[Freddy]

I'm glad it's going well. I donated too but have not caught up with the project for a while. It appears to be taking longer than expected, but I suppose you cannot rush the creative process.

[ranch vermin]

Ivan, your ideas are good.  How far away from actually coding a system yourself are you?   Because thats the proof of your idea,  without it,  people wont be so interested to hear your ideas... and even if you do, still hardly guarantees it,  but its definitely a more professional thing to do.   To have the proof of your idea, no matter if people are interested or not, its better.

[ivan.moony]

@ranch vermin
Sorry, I didn't notice the question before...

How far away from actually coding a system yourself are you?

I'm not yet finished as to general AI representation algorithm. There are some ideas, but I'm still working on it. I'm having some thoughts on copying outer behavior and analyzing input for general AI behavior rules.

The plan is to firstly build a programming language that can hold AI thoughts, then to implement AI algorithm (explained here) in it. A parsing part of programming language is finished for a while now (parser.moonyweb.com - UI doesn't work on internet explorer yet, I'll fix this soon). I plan to step to interpreter part very soon if it turns out that ideas presented here can be held by programming language concept.

But even then, when the language is done, I'm at the beginning of coding AI, maybe it will take ages (or never) for me to finish it. At least I'll have a finished language which might give an inspiration to others (closer real world systems description mechanism than in, say Java or C++).

So, basically, I'm just hanging around and brainstorming and I don't expect that will end up soon.

[DemonRaven]

How is his project coming along? I tried to find some newer info but struck out so if you know of any place let me know.

[Don Patrick]

You've been brainstorming for half a year...

[ivan.moony]

I'm sorry Don Patric, I didn't realize I am getting annoying...

[Art]

It's OK to keep tossing the ideas around but sooner or later one needs to start building upon the ideas to fabricate or fashion the creation.

If all I do is talk about building a nice wooden wine rack and spend all my time exploring all the possibilities (number of shelves, wood types, height, width, drawers or not, top or open, etc.), it might never get built. Of course, if it was something my wife wanted, it would have had to have been completed yesterday! Sort of an analogy.

[Freddy]

I'm on Ivan's side here. It takes a long time to work out this kind of complex thing. Some people have been working on it for years. Ivan has already posted some papers and other reading and that takes time to do too.

Ivan, keep at it, I think you should take all the time you need and you are definitely not being annoying to me.

[Art]

I certainly understand Freddy but on the physical side of things, I have a friend who, a few years ago, said, "I'm going to get a computer!"

When I asked him why he hasn't yet purchased one, he replied, "With all the upgrades and updates, I'm waiting until they get it right! Then I'll get one!"

Having heard this, I naturally assumed that he will most likely never get a computer!

Analogy again but sooner or later one has to produce. Wait too long and your idea will be old when it's released. I'm just sayin....

[ivan.moony]

Please don't argue because of me, I hate to produce negative emotions. Maybe I should just stick to milestone papers without intermediate brainstorming, that would make me more serious.

Edit:
The reason why I didn't produce anything yet is simple: I don't have all the answers. Brainstorming gets me fresh ideas, but I can do it in private, no harm done.

[Don Patrick]

No worries, I can always choose not to read, but my advice to you is that programming the occasional prototype function can be very insightful, because you just can't think of everything beforehand. I've seen people theorise left and right for years and still not have built anything, because without experimentation and testing it's hard to determine which direction is promising and which direction is a dead end.

[Art]

Thank you for putting that sentiment in words. I certainly do not, for one minute believe we were arguing. Rational discussion of an idea, product or theory is the hallmark of this forum and most any forum worth its salt. We are mostly adults here and as such we must respectfully agree to disagree on viewpoints, topics, ideas of theories. This exchange is what makes up human. The ultimate goal is for each of us to walk away with a refreshed perspective or some new ideas or knowledge. Sharing ideas is productive and just because we differ in perceptions, does not make us enemies!!

Post your ideas and sound off. If we have any of our own ideas or possible improvements, we should likewise post them. It IS a forum! Be Friendly, Be Fair!
 

[ivan.moony]

If someone reads what I post here in this thread, I have a good feeling about copying human behavior as an AI behavior process.

If computer says A and a human responds B, then a computer will remember it and it would respond B whenever human says A. This, of course, should have a deeper meaning where B is a function of A (written B = f(A)). This replaces esoteric "intelligence detector" from initial Sky algorithm by more formal analyzing of action and response from textual terminal. Function f would be guessed and checked systematically by a genetic algorithm.

Observing further, human will say "A->B" in one moment (in his own words) and after, when computer says A, human will respond B. In that moment computer will learn the meaning of "->" and what to do when its left side raises in conversation  (to respond the right side). This being learnt, when a human writes "C->D" and says C afterwards, computer would respond D. I think that this could be called higher order reasoning. With higher order reasoning we can upload a whole book of knowledge with a lot of X->Y sentences, making a fast learner after initial face to face raising of AI.

So, what when we have a criminal human saying a criminal deed B (say, steal) on mentioning A? We will also have a victim saying C (say, don't steal) on mentioning A. Then, when A is said to computer, we have answers B and C in contradiction on which we can apply Sky conflict resolver which finds a solution D, satisfying both criminal and victim, thus avoiding the criminal output.

The more people that computer interfaces and the more knowledge it collects, the safer response will be. The best scenario is where an AI is being taught over web site by a lot of people to be safe enough. The worst scenario is when an AI is being taught by a single criminal, fed by wrong books. Maybe single teacher scenario should be forbidden by the law itself when using behavior copying algorithm. On the other side, when passing a whole decent book into computer, AI learns a lot of responses which would also make it safe. So, it is possible that AI, being taught by a criminal, grows up a good personality by reading right books.

Anyway,  happens with a criminal teaching AI and passing to it mean-thought books. After that, one right book and it turns to nice personality. Maybe the only acceptable solution with copying behavior is a crowdsourced-site taught AI.

[Ultron]

If computer says A and a human responds B, then a computer will remember it and it would respond B whenever human says A. This, of course, should have a deeper meaning where B is a function of A (written B = f(A)). This replaces esoteric "intelligence detector" from initial Sky algorithm by more formal analyzing of action and response from textual terminal. Function f would be guessed and checked systematically by a genetic algorithm.

Observing further, human will say "A->B" in one moment (in his own words) and after, when computer says A, human will respond B. In that moment computer will learn the meaning of "->" and what to do when its left side raises in conversation  (to respond the right side). This being learnt, when a human writes "C->D" and says C afterwards, computer would respond D. I think that this could be called higher order reasoning. With higher order reasoning we can upload a whole book of knowledge with a lot of X->Y sentences, making a fast learner after initial face to face raising of AI.

I have suggested something like this as well - except in different words. I refer to A and B as patterns (N, N, A, V, N...), where A is input and B (could be B1,B2,B3...) is a suggested output pattern (a form) which is later filled with adequate words.


I have based this upon the notice of human showing similar behavior when learning how to respond to a given situation. There are many unwritten rules in linguistics, and this is exactly how we learn them (fact, not theory). We mostly do it subconsciously. Additionally, we don't recognize that pattern (B) as a suggested response if it caused an undesired response - instead we remember it is a 'don't do it' response.


I am actually planning to implement this in v.02 of my alpha prototype by feeding the program grammatically correct texts, preferably conversations so it can learn from them. I would use a pocket voice recorder but:


A. Speech recognition sucks;
B. I don't live in England or the US;
C. It won't record well unless I hold it like a d*ldo while I talk.


Suggestions?

[Don Patrick]

I'm going to have to say no go on the voice input until software improves, even with training and a decent headset. Theatre plays and novels contain written conversation, distinguishable by quotation marks. http://www.gutenberg.org has an abundance of english language novels in digital formats.
Cleverbot learns question-answer pairs from online conversations, but that wouldn't get you grammatically decent responses.