[RESOLVED] artificial intelligence

[stlaural]

You know what, you should make that a Web Application and let other people "talk" with your so called "AI". This way you could keep your "algorithms" on the server side so that nobody access the code but people could still run some tests by themselves instead of watching those videos.

[Arve K.]

I don't understand it... What does all those random letters mean??

[moti barski]

11 letters represent the dark areas on the image captured (the outline of the figure seen by v7)
each letter corresponds to 5 points on the image, the dark areas are marked with green dots,
for each 5 points there is a letter assigned to any combination of dark(or green marked) points.

therefore each letter can be assigned any action robotic, visual or code, if the program
feels like it she will execute the code assigned to the letter.

for example if she is fighting and she predicts a left punch (perceived as x for example), x is predicted and executed, for x can be assigned the action shoot/block/hit to left.

the numbers represent the volume level of the audio input.
1 = yes 2-4 = maybe 5-6 = no or angry

this way little memory is used, she can fight more efficiently (because talk combination are lowered )

she can use speech to text but at this time the technology is not advanced and i get only 10% accuracy

[minitech]

I still don't understand the original (v5) program after watching all the videos... some labeling would help a lot. You have one textbox above another, you type in the first textbox, sometimes the second textbox displays the last thing you typed, other times it's the last thing the "AI" parroted back from earlier... that took you 13/2 years to make?

[moti barski]

is the lable right to that textbox visible to to you?

[minitech]

There is no label to the left of the textbox... the textbox is at the left edge of the form..

[moti barski]

there is a label betwin a text box and the volume meter, that outputs stuff can everyone see it ?

[weirddemon]

I'm just praying for the day when the OP comes to the realization that this is not actually AI.

Give me an hour and I could not only replicate the actions of this "AI", but make it more believable.

[moti barski]

I'm just praying for the day when the OP comes to the realization that this is not actually AI.

Give me an hour and I could not only replicate the actions of this "AI", but make it more believable.

i talk she answer
she talks loud if she wants
she replays differently to the same input according to the situation
she has emotions

maybe you don't see the label and you refer only to the 2nd textbox from the top check out the label
with the output (near the vumeter)

what WHAT ?

[weirddemon]

i talk she answer
she talks loud if she wants
she replays differently to the same input according to the situation
she has emotions

maybe you don't see the label and you refer only to the 2nd textbox from the top check out the label
with the output (near the vumeter)

what WHAT ?

You're not making sense and the UI makes no sense. And you're delusional too =/

[moti barski]

you used the word praying, is A.I against your religion are you not objective ?

[formlesstree4]

you used the word praying, is A.I against your religion are you not objective ?

Now you're taking him too literally. I can use the word science, destruction and Matrix, but none of them are completely relevant in a whole (perhaps Matrix might be, but that's a different story).

I believe a bit of proof in the forms of coding have been requested. If you have created a true artificial intelligence, I'm sure the folks at A.L.I.C.E. would like to chat with you.

[NickThissen]

i talk she answer
she talks loud if she wants
she replays differently to the same input according to the situation
she has emotions

So you have a random response generator. How is that in any way even similar to AI? Is a 'magic 8 ball' toy also AI?

what WHAT ?

What?

[moti barski]

some of you say its parroting some of you say its random.
i think in each activation her intelligence has increased and the responses are better, she cusses
when i cuss, she greets me when i greet her, she talked about the appropriet character for the game and shows i typed.

wouldn't it be awesome to play gta san andreas when the characters answer back, to play gomoku
and the pc would cuss if it loses.

considering the input the replays are logical.

[BackWoodsCoder]

This whole thread reminds me of this saying.

"Never argue with an idiot. They'll just drag you down to their level and beat you with experience."

[stlaural]

"Never argue with an idiot. They'll just drag you down to their level and beat you with experience."

I love it!

[stlaural]

Seriously moti, videos do not prove much plus it looks like we don't see the whole GUI so NO, nobody sees the label beside the text boxes. If you really want some feedback, let people try it by themselves and publish the code source so we can all see what it actually does because it definitely looks way too simple to be a real complex AI with learning capabilities.

[moti barski]

the label doesn't always apear but its to the right of the second text box from the top and to the left of
the full progress bar

[moti barski]

tell me what you think about the upgrade
http://www.youtube.com/watch?v=6eWHWS9cAy0

[minitech]

I still see no signs of intelligence.

[stlaural]

Same here, looks more like a random string generator to me. I simply can't understand what is happening and what is being displayed.

[FunkyDexter]

I still see no signs of intelligence.

In the links or in the thread?

[minitech]

In the links or in the thread?

Both...

[moti barski]

now she knows colors

minitech what would you aknowledge as signs of intelligence ?

FunkyDexter : the joke was also about me but i admit it was so funny i choked and i mean it

[minitech]

If I can judge language recognition features to be better than the average CleverBot thing, I'll recognize it as intelligence. In fact, if you come close to half of CleverBot's "intelligence" I'll acknowledge signs of intelligence, but this is not an AI as far as I can see. Can you set it up as a web service everybody can test instead of posting vague videos on YouTube?

[moti barski]

i would like to suggest an intelligence test, if an A.I prog were to successfully complete a game very simillar to pokemon red (for example) would you consider it intelligent ?

by the way here is a script of a chat with cleverbot :
i said hello

1st
Hello.
Apple?
How are you ?
Not good. You?

What tv shows do you like ?
I like Psych.
What is it ?
'Bad Translator has infiltrated the school system? '.
What are you talking about ?
What are YOU talking about?!
Hello.
What was the last recorded important creative act?
How are you ?
Oh i'm 15.
I am fine.
Good, what are you doing?
Playing on the computer.
Is it fun?
Kind of.
Is it red?
Mortal kombat.
Nothing? Nothing?
Hadoken.
What is Hadoken?
Souyuken.
What is a car?

IT SEEMS TO DISRAGARED THE INPUT AND GIVE ANSWERS THAT REPLY TO ANYTHING (NOT SPECIFIC)
they might as well write "that's what she said" as a reply to any input

[minitech]

No, you would program it specifically to beat said game.

If you really want a test that only you can do, let's set up a chat in real time and I'll give you questions from the latest Jeopardy!. You can video-capture the screen. Of course, we wouldn't expect your AI to give correct answers, just that it correctly identify the question being asked and respond sensibly.

[minitech]

As for the chat with CleverBot... if you could write something like that I would probably be impressed.

[moti barski]

do you have a facebook account ?

[minitech]

No. Why do you ask?

[moti barski]

originally i wanted to do the chat (you suggested) via facebook chat but then i thought to my self
i should get someone official to validate a chat with the yotamarker like a news guy

[honeybee]

Well, how about you have your girlfriendbot post replies here instead of you? If we understand them, it proves she/it/whatever is intelligent. If we don't, we shall just assume you are posting under her name.



.

[moti barski]

Well, how about you have your girlfriendbot post replies here instead of you? If we understand them, it proves she/it/whatever is intelligent. If we don't, we shall just assume you are posting under her name.



.

lets hypothetically assume she posted something like "$ its all good" ($ represents wake up in the program) and other stuff simillar that wouldn't be enougth proof for someone like minitech, i think minitech's high conditions are good i will set them as the bar

[FunkyDexter]

FunkyDexter : the joke was also about me but i admit it was so funny i choked and i mean it

It was meant to be funny rather than nasty so I'm glad you took it the right way.

You need to find a way of making your demos understandable. Honestly, the last link you posted brings up a youtube link called AI blood detection but I can't see anything in it to do with blood... or detection for that matter. It just seem to be low quality video of you moving around the room. I'm mystified as to what your actually demonstrating. I'm afraid the previous links have been no more enlightening either. Seriously, look back through the replies to this thread. The overriding message is : "errrr, what?".

At the point where nobody here understands what you're demonstrating it's a pretty fair bet that it's not us that's at fault, it's you.

And don't post back something along the lines of "How do I make it clearer?" We don't now because at this stage none of us have the first clue what it is you're trying to make clear.

[moti barski]

sounds to me like you don't read video descriptions

the color detection is done for the top line of points of the eye (seen as green squares)

top label below image : r = red represent blood or red trafic light stop
red overides green

top label below image : g = green represent green trafic light go, vegatation
top label below image : n = not green or red

! = fear emotion
^ = anger emotion
0-4 = curious emotions

this is a reflex type addetive for the eye class

'TODO green walk code only, not methode like curious triggered walk
'TODO fine tune color detection calibration
'TODO detect fast vehicles from sides to pause walk action

those strings in the v7 videos are thoughts each char in the string is a part of a captured image, the chars are to be connected to actions (robotic or virtual ) , thous generating an action to respond to the real world(image captured), in other words travel, fight, look arround

to analog : string = thought, or brain electric signals

completely unrelated to this thread, anyone knows how to fix a computer mouse that clicks twice instead of once ?

[honeybee]

... Honestly, the last link you posted brings up a youtube link called AI blood detection but I can't see anything in it to do with blood... or detection for that matter. ...

You mean you found something in it to do with AI? Now, that's a discovery!

.

[honeybee]

completely unrelated to this thread, anyone knows how to fix a computer mouse that clicks twice instead of once ?

If you want a cheap and quick fix, switch the mouse buttons, so the right button does a left click.

The only permanent fix is to replace the mouse. I have had about six mice changed for this problem. (over a period of four years, btw)

.

[minitech]

What???! I've kept the same cheap ($10) mouse for 6 years (and I like it better than the new mice too!)

[sapator]

What???! I've kept the same cheap ($10) mouse for 6 years (and I like it better than the new mice too!)

You obviously haven't clicked enough

My mouse record is the Microsoft optical i have now and it's on it's second year.

[honeybee]

What???! I've kept the same cheap ($10) mouse for 6 years (and I like it better than the new mice too!)

Well, we have a mouse which visits our home every year and breeds a whole new family of mice, all completely free! Now beat that!

.