[TheMikh28]

I actually read a long thread, which I think is yours. Its on another forum (F14 tomcat or similar). Is that yours?

Indeed it is.

Most of it (somewhat vaguely) details the dull day-to-day progress of theory behind the project moreso than actual "concrete" developments, but it's the closest it really gets to any sort of project blog or anything like that.

Wow there's a lot of developers here. If I'm keeping this straight, TheMikh28, Bragi, Irh9, sunama, TikaC, datahopa, c.j.jones all have ongoing projects. All these would make good additions to the Member's Blogs, even if the link goes to a local thread like in the case of Irhr9. Also there are artists like claude2 who would make a good addition, and he could link to his youtube page.

Developer?  I'm flattered. 

I agree. There are a lot of people who developing their own AI's and going in their own directions.

The problem is that lets say you have 100 developers. Lets say they all develop their own programs. Each of those 100 people will probably be working on things that are very similar, if not identical. Using this "individual approach", each AI will probably not go far. If all 100 people however, pooled their knowledge resources, then I feel we could all go a lot further and a lot faster. This is very reason why I believe collaboration is the key to producing something ground-breaking, special and something that would have a better chance of being commercially successful.

A lot of users are taking radically different approaches to their projects.  While the fact that developers' efforts are not focused means that there is no single rapidly advancing project from the pooled minds, it means that there are a lot of unique ideas being generated in these respective projects -- and a lot of projects have yielded highly impressive results in different areas.

I believe collaboration has the potential to bring about amazing results, but there's always the potential for conflicting interests with regard to the direction of the project (or aspects of the project), as well as for sporadic workers who may not contribute frequently due to a number of circumstances.  Even the best of the developers present in the community are sometimes on-and-off with their projects.

The problem I feel is that most developers don't want to openly discuss their ideas in case someone else steals their idea. I don't mind exchanging my ideas with other developers as I know that to implement each idea takes a heck of a lot of time, so if they want to rip my idea off, they will still have to do an immense amount of work to engineer it. I know this from experience as to even add a small feature to the program, takes many lines of code, a lot of time and masses of hard work.

I think a lot of users are hesitant to share their more significant ideas ::less:: because they think other users will steal them, but rather fear that their ideas will be stolen by profiteering interests (e.g., corporations) -- and if these interests really desire to rip off an idea, they may very well have the resources to do so far quicker and more efficiently than the average individual.  Then there are some others who would simply rather keep ideas to themselves so they may reserve the right to to profit from them in the future.

Perhaps a new section entitled, "AI Concept Discussion" or similar. We could then create threads which discuss in depth, the best ways to deal with specific capabilities.

Eg.
Language processing
text to speech techniques
user identification
etc

I agree, that would be a great addition to the forum.

Well Freddy I like the "Member's Projects" idea because I myself am interested in seeing how these projects proceed and develop. So having a page where they are easily accessible would be the ideal. So I agree Freddy, it is all about "the grand project" of the developer (no matter if it's just a website, etc). That's what interesting to many.

This idea could actually coincide with an AI Concept Discussion section as described, yet also linked to the particular project that they have been developed for/in. Then it's possible for others to implement good algorithms in a completely different way for their/other projects.  But the project producing the concept would get the credit. The only rules would be don't tell any of your technology you don't want in the open source, public domain. Open source projects might rise to the level of profit generating product, and the developers would split profits IAW code effort put in.

There are a lot of similarities in our pursuits yet with different methods, purposes, implementations, etc.
It would be interesting to see them all organized together, and have a section we could discuss their concepts in, by abstract capability, and in direct reference to a project implementation. It might be popular; hard to say.

That's another great suggestion.

[Bragi]

Generally, people don't give a rats ass about ideas, they care about money, power and attention. As long as your project has non of these, your save to say whatever you want, most people will not listen anyway. So, that has actually been my approach: I have been hiding my AI ideas in plane site, so people can simply ignore it, and I have a proven record.

@TrueAndroids: That actually makes sense. I had already been wondering if that had anything to do with your choice. I must agree, 'Android' was a strange name to pick by google. Though, perhaps it says something about the ambitions they have for the thing.

@sunama: I think you are going to have a hard time finding developers willing to simply abandon their own project to come and help with yours. Especially if you consider that most projects here already appear to be more evolved compared to what you are proposing. It might help if you explain the core concept of your AI. At least than people might have an idea of the direction your project can be heading towards. Now, there is only a dream and a vague promise of participation in the spoils, you want others to create for you.
Don't forget, a project is better of with 1 grand master willing to go to far for to long instead of a 100 monkeys who don't know a thread from a processor.
Furthermore, it doesn't help if the grand master of a project is already stating that it is to big. AI software is no knock-off like a browser, word processor or imaging tool. These are all known algorithms that people can study, understand and get into coding. AI isn't, every AI project is a new adventure with an uncertain destiny and future. Something like that needs 1 guy pulling the whole thing forward 1 step at a time.

[sunama]

Generally, people don't give a rats ass about ideas, they care about money, power and attention. As long as your project has non of these, your save to say whatever you want, most people will not listen anyway. So, that has actually been my approach: I have been hiding my AI ideas in plane site, so people can simply ignore it, and I have a proven record.

Do you have a website for your project/ideas. I am very keen to get/discuss ideas with people, so that I don't waste time developing my program in the wrong direction.

@sunama: I think you are going to have a hard time finding developers willing to simply abandon their own project to come and help with yours. Especially if you consider that most projects here already appear to be more evolved compared to what you are proposing. It might help if you explain the core concept of your AI. At least than people might have an idea of the direction your project can be heading towards. Now, there is only a dream and a vague promise of participation in the spoils, you want others to create for you.

Recruiting for an "out there" project, with huge ambitions, is always going to be difficult. All inventors have this problem. However, the inventor has to keep pushing his idea through and if the idea is good and luck goes his way, then the project can well take off. I won't give up though.

Don't forget, a project is better of with 1 grand master willing to go to far for to long instead of a 100 monkeys who don't know a thread from a processor.

I think thats a little unkind. Those monkeys, may well bring their own ideas, which may turn out to be better than the "grandmaster's".

Furthermore, it doesn't help if the grand master of a project is already stating that it is to big.

This an ambitious project and I'm making no excuses in saying that if 1 person developed it, it would take years. If other people come on board, the project development will accelerate. To me, project development speed is very important.

AI isn't, every AI project is a new adventure with an uncertain destiny and future. Something like that needs 1 guy pulling the whole thing forward 1 step at a time.

I disagree. Unless that 1 guy is the most intelligent man in the world, he needs other ideas. He needs other people questionning whether his idea/algorithm is the right way to go about things or whether a different (potentially better) algorithm could be developed. I strongly beleive that to produce ground breaking software, you need more than 1 mind/person and collaboration.

[sunama]

I have been hiding my AI ideas in plane site, so people can simply ignore it, and I have a proven record.

I'm reading your blog right now, while watching the Turkish GP.

[Freddy]

Decided it might be best at this point to start a fresh thread on the new features for the forum so we don't over complicate this thread and we can leave this thread to be about Sunama's project :

Click here for the new features discussion...

[GT40]

Poor Webber... 

[sunama]

Right gents.
I need help.

I am currently in the process of loading my AI with some general facts about the world and life.

At present I have created the following list of general facts:
https://docs.google.com/document/d/1EavRjE57QuHkNYh81ojxhWIcrqwzAVL27FM6ZJK2yDo/edit?authkey=CMbLyfwF&hl=en_GB#

I need people to offer up some more general facts (keep them short and simple).
Should people wish to add to my current list of general facts, then please do so by replying to this thread with the new general facts.

Once I read them, I shall add them into the google document.

Thanks guys.

[DaveMorton]

Hey, Sunama. Nice list, but it needs a bit of correction.

Some of the facts you have that are incorrect:

Omnivores do not eat meat.

Herbivores don't eat meat.
Omnivores and Carnivores both eat meat.
Omnivores also eat plants, in addition to meat.
Humans are Omnivores.

Trees have leaves.

Most trees have leaves.
Some trees have needles.
Deciduous trees have leaves.
Coniferous trees have needles.

John Fitzgerald Kennedy when he was 46 years old.

John Fitzgerald Kennedy was assassinated when he was 46 years old.

In the USA, colour is spelled color.
In the USA, aeroplane is spelled airplane.
In the USA, centres is spelled centers.
In the USA, Solicitors are salespeople.

(not correcting you here. Just adding facts. )

Barrack Obama became president in 2009.

Barrack Obama was elected President in 2009.
Barrack Obama became President of the USA on January 20th, 2010.

I hope this helps.

[sunama]

Many thanks for pointing out the errors. I have made the alterations, which are now shown in the google document.

[TheMikh28]

Hey, Sunama. Nice list, but it needs a bit of correction.

Some of the facts you have that are incorrect:

Omnivores do not eat meat.

Herbivores don't eat meat.
Omnivores and Carnivores both eat meat.
Omnivores also eat plants, in addition to meat.
Humans are Omnivores.

Trees have leaves.

Most trees have leaves.
Some trees have needles.
Deciduous trees have leaves.
Coniferous trees have needles.

John Fitzgerald Kennedy when he was 46 years old.

John Fitzgerald Kennedy was assassinated when he was 46 years old.

In the USA, colour is spelled color.
In the USA, aeroplane is spelled airplane.
In the USA, centres is spelled centers.
In the USA, Solicitors are salespeople.

(not correcting you here. Just adding facts. )

Barrack Obama became president in 2009.

Barrack Obama was elected President in 2009.
Barrack Obama became President of the USA on January 20th, 2010.

I hope this helps.

Aaaaaaaaactually Barack Obama was elected in 2008 (Nov 4) and inaugurated on January 20th, 2009.

With that said, developing some method of identifying alternate spellings and/or synonyms would be an excellent addition to the software.

[sunama]

LOL. Yes. I've made the alteration.

Obama is creating a big problem here.

Synonyms
As yet, I haven't developed a technique to "create" synonyms, but I have some idea. This is not a high priority for me though.

The most important thing at present is to get the AICore to be able to identify general facts from sentences, automatically, by itself, without assistance.

The more facts that I enter manually, the higher the success rate of the AICore will be, to identify general facts, by itself.

[DaveMorton]

ok, no more posting for Dave while on pain meds.

[ivanv]

long time ago, but maybe someone finds this post usable...

I have already found a way for AI to relate the word 'colour' with 'blue', so that I get the correct answer. However, I need to find other ways/methods/algorithms to find the associations. Does anybody know of any sources of information which may give me ideas on how to associate words with other words?

Let me brainstorm here:
Words can be divided to subject and predicates, etc. Predicates take any number of parameters (2 - 3 in spoken languages). Some subjects are sets of other subjects. These sets can be identified by analyzing instances of particular parameter of some predicate.

ie.
blue is color
yellow is color
-------------------------
conclusion: predicate "is" relates subject color to subjects {blue, yellow}

Brainstorming further, predicate "is" mostly represents set->element or set->subset relation. Predicate "is" also represents syllogism. Other predicates mostly don't. Other predicates also can form a grouping of their's instances of particular parameter.

If anyone is interested, I took a brainstorm base from this short video about knowledge structures. Represented KB concept lacks dynamic dimension, but here it is:

i.

[victorshulist]

Words can be divided to subject and predicates, etc.

Predicates take any number of parameters (2 - 3 in spoken languages).

"Words can be divided to subject and predicates, etc. " 
    no, rather.....
"Sentences can be divided to subject and predicates, etc.

ie.
blue is color
yellow is color

You should realize that these statements are an abbreviation.    What you really mean is:

"blue is an example of a color"
"yellow is an example of a color"

simply stating "blue is color" would mean that literally the word "color" can be subsituted for "blue".    Which would mean you could convert:

"What color is your car?"   ,   (if  "blue is color" were true), then this would mean this question would be the same...

"What blue is your car?"    dosn't make sense, thus  "blue is color" is FALSE.

-------------------------

conclusion: predicate "is" relates subject color to subjects {blue, yellow}

FYI:  "is"  is the verb inside the predicate.   The predicate is "is color" (or, should be "is an example of a color").

Also, "blue" and "yellow" are NOT subjects, they are the direct compliment nouns of the verb "is",  (and "is blue" and "is yellow" are the complete predicate).

Sentence :=  subject + predicate,   or, if you prefer

Sentence :=  subject + predicate.verb + predicate.direct-objects

there are many, many other possiblities though 

Brainstorming further, predicate "is" mostly represents set->element or set->subset relation. Predicate "is" also represents syllogism. Other predicates mostly don't. Other predicates also can form a grouping of their's instances of particular parameter.

1. the word "is" by itself, tells you nothing really.   Only by considering the associated information of the subject words and the direct compliment words can you infer what the 'is' means.    That is, without considering any associated information that X has or that Y has,  "(X) is (Y)" will get you no where.      Also, don't ignore words like "the" and "a".    Even the word "a" (indefinite article) can give vital clues.   "(X) is a (Y)" means something quite different even than "(X) is the (Y)"  or "(X) is (Y)".     With "(X) is a (Y)" this suggests that  (Y) is a named collection, of which (X) is one member, among others.

2. "their's" --->> "their"  (don't need  apostrohe s)

[ivanv]

yeah, yeah, very funny ...... to take a laugh over bits of bytes... you could say also "why do we need true and false if we got boolean"... 
alright, i admit the subject-object-complement-something error for generalizing these notions under term "subject" , but i ask for little bit of collaboration from You  . please excuse my piggin english?

  well, i'm talking about recognizing list of notions described by words being elements of some set:
1. recognize predicate
2. enumerate combinations of words for each parameter of predicate (correcting -> 0, 1, 2 or 3 parameters  )
3. to use OOP terms, enumerated words could be objects in properties of some class (if predicate is a class)

by the way, do you know that predicate and higher order logic has only predicates and variables (variables represent subj-obj-compl...), yet it's so expressive.

i could talk forever about things such is "parameter of a predicate can be considered to be a predicate also", but i'll save it for my justaboutrightnowreadytobepublished (meanning never) AI site ...koff...kof...anyone...kov...interested?