[sunama]

Strictly speaking, a lot of the posts in this thread should be posted in an "AI Concepts" thread. That way other people who developing their own AIs will be able to pick up some information that might prove useful to their own program, especially when it comes to testing.

I've copied 'the words which are pronounced the same', in the post above mine, in order to test my own AI, when I get to that point.

[Bragi]

Thanks for the examples Art. I had just started working round 'blue'. Stay tuned.

[Art]

Another of the problems would be words that sound the same when pronounced but have different meanings depending upon the context with which they were meant or given.

Such would be your blue = color and a feeling of sadness.
hose = a rubber or vinyl tube for delivering water or air or my computer got hosed with a virus
rub = a mixture of dry seasoning to add spice to various meats or rub a muscle when you get a leg cramp.
lime = color and a fruit as with lemon, orange, tangerine and avocado...etc.

There are simply so many words that fall into so many categories that it is one reason so many AI enthusiasts have a difficult time instructing the bot how to parse / handle words like these. Also how difficult it is for the bot to try to understand the context in which a word was meant.

[Freddy]

Interesting, there's also lime the building material and lime trees.  Must be hundreds of these.

Disambiguation !

[lrh9]

But how do we know which is which? Is context processing really that difficult? Or has someone not given proper consideration and time to the problem?

I don't know. But I do know that sometimes it depends on further interaction. I've got to be honest. If you were cooking meat and asked me to go buy the rub I'd have to ask you what the hell you were talking about. 

[sunama]

I've got to be honest. If you were cooking meat and asked me to go buy the rub I'd have to ask you what the hell you were talking about. 

LOL. My thoughts exactly.

[Bragi]

Art, were you thinking of something like this:

Notice I got the verbs working, near the end 


I think I'm going to work on the 'copy' example now, as in:

copy 'x' to 'y'
delete the copy

It's a bit similar to Art's 'record' example, but a bit closer to the pc world
This has some nice new challenges I haven't fully implemented or tested yet, like parsing file paths and the link back to .net: actually copying and deleting files.

[TrueAndroids]

very nice Bragi. That's enough evidence to say your AICI program is indeed thinking, and therefore an AI. If we use the networkdictionary.com definition of AI:

Artificial Intelligence (AI) - "can be seen as an attempt to model aspects of Human Thought on computers." From : www.networkdictionary.com

The thoughts here are classifying thoughts (genus / species). In your example, color(white, blue, red) and feeling(blue, happy). The thought answers the question "What <genus> is/are me/you?" And the thought is "I/you are <species>". And a question thought is also used to get the external information: "What is <species>?" And then a definitional thought is understood: "Species is genus." Keeping to this genus/species arrangement, this should be open-ended expandable. The user could then teach any genus/species arrangement and the AICI answer any g/s query (it has thoughts on).

** I also agree that this thread would be a good addition to the Concepts section Freddy!  

[Bragi]

Thanks TrueAndroids. Think about this for a second: I'm just getting started, these are the easy bits.

[sunama]

Those are good answers which your AI is coming back with. Very good.

[sunama]

Art, were you thinking of something like this:

Q1. What sort of duration is taking for you to get your AI to reply to your questions? Less than 1 second?
Can you give us some idea please?

Q2. Also, is your AI written in Prolog?

Q3. Can your AI relate words yet, in the follow sense:

user: my eyes are green.
user: what colour are my eyes?

Can you AI relate that green is a colour?

[TrueAndroids]

Art, were you thinking of something like this:

Q3. Can your AI relate words yet, in the follow sense:

user: my eyes are green.
user: what colour are my eyes?

Can you AI relate that green is a colour?

ANS>Q3 user: my eyes are green.
(*AICI either understands green from prior input, or asks PC:"what do green mean?" and user answers user: "That is a color."
User: What color are my eyes?
PC: Your eyes are green. (: as learned from prior, received external (user) knowledge as shown:)

Can you AI relate that green is a colour?

We see the AICI program do:
PC "What do red mean?"
User: "That {red} is a color."

So it could do the equivalent -
PC "What do green mean?"
User: "That {green} is a color."

**fixed somewhat, from below.

[Bragi]

Yep, TrueAndroids is getting it. If 'green' was known, it would simply store the data. If 'green' was unknown, it would ask what it is.
1 difference, to request for the info again, You'd have to ask, what color are my eyes.

[TrueAndroids]

Yep, TrueAndroids is getting it. If 'green' was known, it would simply store the data. If 'green' was unknown, it would ask what it is.
1 difference, to request for the info again,

You'd have to ask, what color are my eyes.

Ok I see, so generically it is "what <genus> is/are/am <object> (eg my color, my eyes)" (*where the correct answer is the assigned species by another thought)

Example: user: What mood (=feeling) is the boss in? PC:Blue.

[Bragi]

Yep, except it's fairly flexible in the word sequence (it parses it, a bit like a file path). You could also say something like:
My sister's color is blue or the brother of your wife is red