Artificial Intelligence

I think the problem is that humans essentially have a massive database imbedded in their brains. It's a lot harder for a computer to hold and distinguish between the good data and the junk.

We'll look at my escaping the room situation. What material is the propane tank made of? Is the propane flammible? What's teh hammer made of? What happens if you throw the hammer into the lawnmower? How hard is the rope compared to your skin? If the rope is 1/8th of an inch thick, about how much weight do you think it could hold up? What's the softest material: steel door, drywall, or brick wall?

All that information is already imbedded in a huge database in your brain and can be accessed almost instantaneously. If you want to simulate a human brain, it needs all that information as well.

Damn, I can't answer half those things...lol (jk)

Either way, computers are only 50-ish years old, and human beings have a head start by atleast thousands of years.

Eventually there will be a commonly understood language for computers. An algorithm of data that is commonly understood by computers. It will hold all of the facts like how hard brick is, or how badly it'll hurt itself if it falls 10ft. eventually, it'll have all the facts in its mind.

I'm starting to not feel well at all, head is hurting badly.... This is where an Expert System comes in handy. They help doctors diagnose sicknesses, its basically a DBase of information... you type in a symptom, and it'll tell you what it thinks is wrong.

I have something to add to this conversation, and I'll keep it brief. What role does instict play in this? Perhaps we need machines with instince before we can make intelligent ones.

Before a computer can become intelligent, it will definatly need a huge database of stuff it's learned. But humans are no different in this respect. If you were to replace the mind of the AI with an infant, my personal bet would be that it still wouldn't be the first to escape the room. It has no experience with hammers, propane, drywall, etc.

The great thing about humans is that we have 5 senses with which to constantly gather data. A computer on the other hand, is usually only programed to use only one or two inputs that have data manually entered. AI based spam filters have a manual "This is spam." button which must be clicked for it to observe. AI based 20 questions (click the link and play a few rounds, it's really fun! ) has a person's thought and their description through the 20 questions as input.

I don't belive the biggest problem with AI will be getting the database to be "wandered" (propane burns, burning creates heat, hot things melt metal, this door is made of metal), but rather the database not growing large enough to be of much use. Secondary to that problem is storage space for the database, and then only third comes the problem of relational thinking (which I will admit probably is a pretty large problem ).

JigPu

Its not the DBase we're lacking. It should be able to create its own database. An infant may not have a DBase, but it makes one over the years of its developement.

Once the AI can create its own DBase, and figure out a way to store the information, even if it does not have huge arrays of 300gb HDs......Then I think that we'll be able to say we have created an intelligence.

"20 Questions is too busy for new players"

hmm i dont think AI is as much being "Intelligent" as it is having free will and understanding emotion...or is that being human..

i know honda is working hard on AI and so far they have th most advanced AIs as of now.

and in my opinion AI will be purley software, and really have nothing to do with more advanced hardware, im sure if someone makes an AI program that is truley intelligent itll run just as good on our rigs as it will on some 10 million dollar robot.

Honda has some nice things, but so does Sony(I think its Sony atleast).

AI will not be purely software, it will be mixed with hardware also, but it depends on what the AI is trying to accomplish.

I know that there is a person who made his AI, and it has some cameras on it, and hooked to the mouse and keyboard. It is able to play a computer game and not run into the walls and such, and can actually fight inth egame. One of the largest problems for that though is for it to differenciate between the game, and other things happening in the room. Things that our natural neural network takes care of for us.

JigPu said:

I once had to do a programming project when I was a freshman in high school, where I had to write an AI that could play tic-tac-toe. It was freaking hard, but it worked eventually (and it was even a real simple AI that only gathered a database of knowledge, checked out what you were doing, and determined if it had ever seen anything like what you are doing). Like you said, the real problem IMO is not the programmer, but the lack of knowledge of how learning and intelligence actually works.

Good luck learning the languages and with AI. It's an extremely interesting field of research, and likely to eventually gain some more "real world" use

JigPu

Click to expand...

LOL my final programming project was PONG using java. I decided to do vs the COmputer...lol my ai...paddle follows the y coordinate of the ball at a certain speed.,. WOOT some ai

Zoltan

Foxie3a said:

Its not the DBase we're lacking. It should be able to create its own database. An infant may not have a DBase, but it makes one over the years of its developement.

Once the AI can create its own DBase, and figure out a way to store the information, even if it does not have huge arrays of 300gb HDs......Then I think that we'll be able to say we have created an intelligence.

"20 Questions is too busy for new players"

Click to expand...

But, humans don't create their own DB's. They rely on what they've been taught about how things work, and how things should be since the day they're born. AI is no different. You tell it what the rules are, and give it as much input, as you would a child as it he/she grows up. It will work within the rules it's been given. Shelter a child enough and they will react the same way computer AI does. You tell them from day one that 2+2=5 and they will belive that to be the truth. Teach them never to question it (or never give them the option to), and no one will be able to tell them any different. The only limitations to computer AI are the ones we impose upon it. It only learns as fast as we can teach it to by rules we've given it (and ourselves consequentially), just like a child. It's input relies on how fast people code to impose rules and restrictions on it. Wanna see the prefect example? 2 times 2 is what? 4 right? Easy enough...
...don't use a calculator...
...just do it in your head...

2345632 times 98 equals what?

Couldn't do it in your head huh? (neither could I)

answer is 229871936.

Kinda funny how after years of learning arithmetic, and going to school for 20+ years, and we can't just spit out the answer. Too many digits? Not a fair test? Ok.

1000000 times 10 equals what?

10000000 right?

Yeah, but garanteed 99.5% of you cheated, and just added a 0 onto the end of it. Why? Beacause you've been PROGRAMMED to do it that way. 0.4% of you sort of did the calculation in your head, but more than likely repeated it out of something you learned years ago. "10 X 1= 10, 10 X 10= 100, 10 X 100= 1000, etc...". Once again, programmed into your head. Someone programmed a computer to do that calculation a long time ago, and I'll bet the computer didn't take any shortcuts.

hmm i dont think AI is as much being "Intelligent" as it is having free will and understanding emotion...or is that being human..

Click to expand...

There's plenty of humans that have been programmed into beliving they have no free will. The emotion part. Well, it's only a matter of time before someone figures out how to code to take that into account.

well, the current problem is not the creation of databases(modern AI can already do this), it is the association(semantics) of the data. For example, an AI might know what the word "what" means; however, without appropriate semantics, it will not be able to properly use it in a sentence. The major problem with AI is decision making, as a computer can tell you whether something is right or wrong, and even why, but cannot give you an answer in the gray area(for example, abortion, it couldnt say that depending on the circumstances it is great, but is illogical otherwise).

ok, then humans and AI can both be programmed.... so where do we seperate AI from "stupid" code?

ok, then humans and AI can both be programmed.... so where do we seperate AI from "stupid" code?

Click to expand...

Evolution my dear Watson, Evolution.

Evolutionary programming(by genetics progamming) will probably be the key there. Say if you have an AI, and has to learn things. What if it learns that doing something that can kill it is good... it dies, and fails to spread it's stupidity on. Of course the idea of using an evolutionary algorithm with knowledge as a "genetic factor" could become rather interesting, the AI learns that being a mean SOB is the only way to be successful in this world.