 2018, 12:04:21 am
But what if you changed the beginning number! 0O?
Title: Re: KorrBot
Post by: korrelan on May 23, 2018, 11:48:00 am
I bit more work done on KorrBot.

https://www.youtube.com/watch?v=l-7NmyhirJw

I'm slowly getting my head around this.

Starting to extract a lot more information from short complex sentences.

 :)
Title: Re: KorrBot
Post by: LOCKSUIT on May 23, 2018, 12:10:13 pm
Another video to add to the base!

:)
Title: Re: KorrBot
Post by: ranch vermin on May 23, 2018, 12:24:05 pm
wow first time ive seen this,   done a very good job, showing off the power of assignment.

This would be cool for a supervised learning/implanting technique.

what I would think to do if I was making this would be to combine it with machine vision, and then the robot could search through the video for the labels and maybe you could get the robot to have "implantable methods" - that are just written in plain english to it.

Cause how else would you do it?  You need labels or you cant tell the robot anything.  otherwise its just stuck in its primitive goal or your doing basicly this but in a more primitive way.

Why cant you just tell the robot and then it learns - because theres the problem where it cant distinguish the language labels from the actual thing that it associates with, and this would skip past that issue.
Title: Re: KorrBot
Post by: LOCKSUIT on May 23, 2018, 12:28:56 pm
Yep that's better in all ways around. Human Speech only developed not long ago. Visual sentences were what all animals used only. "Hammer smashing crystals.". There's no way to interpret it wrongly. It's universal. First you teach it, then it discovers more.

:)
Title: Re: KorrBot
Post by: korrelan on May 25, 2018, 10:04:10 am
Grammatical rules are for the obedience of fools and the guidance of wise men.

Ok… besides the fact that ‘verb’ is a noun, I now have ‘nerbs’ lol. 

This makes my life much simpler, so henceforth… ‘fly’ is a nerb.

 :)
Title: Re: KorrBot
Post by: spydaz on May 27, 2018, 11:38:58 am
Yep that's better in all ways around. Human Speech only developed not long ago. Visual sentences were what all animals used only. "Hammer smashing crystals.". There's no way to interpret it wrongly. It's universal. First you teach it, then it discovers more.

:)

All living beings are born learning....even animals learn from their parents... :-X
Title: Re: KorrBot
Post by: LOCKSUIT on July 11, 2018, 09:05:13 am
"This is the first incarnation of the code; It will be improved along with Korrbot. The link colours will represent the predicate and the nodes the objects and subjects. It runs in 3D as well and has been designed for ease of use on my large touch screens."

Korrelan,    where is KorrBot ?

Is there a new incarnation of the visualization?

Please provide the sauce.
Title: Re: KorrBot
Post by: korrelan on July 14, 2018, 02:55:08 pm
Quote
Please provide the sauce.

Why? Do you want to ketchup with the project… hehe.  There are no ground breaking snippets of code within the bot, it’s just standard Chatbot NLP parsing linked to a custom DB that handles inference, etc.

KorrBot is a side project that I switch too when I’m tired of my main AGI project.

It’s still under development, I have about twenty pages of ideas I wish to implement/ try.  It can hold a decent conversation atm but I can take it further… much further.

I’ll find the time to do an update… sometime soon lol.

 :)

Title: Re: KorrBot
Post by: LOCKSUIT on July 14, 2018, 03:13:15 pm
I won't lie, I had the korrbot visual knowledge graph in that video on my screen for hours studying it until concluding some things :3