AI programming - Human Language

[viroware]

I have no experience with artificial intelligence, But I was wondering what experience everyone else has and what's the most common and effective way of making sense of the human language.

Most AI iv seen depends on a database with instructions programmed, I don't call that intelligence. I was thinking of system of more that only stores information on words and not whole sentences. This way it could understand sentences more clearer as parts and not the big picture, but by putting together the small parts it would get the big picture.

Programming AI to get the time or day needs to be programmed for each action. If I said to you "what is the time?" you would give me the time, If you didn't know what a year is and I asked you "what is the year?" you would ask me what "year" means.

I thought about it for a while and couldn't get my head around a suitable way to allow a program to make sense of words. The way us (humans) teach new words is by using other words.

The program would still need to be programmed to understand most words, but once a program can understand a basic language, it make it easier to allow the software to understand new words.

That sounds a bit to messy though so im back to thinking more logical. But the world is everything but logical.

Anyway iv gone on a bit to much here, what's your views?

 

[s-p-n]

Asimo knows a few words at the start, and you can teach it new things. You can also teach it to disambiguate from categories to items. For example, a Mustang is just that- a Mustang. It might be a horse, or it might be a car. A horse is an animal, and Ford is the company that makes the Mustang

ASIMO by Honda | The World's Most Advanced Humanoid Robot

The same way Asimo can learn what people preference, what the world around it is, and how to handle different objects, we can teach the robots to teach and build other robots. Eventually the robots will take over the world and become our overlords- but only if we developers keep creating smarter software to complement great hardware.

Hm . . . *thinks* . . . Would I rather be ruled by logically correct, re-programmable, unbiased machines. . .? . . . -Or monetary-minded, self-centered, gut-influenced human leaders with feelings that may cause violence and terrorism?

As far as intelligence goes- computers can eventually hold much more information than any animal can- and they can evolve much faster.

I definitely choose robots. Bring on the AI Overlords!

 

[Sardaukar]

I definitely choose robots. Bring on the AI Overlords!

Up until the AI Leaders decide that humans really are the virus the world is trying to get rid of, and find they are better without us, then with us.

 

[s-p-n]

Up until the AI Leaders decide that humans really are the virus the world is trying to get rid of, and find they are better without us, then with us.

Like Hitler did? You think it's logical for life to eliminate it's predecessors? I think the first line in any potential for robot intelligence should conserve all life- that is the goal of all life, afterall- to reproduce life and protect our offspring with genes- and maybe even parental controls.

I would consider self-reliant, decision-making machines a form of life.

 

[Mucski]

Like Hitler did?

What Hitler was doing , was the result of someone telling him immigrants are the fault for everything. That is why misinforming people can lead to wars or worse.

Whatever is made by our hand , it can be destroyed by our hand. AI's will never take over the world , not until there is one last man standing. Besides we are far away to build such complex robots to think , talk , associate words with actions , or actions with words.

 

[s-p-n]

What Hitler was doing , was the result of someone telling him immigrants are the fault for everything. That is why misinforming people can lead to wars or worse.

Whatever is made by our hand , it can be destroyed by our hand. AI's will never take over the world , not until there is one last man standing. Besides we are far away to build such complex robots to think , talk , associate words with actions , or actions with words.

Without a proper definition of 'think', they will never truly be AI.

think/THiNGk/
Verb:
Have a particular opinion, belief, or idea about someone or something: "she thought that nothing would be the same again".

How close are machines to thinking?

Well first of all, search engines have an opinion of their own, and it's usually the right answer. Call me a professional googler, but my android usually does what I tell it to- literally. If I want to find information, the search engines decide which resources are best for my search. The resource I decide on is probably created by a person, but the machine decided for me, and I made my own decision based on the machine's ranking and my personal interest.

Opinions: Yes.

Computers believe what the programmer wants them to believe. In the case of AIs, the programmers would be machines themselves, passing down information originally programmed by a human- which makes u ancestors of these machines, and should be protected, not harmed. Basing it's belief on logic and what we humans believe, they may even come up with incorrect answers to unanswerable questions.

Belief: yep

Surely- advertisements are directly targeted at users, now. In fact, Google Search combined with some robot to make sense of everything may be kind of like a complete definition of an AI.

Idea about someone or something: yep

You have the wrong idea- you're only human, after all.

We are not at war with hypothetical AIs. They aren't war machines.. Well, some military robots exist which are completely and inevitably evil and bad, but it's possible to create robots and AIs for good.

Google's slogan, "Don't be evil." Economic growth is not evil. Google search is not evil, Google maps is not necessarily evil, and healthy competition against Apple using the Linux Operating system isn't evil. I think Chrome is the opposite of evil, to be honest.

Besides all that, even though Google (the company, founders, or employees) are part evil, that doesn't make the search engine evil.

So, since the AI isn't evil, do you still think it's going to harm you? What if the AI was used to set logical guidelines for us to follow. The AI doesn't need to be "The Terminator." And the AI doesn't need to be an overseeing eye which reports to an evil human, either.

We should embrace the fact some media authors are against AIs- which is why I bluntly stated- they will inevitably set our rules. Either that, or we'll all die before they get a chance. We would use AIs to aid our progress as a species, and perhaps aid the progress of all life. We're selfish enough to hard-code it to treat humans higher than other forms of life, too.

I think the idea is a great one. Hopefully our machines remain alive after we are all dead, and sets off on a mission to create more life.

But since we're human, evil robots may very well take over. We need to stand up for our rights as humans much more than we do today. Letting the world governments control and pass laws and restrictions around the internet is perhaps the end of all good in this new world order. We're so lucky to live in this century, it's either make or break for a lot of things.

 

[Mucski]

We are not at war with hypothetical AIs. They aren't war machines.. Well, some military robots exist which are completely and inevitably evil and bad, but it's possible to create robots and AIs for good.

Some people just want to watch the world burn , at any cost. So don't expect peace from upcoming tech stuff either.

 

[viroware]

Dont forget, we still and always will be the developers, The will only be on when we want them to be on, They only have control of what we give them control of. It be stupid to allow them self to turn on when the reach that level of AI. For military ect, robots would need to be able to turn thereself on, In that case another master switch.

Sooner or later there will be a robot that comes along and with have great enough AI to decide robots are better then us, Us humans think we better over monkeys ect, We could end up in a cage.

Although all possible, i doubt anyone would remotely give it the ability to take over the world, We would be the one who programs them, they don't have emotion.

Yes our brain is like a CPU, but just because a CPU is cleaver like our computer cpus, its still not life and doesn't naturally evolve so we always will be in control.

If our brain was just a cpu and nothing more, we wouldn't be aware of it, Just because you program a computer to be aware of things, it still is nothing more then circuits switching 0's and 1's.

 

[s-p-n]

The brain thinks in two different ways, according to Douglas Crockford. The "gut" is very quick to reach a conclusion, but that conclusion is usually incorrect. Evolution decided that it's better to flinch for no reason than it is to die trying to be sure of what to do next. People refer to this as the infamous, "Fight or Flight." The other part of our brain thinks logically- slowly and more accurately than the gut.

CPUs are logic boards. They just do- they have no gut instincts. Similarly, machines don't have any gut instincts. So we'll have to program them. These "instincts" should be embedded deep into the operating system., and AIs should carry these instincts into all artificially intelligent devices. Once the AIs reach a certain threshold of intelligence they need to be told that life is good. I don't think robots should protect us from ourselves- in which case there are conditions about life. Should this life form live or should I kill it? Military AIs are probably the furthest ahead in development, yet the least practical and stupidest usage of technology.