I'm going to do it

[Sterling]

Awesome! I'll be awake for another 30 minutes! Even though it's 2:00 AM, I love jogging a mile and a half around the block in the cold. So I'll be back soon, got to get some jogging done

Wow, it's like 4 here. Man, time flies when you're bullshitting your night on GBAtemp. xD I probably need to get some shut eye. My final move in date to my apartment is today. xD

 

[nryn99]

found it, it was supposed to be a blog for my technical writing class.

Spoiler

My AI Concept Part 1(First Insight)

First of all, this is hard for me to write because I have a hard time remembering what I am thinking in my free time so not all that I had thought in my head is written here. Second, what I have written may still be changed and will be called Second Insight and so on. And lastly, this is just my own opinion, therefore my ideas may be entirely wrong.

I think that the difference between human and AI is not by the human’s creativity, able to judge and to feel. The only difference is the human’s ability to learn from experience or translated for the AI, the ability to create a new line, statement or function from past events. Because it all begins here, humans as a baby, don’t know anything about the world, they only know that they must be fed and their development is slow at first because like in programming, creating software takes longer time than providing updates.
Now here’s how babies learn, they try to walk and when they do, they noticed that they moved forward to a different area. The more they walk, the more they will be accustomed to walking and they will now learn that to go somewhere, they must walk. For an AI, first it must be programmed to try to utilize anything around it including its own body, or what we call curiosity. The events that happened while trying to walk will be written somewhere, probably a .log file; this will serve as the AI’s observation. Next the AI will access the .log file for walking and analyzes it and tries it several more times to prove it’s theory and if proven true then the observation will become an experience and it will create its own code for walking.
Don’t misunderstood me by thinking that walking and moving its legs are the same, babies already know how to move its body parts from the beginning, thus the AI model must also be programmed to be able to move, but the “walk” must not be defined yet. Hence if they do not know how to walk then the AI will not be able to move forward but only roll around and move its limbs like a baby in a bed or maybe crawl, but they also need to learn that.

Hmmm, right now I forgot what I’m going to write next, it’s really hard for me to concentrate this much on one topic. I guess I must take a break and continue this some other time…

My AI Concept Part 2(First Insight)
Now for the continuation…

What I have said before sounds easy, but is difficult. Programming an AI to learn is how it will be able to create its own codes from its observations/experiences (.log), like a child learning that tripping hurts, or even a teen learning that love hurts. But a harder problem is how it will be able to make codes from external observations/experiences or from a person or an AI that is coded using a different programming language. If it cannot understand the language, it cannot create codes from outside data. Therefore the AI must also be able to recognize the pattern or rules of a foreign language.
My point here is coding for an AI to walk might not be that hard, but coding an AI to be able to create its own codes/ functions is harder, for it must be able to create or understand words so that it can associate objects, actions, and other things that is unknown to the AI. An example is a baby calling his dad “dada” and his mom “mama”. If a lot of people say “dada” the baby will associate the word to everyone, but if only dad says “dada” the baby will associate the dad (object) to “dada” (object name).
Thus developing the ability to name, speak, and eventually communicate.

i wrote this w/o much study about ai, so this is just simple ideas

 

[Zetta_x]

Awesome! I'll be awake for another 30 minutes! Even though it's 2:00 AM, I love jogging a mile and a half around the block in the cold. So I'll be back soon, got to get some jogging done

Wow, it's like 4 here. Man, time flies when you're bullshitting your night on GBAtemp. xD I probably need to get some shut eye. My final move in date to my apartment is today. xD


Yeah, I think my greatest weakness in my life is this girl. I'm really going to put myself through a lot of stress ignoring her for the next 45 hours. But I'm pretty sure if she loves me she will see the same thing, a life where I'm not there. Part of this blog is to convince myself not to give in to texting/talking/seeing her because it's going to ruin my experiment. I know within the first 36 hours she is going to get very depressed and text me with all the pain I'm making her go through. I just have to be strong enough to carry it for two days straight. On December 18th, 11:45 PM I will send her a text explaining some things and ask her if she has gone through as much pain as I have. From there we will figure things out.


QUOTE(game01 @ Dec 17 2010, 11:05 AM)

found it, it was supposed to be a blog for my technical writing class.

Spoiler

My AI Concept Part 1(First Insight)

First of all, this is hard for me to write because I have a hard time remembering what I am thinking in my free time so not all that I had thought in my head is written here. Second, what I have written may still be changed and will be called Second Insight and so on. And lastly, this is just my own opinion, therefore my ideas may be entirely wrong.

I think that the difference between human and AI is not by the human’s creativity, able to judge and to feel. The only difference is the human’s ability to learn from experience or translated for the AI, the ability to create a new line, statement or function from past events. Because it all begins here, humans as a baby, don’t know anything about the world, they only know that they must be fed and their development is slow at first because like in programming, creating software takes longer time than providing updates.
Now here’s how babies learn, they try to walk and when they do, they noticed that they moved forward to a different area. The more they walk, the more they will be accustomed to walking and they will now learn that to go somewhere, they must walk. For an AI, first it must be programmed to try to utilize anything around it including its own body, or what we call curiosity. The events that happened while trying to walk will be written somewhere, probably a .log file; this will serve as the AI’s observation. Next the AI will access the .log file for walking and analyzes it and tries it several more times to prove it’s theory and if proven true then the observation will become an experience and it will create its own code for walking.
Don’t misunderstood me by thinking that walking and moving its legs are the same, babies already know how to move its body parts from the beginning, thus the AI model must also be programmed to be able to move, but the “walk” must not be defined yet. Hence if they do not know how to walk then the AI will not be able to move forward but only roll around and move its limbs like a baby in a bed or maybe crawl, but they also need to learn that.

Hmmm, right now I forgot what I’m going to write next, it’s really hard for me to concentrate this much on one topic. I guess I must take a break and continue this some other time…

My AI Concept Part 2(First Insight)
Now for the continuation…

What I have said before sounds easy, but is difficult. Programming an AI to learn is how it will be able to create its own codes from its observations/experiences (.log), like a child learning that tripping hurts, or even a teen learning that love hurts. But a harder problem is how it will be able to make codes from external observations/experiences or from a person or an AI that is coded using a different programming language. If it cannot understand the language, it cannot create codes from outside data. Therefore the AI must also be able to recognize the pattern or rules of a foreign language.
My point here is coding for an AI to walk might not be that hard, but coding an AI to be able to create its own codes/ functions is harder, for it must be able to create or understand words so that it can associate objects, actions, and other things that is unknown to the AI. An example is a baby calling his dad “dada” and his mom “mama”. If a lot of people say “dada” the baby will associate the word to everyone, but if only dad says “dada” the baby will associate the dad (object) to “dada” (object name).
Thus developing the ability to name, speak, and eventually communicate.

i wrote this w/o much study about ai, so this is just simple ideas

Very interesting.

I agree that the future is heavily dependent on the past. We always consider ourselves that we can be random. However, if I asked to pick a number between 1 - 5, my theory is that the number we pick may seem random, however it has obviously had some greater significance then the other numbers which is why we picked it that roots the past.

I liked the idea of trial and error. A .log file is similar to our mind, we take are failed counter examples and we learn by them. This heavily relates to my statistical studies. The AI can gain lots of observations and as the number of observations increase it can make a stronger inference on a certain subject. Sort of like walking. Out of the 100 attempts to walk, 80 were successful in one step, 5 failed by curling the heel, 10 failed by trying to step twice in a row with one foot... therefore (by some statistical model) we can be 95% confident that the procedure that gained 80 successes is the best way to walk which can be something completely isomorphic to the human mind.

I think the best way to simulate an AI to represent a human is to make inferences. Unfortunately, a computer is pretty consistent. If you push a on a keyboard, a is going to appear on the screen. An AI can learn a language by first trying to communicate and looking at the reactions. From the reactions, it can decide weather or not it was a successful act of communication. For example, if an AI tried speaking and the speaker says... "wtf"... mark it down as error. Try again. Eventually with so many observations, you can make inferences and 'guess' with some level percent of confidence what the right move is

Those are my thoughts =)

 

[Wabsta]

Look at me, I'm on the internet and I can predict minds, I must be special and totally different from other people!

Grow up, you're, like the rest of us, just human.

 

[nryn99]

I think the best way to simulate an AI to represent a human is to make inferences. Unfortunately, a computer is pretty consistent. If you push a on a keyboard, a is going to appear on the screen. An AI can learn a language by first trying to communicate and looking at the reactions. From the reactions, it can decide weather or not it was a successful act of communication. For example, if an AI tried speaking and the speaker says... "wtf"... mark it down as error. Try again. Eventually with so many observations, you can make inferences and 'guess' with some level percent of confidence what the right move is

Those are my thoughts =)

while reading your post i don't remember more than half of what I've written. anyways my idea for communication is not like trial and error, instead it will ask you what it means and it adds that word to his "dictionary". its like if the ask you what does lol mean, you'll tell me it's "laugh out loud" now the ai will reference these three words in his "dictionary" and gets what it mean, he then adds the "lol" to his "dictionary" with the meaning laugh out loud. just like how we learn. i ask you "what's lol?" you say "its laugh out loud" me: "ohh, so its an acronym for laugh out loud. thanks"

 

[Zetta_x]

Look at me, I'm on the internet and I can predict minds, I must be special and totally different from other people!

Grow up, you're, like the rest of us, just human.


Yay, a prime example why I have so much hatred towards society.

It's funny, because you act like you know what it is like being human. You pretend humans have a certain way of living. Based off that one sentence I can tell how much you plagued the concept of what it is like being human. Have you ever considered what it is like being human years of civilized existence? It would be nothing like what you say I have to "grow up" to. As far as I'm concerned, if I need to believe in society's biases to grow up, then I'd rather stay a MFC (Mother F***ing Child)

QUOTE(game01 @ Dec 17 2010, 11:38 AM) while reading your post i don't remember more than half of what I've written. anyways my idea for communication is not like trial and error, instead it will ask you what it means and it adds that word to his "dictionary". its like if the ask you what does lol mean, you'll tell me it's "laugh out loud" now the ai will reference these three words in his "dictionary" and gets what it mean, he then adds the "lol" to his "dictionary" with the meaning laugh out loud. just like how we learn. i ask you "what's lol?" you say "its laugh out loud" me: "ohh, so its an acronym for laugh out loud. thanks"

Hmm, I see. It looks like your trying to create an AI that is near isomorphic to a way a human learns from its past. It looks like you are heading in the right direction, you learn as much as you possibly can that a human would react and how the mind would perceive it then you try creating a map from our mind into a AI's mind with the same exact structure!

I'm heading to bed but feel free to post any other ideas, I know what its like coming up with stuff then after I'm done I kind of do away with it without getting it written in detail

 

[GH0ST]

Where is the baby ?

 

[nryn99]

Where is the baby ?

what baby?

 

[Y05h1]

Chiming in on this discussion about AI:
The reason humans can't at this point create proper AI is because of the physical complexity of the human brain. The number of neuron connections within the brain is staggering. Not only that, but rather than acting as simple logic ports, the neurons act more like processors. Even if every single computer in the world was used in an attempt to recreate a digital version of the human brain, it would fall significantly short of the goal. It's why there is so much potential in biocomputing (although currently the next advancement is quantum computing).

I agree with the comparisons of computers to the mind. Perhaps it's not really coincidence, but more of an intentional development when computers were first being invented. But it also relates to this thought on no thought truly being random. It's the same with computers, they cannot produce a random number, only pseudo-random ones, which are chosen by some sort of algorithm. And in the same way, the algorithm draws from different sources in its memory. Time is often a part of it. It can have things like "x number of inputs" to affect the random number.

The human mind really is very structured. I've also noticed it. I use the knowledge to manipulate people as well, with simple things like planting ideas, or playing with their emotions. More proof in that human minds are structured comes from MBTI personality assessments. Everyone seems to fall into one of the categories, and that fact goes to show that there are essentially 16 types of people. Of course it's possible to fall in between personalities, but the personality of a person can still be charted using percentages of how much they fall into a category. And once someone has been classified, it's easy to know nearly everything about them. The only thing making them unique is their memories, and the decisions they make using these memories (of course aspects like IQ come into play as well).

Something else that really interests me is the subconsciousness. It's a part of our minds we seemingly don't have control over. But I've come to notice how simple it is to train it, and use it. It can become apparent in something like a test. You are looking at a question, but you don't know how to go about solving it. Yet for some reason, you keep thinking the answer is C (lets say its a multiple choice question). And the reason will often be that your subconsciousness knows the answer. It's like it has a better connection to your memory than your conscious mind. And it's important to harness this.

Then dreams. They are our best way of seeing into our subconsciousness. The average dreamer does not have control of his or her dream, it's their subconsciousness doing all the work. But, as you were discussing about lucid dreams, they are a key component to exploring our subconsciousness consciously. And it's techniques like this that let us truly exploit the full capabilities of our minds.

Emotions. Learning what they are, where they stem from, will give you control over them. And having control over them will also maximize our potential in doing things. Usually positive emotions will be the greater benefit, and can help in making clearer decisions. But even emotions like anger can be useful, in, lets say, a fight. It can build up testosterone, adrenaline and whatnot, which will subdue pain and bring out the aggressiveness necessary to not hold back.

I hope I didn't write down too much... I'll keep the second half for further discussion

 

[GH0ST]

Where is the baby ?
what baby?

I had to break up with my girlfriend. As much as I loved her, this relationship was causing so much stress in our lives.

She took her baby and ran away...

QUOTE(Zetta_x @ Dec 17 2010, 09:31 AM)

My relationship with my girlfriend has been shaky from the get go.

We both like to mess with people's minds and in the 4-5 months we have been together...

 

[nryn99]

Spoiler

Where is the baby ?
what baby?

I had to break up with my girlfriend. As much as I loved her, this relationship was causing so much stress in our lives.

She took her baby and ran away...

QUOTE(Zetta_x @ Dec 17 2010, 09:31 AM)

My relationship with my girlfriend has been shaky from the get go.

We both like to mess with people's minds and in the 4-5 months we have been together...

oh didn't read that part. or i just forgot.

 

[exangel]

If any of you fellows talking about wet computing have interest in reading science fiction, I highly recommend the book "The Diamond Age" by Neal Stephenson. It's an amazing cyberpunk author's work reflecting on everything from social constructs and individual power to the use of extraordinarily bizarre turing machines as plot devices within the in-story "book" the main character reads. And all kinds of nanotech and metatech and lucid-dream level weirdness.

 

[nryn99]

The human mind really is very structured. I've also noticed it. I use the knowledge to manipulate people as well, with simple things like planting ideas, or playing with their emotions. More proof in that human minds are structured comes from MBTI personality assessments. Everyone seems to fall into one of the categories, and that fact goes to show that there are essentially 16 types of people. Of course it's possible to fall in between personalities, but the personality of a person can still be charted using percentages of how much they fall into a category. And once someone has been classified, it's easy to know nearly everything about them. The only thing making them unique is their memories, and the decisions they make using these memories (of course aspects like IQ come into play as well).

hmm... i think why one can tell nearly everything about a certain person because personality and experience is related, or personality is based on experience. for example[case a] a child who grew up in a violent environment, he to becomes violent. but[case b] what if he doesn't talk to much to other people in their area instead he just watches a peace loving channel, he will soon questions his surroundings.

damn... i lost my idea again

i remember now, that being said you can program your child to what you want him to be by exposing him to the right elements. people with different personalities have experienced more than just one nature and understood why is it like that.

 

[boombox]

ok...what the hell is wrong with you? nobody else thinks this guy is an absolute weirdo and a total FREAK to purposefully f**k up someone he supposedly (and in a perverse way) loves...i don't think she will contact you. you sound like an ass-hat to be quite honest...

 

[Kickstarts]

ok...what the hell is wrong with you? nobody else thinks this guy is an absolute weirdo and a total FREAK to purposefully f**k up someone he supposedly (and in a perverse way) loves...i don't think she will contact you. you sound like an ass-hat to be quite honest...

Actually, that's a good point.

 

[exangel]

The human mind really is very structured. I've also noticed it. I use the knowledge to manipulate people as well, with simple things like planting ideas, or playing with their emotions. More proof in that human minds are structured comes from MBTI personality assessments. Everyone seems to fall into one of the categories, and that fact goes to show that there are essentially 16 types of people. Of course it's possible to fall in between personalities, but the personality of a person can still be charted using percentages of how much they fall into a category. And once someone has been classified, it's easy to know nearly everything about them. The only thing making them unique is their memories, and the decisions they make using these memories (of course aspects like IQ come into play as well).

hmm... i think why one can tell nearly everything about a certain person because personality and experience is related, or personality is based on experience. for example[case a] a child who grew up in a violent environment, he to becomes violent. but[case b] what if he doesn't talk to much to other people in their area instead he just watches a peace loving channel, he will soon questions his surroundings.

damn... i lost my idea again



Even with the personality assessments those can't be used to know nearly everything about an individual long-term. The amount of life experience an individual gains after taking the test will cause the individual to integrate different "functions" into their hierarchy. Perhaps there's already typical progressions for those types, but my father was discussing this with me recently and he said:

My score: INFJ, score 78-100-25-11. That's 100% on insight. Classification: "counselor". Somewhat similar testing about 6 years ago with different descriptive classifications had me as mastermind-architect with a minor in healing.

Now, I know that the test he took 6 years ago was a slightly different one but it implies he made a pretty major shift from logical/rational thinking needing sensory validation to being a very intuitive, insightful, realizing-before-actually-understanding type. both ways he is still an introvert but he responds to things very differently now than he would have 6 years ago.

edit:

QUOTE(boombox @ Dec 17 2010, 12:16 PM)

ok...what the hell is wrong with you? nobody else thinks this guy is an absolute weirdo and a total FREAK to purposefully f**k up someone he supposedly (and in a perverse way) loves...i don't think she will contact you. you sound like an ass-hat to be quite honest...

Actually, that's a good point.

I think it's fucked up but I don't care about his relationship's "health" that much. It his age and his girlfriend's age they need to learn, and will learn a lot from this situation regardless of how they react to each other's decisions. If things work out in a mutually beneficial way, so much the better. If he winds up completely dominating her, he seems like the kind of person that does intend to relinquish that complete power out of actual love.

 

[Y05h1]

Nature vs. nurture. Both play a part. I still think that despite any environmental aspects, the underlying genetic components remain largely unaltered, and that it would be very easy to revert back to ones "original" personality. Of course should a persons personality change, the fact still remains: it can be classified. Meaning there isn't really as much room for individuality as we would like.

 

[ProtoKun7]

To be honest, I always relate the human mind similar to the structure of a computer.

I often do the same.

 

[Zetta_x]

Where is the baby ?

She is safe, her parents are really her baby's guardian angel. They won't let anything happen to her.

QUOTE(boombox @ Dec 17 2010, 12:16 PM) ok...what the hell is wrong with you? nobody else thinks this guy is an absolute weirdo and a total FREAK to purposefully f**k up someone he supposedly (and in a perverse way) loves...i don't think she will contact you. you sound like an ass-hat to be quite honest...

Lmao, no shit Sherlock. From a typical society person, I am one messed up person. I love messing with people's mind. I love trying to exploit the human weaknesses for personal gain. I have absolutely no moral boundaries. However, your basing everything on biases. You believe there is a certain way someone should act and since I'm not acting that way I must be an "ass-hat". When I first met her, I didn't hold back my thoughts and that is what she fell in love with in the first place. I have dated a number of girls, all of them really fail in comparison because she is the only one who accepts me for who I am and I don't feel like I have to wear a mask to hide who I am. The funny thing is, I'm not some obese guy in the basement who dwells on his agony. I am a pretty good looking person with a strong athletic body who is a leader at my university and I prey on the weak boo hoo.

It's two simple things, she is either going to try and go for me or she isn't. If she does, I will work with her. If she doesn't, then why bother anymore?

---

It's amazing. I used to be the nicest person before a girl broke my spirit a few years ago (high school drama lol). If anyone wants to know the update on this. I didn't text her and she showed up at my house completely broke down in emotions and she apologized. Like I said, I knew our love was going to pull through

 

[Sterling]

ok...what the hell is wrong with you? nobody else thinks this guy is an absolute weirdo and a total FREAK to purposefully f**k up someone he supposedly (and in a perverse way) loves...i don't think she will contact you. you sound like an ass-hat to be quite honest...

Actually, that's a good point.

Lol, I just ignored that. Honestly it's the internet. I kinda disapprove of the idea, but honestly it's kinda entertaining because I am not the one doing it for one, and two it's a very interesting subjet that he is going on about. heh.