Artificial Intelligence Messaging
+3
iConnor
awasteoflife
John(GmMkr-Admin)
7 posters
Page 1 of 2
Page 1 of 2 • 1, 2
Artificial Intelligence Messaging
Well I'm sure we all know what AI is. So let me get to the point.
I would like to discuss AI messaging. Like having a conversation with your AI.
We would first have to program the base knowledge for the AI such as greetings.
First we would have to program the AI to know when it is being greeted.
Depending on your greeting it would greet back.
You: Hi there.
AI: Hello.
The AI simply just checked for the word 'hi' seeing how it was the greeting. 'There' was just an extra.
So if you were to say.
You: Hi bob.
AI: Hello.
It would respond in the same way.
Now here is where it gets more complicated.
What if we also want the AI to check the other words.
In the above scenario we would want the AI to check if its name is Bob.
Although this would be possible with the 'get_string()' function it would
be extremely time consuming and you would make very inflexible AI.
Now, you also want the AI to learn. Well since it really can't learn, it just
stores information. That would be a little too perfect.
You wouldn't want this.
You: Remember this- Bill enjoys lots of candy.
You: Now, what did I tell you to remember?
AI: Bill enjoys lots of candy.
As much as you would want this.
You: Remember this- Bill enjoys lots of candy.
You: Now, what did I tell you to remember?
AI: Bill likes a lot of candy.
That would be difficult to program but it would make the AI
seem more human.
We would need to check for a lot of input. Here is some color-coding.
You: Remember this- Bill enjoys lots of candy.
You: Now, what did I tell you to remember?
AI: Bill likes a lot of candy.
The red section tells the AI that it needs to store information.
The green and blue sections are parts that the AI can change if it wants
to words with the same meaning.
Although choosing 'random()' isn't the way to go it would make it work.
You would get very human-like answers.
For those who take psychology you should know this theory.
Thoughts are put to words. Words are heard. Then made into thoughts again.
What this means is that we don't catch the words perfectly, only the meaning. Then
we make our own words.
You may get a variety of answers from the AI and that makes it more human-like.
I would like for all of you to expand on this. Other things to consider to add onto the AI.
I know there are hundreds. Also I would like to hear some opinions.
I would like to discuss AI messaging. Like having a conversation with your AI.
We would first have to program the base knowledge for the AI such as greetings.
First we would have to program the AI to know when it is being greeted.
Depending on your greeting it would greet back.
You: Hi there.
AI: Hello.
The AI simply just checked for the word 'hi' seeing how it was the greeting. 'There' was just an extra.
So if you were to say.
You: Hi bob.
AI: Hello.
It would respond in the same way.
Now here is where it gets more complicated.
What if we also want the AI to check the other words.
In the above scenario we would want the AI to check if its name is Bob.
Although this would be possible with the 'get_string()' function it would
be extremely time consuming and you would make very inflexible AI.
Now, you also want the AI to learn. Well since it really can't learn, it just
stores information. That would be a little too perfect.
You wouldn't want this.
You: Remember this- Bill enjoys lots of candy.
You: Now, what did I tell you to remember?
AI: Bill enjoys lots of candy.
As much as you would want this.
You: Remember this- Bill enjoys lots of candy.
You: Now, what did I tell you to remember?
AI: Bill likes a lot of candy.
That would be difficult to program but it would make the AI
seem more human.
We would need to check for a lot of input. Here is some color-coding.
You: Remember this- Bill enjoys lots of candy.
You: Now, what did I tell you to remember?
AI: Bill likes a lot of candy.
The red section tells the AI that it needs to store information.
The green and blue sections are parts that the AI can change if it wants
to words with the same meaning.
Although choosing 'random()' isn't the way to go it would make it work.
You would get very human-like answers.
For those who take psychology you should know this theory.
Thoughts are put to words. Words are heard. Then made into thoughts again.
What this means is that we don't catch the words perfectly, only the meaning. Then
we make our own words.
You may get a variety of answers from the AI and that makes it more human-like.
I would like for all of you to expand on this. Other things to consider to add onto the AI.
I know there are hundreds. Also I would like to hear some opinions.
John(GmMkr-Admin)- Admin
- Number of posts : 501
Warning :
Reputation :
Registration date : 2008-07-17
Re: Artificial Intelligence Messaging
To be like true AI, the computer would have to learn. Of course, it's not actually going to be able to "learn", but we can use variables to add to the information the computer has at it's disposal. this could also effect sayings and pronouns etc.
Example 1:
Here the computer would recognise from the first statement that:
a)The user is talking about him or herself.
b)The user has a positive experience of the subject.
c)The subject is custard.
From the information given, the computer could evaluate that:
a)A new word has been learned (or saved.)
b)Not enough reliable sources have confirmed the information about custard.
c)Everyone asked has a positive experience of custard.
d)Custard is a yellow liquid.
e)Custard is edible/drinkable. (The computer presumes drinkable, due to liquid.)
Here the computer would recognise from the first statement that:
a)The user is talking about him or herself.
b)The user has a positive experience of the subject.
c)The subject is custard.
d)100% of the people asked like custard.
From the information given, the computer could evaluate that:
a)The info gained previously has been confirmed, making it reliable.
b)New info has been learned, which is not as reliable.
Player 3: What do you like to drink?
Computer: I enjoy drinking custard.
Player 3: What's that?
Computer: It's a yellow liquid, I think it's made from cornflour..
When asked about drinks, the computer responds by:
a)Selecting a drink that the majority of people like. In this way the computer is it's users.
b)Stating the confirmed info as fact.
c)Expressing an unsurity over the reliablity of the unconfirmed info.
Example 2:
User 1: ROFL!
Computer: What?
User 1: Rolling on floor laughing.
Computer: What does that mean.
User 1: I find it funny.
User 2: *insert correctly typed joke here*
Computer: Hahaha
User 3: *insert badly typed joke here*
Computer: ROFL
User 3: OMGz teh com is 1337.
Computer: lulz, i am teh 1337est!
Explanation coming later...
EDIT^2: Corrected grammar and spelling.
EDIT^3: I wanted to ask: Is there a way to analyse individual parts of strings, and if so how could it be done? I know in C++ you can save every word as a different variable, is that possible im GML?
EDIT^4: Cleaned up the first example and strted on another.
Example 1:
- Code:
Player 1: [color=blue]I like[/color] [color=red]custard[/color].
Computer: What is that?
Player 1: It's this [color=green]yellow liquid that tastes great.[/color]
Here the computer would recognise from the first statement that:
a)The user is talking about him or herself.
b)The user has a positive experience of the subject.
c)The subject is custard.
From the information given, the computer could evaluate that:
a)A new word has been learned (or saved.)
b)Not enough reliable sources have confirmed the information about custard.
c)Everyone asked has a positive experience of custard.
d)Custard is a yellow liquid.
e)Custard is edible/drinkable. (The computer presumes drinkable, due to liquid.)
- Code:
Player 2: I like [color=red]custard[/color]?
Computer: I've heard about it. What is it like?
Player 2: It's this [color=green]yellowy stuff made from cornflour[/color].
Here the computer would recognise from the first statement that:
a)The user is talking about him or herself.
b)The user has a positive experience of the subject.
c)The subject is custard.
d)100% of the people asked like custard.
From the information given, the computer could evaluate that:
a)The info gained previously has been confirmed, making it reliable.
b)New info has been learned, which is not as reliable.
Player 3: What do you like to drink?
Computer: I enjoy drinking custard.
Player 3: What's that?
Computer: It's a yellow liquid, I think it's made from cornflour..
When asked about drinks, the computer responds by:
a)Selecting a drink that the majority of people like. In this way the computer is it's users.
b)Stating the confirmed info as fact.
c)Expressing an unsurity over the reliablity of the unconfirmed info.
Example 2:
User 1: ROFL!
Computer: What?
User 1: Rolling on floor laughing.
Computer: What does that mean.
User 1: I find it funny.
User 2: *insert correctly typed joke here*
Computer: Hahaha
User 3: *insert badly typed joke here*
Computer: ROFL
User 3: OMGz teh com is 1337.
Computer: lulz, i am teh 1337est!
Explanation coming later...
EDIT^2: Corrected grammar and spelling.
EDIT^3: I wanted to ask: Is there a way to analyse individual parts of strings, and if so how could it be done? I know in C++ you can save every word as a different variable, is that possible im GML?
EDIT^4: Cleaned up the first example and strted on another.
Last edited by awasteoflife on Thu Jul 31, 2008 5:06 am; edited 1 time in total
awasteoflife- Junior Member
-
Number of posts : 174
Age : 30
Location : My house
Job/hobbies : Prohramming, Gaming and Paperrounds
Warning :
Reputation :
Registration date : 2008-07-25
Re: Artificial Intelligence Messaging
*bump* why is everyone posting in the other topic but ignoring this one
awasteoflife- Junior Member
-
Number of posts : 174
Age : 30
Location : My house
Job/hobbies : Prohramming, Gaming and Paperrounds
Warning :
Reputation :
Registration date : 2008-07-25
Re: Artificial Intelligence Messaging
It's do-able. Take the iGod for example: http://www.titane.ca/concordia/dfar251/igod/main.html
Not the smartest Artificial Intelligence, but it's pretty good. It sometimes wanders off topic if it doesn't recognize what you're saying.
But you have the right idea, but making it learn more words, phrases, and using them in the correct context.
Not the smartest Artificial Intelligence, but it's pretty good. It sometimes wanders off topic if it doesn't recognize what you're saying.
But you have the right idea, but making it learn more words, phrases, and using them in the correct context.
Re: Artificial Intelligence Messaging
That was funny!
It's possible to extract the words but it's very difficult.
I would need get_string(); and string_copy();
Both of which are very hard to use for this cause.
get_string(); Gets a WHOLE string.
and
string_copy(); Gets a piece of the string but at a fixed location.
It's possible to extract the words but it's very difficult.
I would need get_string(); and string_copy();
Both of which are very hard to use for this cause.
get_string(); Gets a WHOLE string.
and
string_copy(); Gets a piece of the string but at a fixed location.
John(GmMkr-Admin)- Admin
- Number of posts : 501
Warning :
Reputation :
Registration date : 2008-07-17
Re: Artificial Intelligence Messaging
John(GmMkr-Admin) wrote:That was funny!
It's possible to extract the words but it's very difficult.
I would need get_string(); and string_copy();
Both of which are very hard to use for this cause.
get_string(); Gets a WHOLE string.
and
string_copy(); Gets a piece of the string but at a fixed location.
Yes, I love messing with iGod. It's seriously absolutely hilarious.
I'll post a thread where people post their conversations with iGod.
Re: Artificial Intelligence Messaging
havnt heard of string copy....
how about, a textbox that when SPACE is pressed, saves a string. infact, if you could be bothered you could do this
if keyb9oard_check(vk_a)
{save the letter a somewhere}
and do that for EVERY letter
how about, a textbox that when SPACE is pressed, saves a string. infact, if you could be bothered you could do this
if keyb9oard_check(vk_a)
{save the letter a somewhere}
and do that for EVERY letter
awasteoflife- Junior Member
-
Number of posts : 174
Age : 30
Location : My house
Job/hobbies : Prohramming, Gaming and Paperrounds
Warning :
Reputation :
Registration date : 2008-07-25
Re: Artificial Intelligence Messaging
That's key-logging.
Although very possible to do, extremely difficult to check upon.
Oh and for future reference it's "keyboard_check(ord('A'))"
Strange right?
Although very possible to do, extremely difficult to check upon.
Oh and for future reference it's "keyboard_check(ord('A'))"
Strange right?
John(GmMkr-Admin)- Admin
- Number of posts : 501
Warning :
Reputation :
Registration date : 2008-07-17
Re: Artificial Intelligence Messaging
what do you mean? i use vk_key...
if my mouse was working i would construct a demo about how this could work...
if my mouse was working i would construct a demo about how this could work...
awasteoflife- Junior Member
-
Number of posts : 174
Age : 30
Location : My house
Job/hobbies : Prohramming, Gaming and Paperrounds
Warning :
Reputation :
Registration date : 2008-07-25
Re: Artificial Intelligence Messaging
Unless you have that extension, I can't use vk_a or anything.
Not a valid variable name.
Not a valid variable name.
John(GmMkr-Admin)- Admin
- Number of posts : 501
Warning :
Reputation :
Registration date : 2008-07-17
John(GmMkr-Admin)- Admin
- Number of posts : 501
Warning :
Reputation :
Registration date : 2008-07-17
Re: Artificial Intelligence Messaging
Game Maker Language 7
The seven is optional, I guess.
But I think the language varies from different versions of GM. Like GM6...
The seven is optional, I guess.
But I think the language varies from different versions of GM. Like GM6...
Re: Artificial Intelligence Messaging
I have GM7 Pro.
Also vk is used for keys such as space.
Also vk is used for keys such as space.
John(GmMkr-Admin)- Admin
- Number of posts : 501
Warning :
Reputation :
Registration date : 2008-07-17
Re: Artificial Intelligence Messaging
i thought you could use it for letters as well, i did just make up that bit of code...
awasteoflife- Junior Member
-
Number of posts : 174
Age : 30
Location : My house
Job/hobbies : Prohramming, Gaming and Paperrounds
Warning :
Reputation :
Registration date : 2008-07-25
Re: Artificial Intelligence Messaging
Oh, heh. Well no you need the ord().
So it would be:
keyboard_check(ord('A')), It also has to remain capital.
I don't know why.
Okay, back on topic. Any more additions to the AI Messaging?
So it would be:
keyboard_check(ord('A')), It also has to remain capital.
I don't know why.
Okay, back on topic. Any more additions to the AI Messaging?
John(GmMkr-Admin)- Admin
- Number of posts : 501
Warning :
Reputation :
Registration date : 2008-07-17
Re: Artificial Intelligence Messaging
are you actually considerring making this or is it more conceptual?
awasteoflife- Junior Member
-
Number of posts : 174
Age : 30
Location : My house
Job/hobbies : Prohramming, Gaming and Paperrounds
Warning :
Reputation :
Registration date : 2008-07-25
Re: Artificial Intelligence Messaging
It's conceptual.
I don't plan on making this but we can learn a lot
from talking about it.
I don't plan on making this but we can learn a lot
from talking about it.
John(GmMkr-Admin)- Admin
- Number of posts : 501
Warning :
Reputation :
Registration date : 2008-07-17
Re: Artificial Intelligence Messaging
You could probably learn way more if you actually created it.
That's how I learn things. From making websites to GML.
That's how I learn things. From making websites to GML.
Re: Artificial Intelligence Messaging
Yes, but we don't even have the basics down.
Maybe later.
I created something simple that when you said hi to it, it said hi back.
Any form of greeting you gave it, it answered back with another one.
Maybe later.
I created something simple that when you said hi to it, it said hi back.
Any form of greeting you gave it, it answered back with another one.
John(GmMkr-Admin)- Admin
- Number of posts : 501
Warning :
Reputation :
Registration date : 2008-07-17
Re: Artificial Intelligence Messaging
No.
You would say 'Hi'
and it might answer 'What's up?'
You would say 'Hi'
and it might answer 'What's up?'
John(GmMkr-Admin)- Admin
- Number of posts : 501
Warning :
Reputation :
Registration date : 2008-07-17
Re: Artificial Intelligence Messaging
I'd say, "CHILL3N LI3K A VILL4IN"
So, what happens if you say something like, "Nothing Much, How about you?"
So, what happens if you say something like, "Nothing Much, How about you?"
Re: Artificial Intelligence Messaging
It would i dentify NM then combine that with the question How about u? and say "Same" or something
hehehe, notice how th first two posts were really indepth and then everything else was just a couple of lines. im going to add some more stuff to my original post
hehehe, notice how th first two posts were really indepth and then everything else was just a couple of lines. im going to add some more stuff to my original post
awasteoflife- Junior Member
-
Number of posts : 174
Age : 30
Location : My house
Job/hobbies : Prohramming, Gaming and Paperrounds
Warning :
Reputation :
Registration date : 2008-07-25
Re: Artificial Intelligence Messaging
if its conceptual why is it in the GM forum???
do you by any chance program in C++ as well?
do you by any chance program in C++ as well?
awasteoflife- Junior Member
-
Number of posts : 174
Age : 30
Location : My house
Job/hobbies : Prohramming, Gaming and Paperrounds
Warning :
Reputation :
Registration date : 2008-07-25
Page 1 of 2 • 1, 2
Page 1 of 2
Permissions in this forum:
You cannot reply to topics in this forum