[Bragi]

Interesting example. Aici can't handle this yet. I'll work on this after the verbs are working.
I would consider this to be a substitution problem: when you say 'I have hands', the 10 fingers should be replaced by 2 hands, which can then in turn point to the fingers (what a strange sentence  )
The problem here is: when do you replace and when do you ask for more info. Let me illustrate with another example that broadens the scope a bit:
I have 8 tiers
a car has 4 tiers
i have cars
I can easily have some cars and some spare tiers somewhere. It's not so certain that, by 8 tiers, I meant those of the car. So it's probably better to ask some more info, like 'Are those 8 tiers the ones on the cars'.
That's where the 'components' from my original example can come back into play. In the 'how many hands to I have'  sequence, fingers and hands are both human bodyparts + the fingers are attached to the hands (a hand has 5 fingers), so, when both are encountered on a human, you can remove the fingers from the object without asking for more info. On the other hand, if one is the component of the other, it's probably best to ask if it can be replaced, otherwise, both relationships need to co-exist.
If you, or anyone else for that matter, know of any other examples that expose still other tricky bits, I'd love to hear about them.

[sunama]

Right now, I am not so advanced in my AI, where I am asking it yet, to work out stuff (eg. I have 2 hands, Each hand has 5 fingers, how many fingers have I got in total?).

Right now, I am concentrating on getting my AI to retrieve data. Once I have this reliably complete, then I shall move onto calculations.

I can give you plenty of examples involving calculations, but just not yet. When I come to carry out those calculations in my own AI, then I will post them up here and ask you to input those questions into your AI, and check the result.

Right now, my AI doesnt classify each single word as a verb, adjective, noun, etc as I feel this is a bit superfluous. This would require immense memory/hard disk/cpu power HOWEVER, it is something I will definitely be looking in to and seeing if by classifying each word in the way you have, will bring any extra functionality.

Right now, even without classifying each word (as you seem to be doing), I am able to get decent output and like I stated before, my AI will be able to answer the questions posted earlier.

It may be difficult, when my AI comes to answer the question:

I have 8 tiers
a car has 4 tiers
i have cars.
How many cars do I have?

...however, I am confident I will be able to get the correct answer from the above input.

I think the main reason why I am able to get these answers without massive changes to the program is because the underlying principles of my program are a lot more basic and simple.

From what I can see your program is either VERY VERY complicated or your explanation is making a simple program sound very complicated. Note: I am not criticising your explanations, as you don't have to explain anything to anybody on these forums...you are doing it as a favour to the forum readers. Of course, when it comes to getting funding, you will definitely have to be able to explain your product, quickly and without confusing the listener.

[Bragi]

Right now, I am concentrating on getting my AI to retrieve data. Once I have this reliably complete, then I shall move onto calculations.

Ok, I understand.

Right now, my AI doesnt classify each single word as a verb, adjective, noun, etc as I feel this is a bit superfluous.

How are you handling multiple 'meanings' of a word? For instance, 'mail' can be a verb or 'noun' and when noun, it can mean multiple things. There are very few english words that only have 1 meaning (even 'be' has multiple). I don't only classify in noun, verb,.. I classify for each single possible meaning.

This would require immense memory/hard disk/cpu power

This is true, might I suggest to keep an eye on this since it is the determining design factor of the core. It's not like you can keep all your data in memory and existing (relational?) database systems tend to be far to slow.

[sunama]

Multiple meanings of a word are not a problem. My AI doesnt try or attempt to understand a word. It only looks for patterns. Thats all.

It works using key words. Once it has constructed a sentence using key words, it then looks for previous input (ie something which the user has previously input, or it has read on the internet) and then constructs a sentence by putting in words between the key words. In this way, the sentence is constructed without the AI actually understand what it has written.

So, for example:

Question: What colour are my eyes
AI's preliminary answer: eyes,,green (my AI already does this)
AI now use word combination patterns to turn the key words into a proper sentence
AI's final answer: [Your] eyes [are] green.

My belief is that the above method will give similar output to what your AI will give, only it will use far less resources and give answers quicker.

The only problem is that if my AI is asked a complicated question such as, "What is the meaning of life." it will struggle. Your AI on the other hand, may well give a decent answer.

Using the above technique, my AI never needs to understand what it is outputting, just that the format/pattern of words which it is using is correct. It knows that it is correct, because that same pattern of words has been used before, by the user or on the internet.

By breaking down the English language and defining each and every word, your AI is definitely more advanced, however, my gut feeling is that my AI shall be able to give similar output to your's, using fewer resources. But only time will tell.

If I feel that my AI needs to break down the English language (in a similar way to yours), then I shall have to add extra code to allow for this. At this point in time though, I don't want to go down that route, due to the sheer amount of code which will be required and the extra memory/hard disk space the program will use.

Note: I'm not saying that your AI is bad in anyway. If I had 2 years to play with, I think I would take the route you have taken, but I want to produce my AI as fast as possible, which is why I want to use different method. For commercial reasons, the faster the program can be produced, the better.

Another point is that my AI already uses multiple algorithms to come up with answers to questions. I think your AI must also have this ability. In otherwords, if it comes up with an answer using your default algorithm, your AI should analyse that answer to see if it makes sense. If not, it can then use a different algorithm and see if a better answer can be found. I'm not sure if your AI incorporates this way of thinking, but it might be something you could look into.

I shall definitely keep an eye on this thread and we can definitely compare answers (ie. of what eachother's program gives, as an answer).

[sunama]

It's not like you can keep all your data in memory and existing (relational?) database systems tend to be far to slow.

What I'm doing is "caching" data in memory.

So, in some instances, I am repeatedly accessing a file on a hard disk. If the current code block is repeatedly accessing a particular file, then I load the entire file into memory and cache it. So, when the program needs information, rather than going back to the hard disk file, it simply reads it from the cached file, in memory. MUCH FASTER.

I also want to develop a technique, where my program can predict which files it will need. Once this is done, AI will load into memory the contents of the file, before it is required, thus improving its speed.

Caching is definitely the way to go, I think.

[Art]

With all due respect a car doesn't have tiers...it has TIRES.

A TIER is a level or layer while a TIRE is a round, rubber, inflatable
device for an auto.

Spelling is very important when constructing and training a bot.
While there are many words that sound alike and some that are
spelled the same but have different meanings it is difficult to
"teach" that information to a bot.

Good luck with that.

[Bragi]

With all due respect a car doesn't have tiers...it has TIRES.

My bad, spelling clearly has never been my major strength. I am not very good at it in Dutch, which is my native tongue, nor am I any good at it in English. I am even worse in French and German, but if needs be, I'll manage. How well do you spell in other languages?

While there are many words that sound alike and some that are
spelled the same but have different meanings it is difficult to
"teach" that information to a bot.

Good luck with that.

Other people's bots I have come across indeed don't appear to be able to handle this very well. Please note, this is not some future feature I would like to build in. This was one of the core design issues. AICI can already  do this, so thanks for the luck, but not needed  .
Bugger, I had a nice example of 'be a be', but that's not right, it's 'be a bee'. If you know a simple example of 2 words using the same spelling, let me know, cause my mind is currently blank.

[Bragi]

got it, let me make a small demo with 'mail' as in :
send a mail to x or print mail and  mail x or something similar.

or what about this:

I am blue -> feel down
my eyes are blue -> color.

This last example is already possible since no special verb.

[sunama]

got it, let me make a small demo with 'mail' as in :
send a mail to x or print mail and  mail x or something similar.

or what about this:

I am blue -> feel down
my eyes are blue -> color.

This last example is already possible since no special verb.

I think that to deal with "special cases", the AI will have to be given a "special case" file, which it searches through to look for input which contains any such special case.

Bear in mind that cases such as "I am blue" are very rare. Most people would simply say, "I feel sad" or similar.

My belief is that the AI must be able to process correctly 50% of all cases. Once we have this, then extra modules will need to written for the "special cases". In this way, with every future release, the figure of 50% will rise and move towards 100%.

I don't believe that you can achieve 100%, without writing code for the special cases.

[Bragi]

Give me a few days to show some results.

[lrh9]

Still interested in this, but I'm sorry I have not had much time to learn about it.

Seems like a dam has broken and I've been getting new ideas for code that I've wanted to try out.

Doing some coding today. I might have an alpha release of the agent tools ready.

I'll take a break after I've completed the alpha and take a deeper look at this.

(The break is well placed, because I'm going to have to do some tests and then refine the design and code to make it developer and user friendly. The result of refactoring design and code is joy. The effort is hell. Hopefully I won't be too embarrassed by what I have so far and will upload the source here.)

[Bragi]

The break is well placed

Are you a student? and if so, what courses are you taking?

[lrh9]

I'm not a student.

[sunama]

I'm not a student.

"Then what are you...A lost tourist?"

 

I couldn't resist. 

[Art]

Tourist...

@Bragi
Two words w/ same spelling...but pronounced differently...
There's :

Wind = air blowing / Wind = to encircle an object with string or to wind a mantle clock with a key.
Record = a vinyl disk containing music / Record = to make a permanent copy of ones writing, music, notes, etc.
Wound = a bodily injury to the skin and surrounding tissue / Wound = that which is or has been wrapped around another object.
Tear = a drop of saline from the corner of one's eye / Tear = to rip apart  a piece of paper.
Lead = a heavy metallic element / Lead = the guiding or directing of one or more individuals.
Read = To eyefully digest printer matter (my definition) / Read = a book or printer matter one has already seen and visually digested. Rhymes with RED.

There are many more as only these came to mind without resorting to a thesaurus.  I still contend that the real difficulty is getting the computer program to KNOW what a word is and then it's usage. A computer will never know what blue is other than a measure of visible light with a specific frequency on the spectrum and perhaps how it compares to other visible lights, etc. How it is used is another issue altogether. How difficult would it be to teach a computer to create poetry with all the nuances. Not just the simple moon, june, spoon words but using more diversity and even some near rhymes (except for silver, month and orange).

Let's face it, computers presently rely on recalling a series of words or phrases that are available to them instead of applying any knowledge of the word's actual meaning. More things to think on....

And a little luck was offered in good faith. If you don't need it...be that way....