New developments in AI

Is AI dead per-se? i am not hearing anything about AI these days; are scientists giving up on AI?

Please Register or Log in to view the hidden image!

that would be bad news

Rick

 

Log in or Sign up to hide all adverts.

actually they are giving up on artificial intelligence. the government won't give any funding to new AI theories because they think its a waste of time.

 

Log in or Sign up to hide all adverts.

That's too bad. I'm interested in AI, but haven't had time to seriously research it.

What is the problem with creating AI on a self-propagating level, I think with enough effort and time you can make a self-learning machine. Can anyone give a brief explanation?

 

Log in or Sign up to hide all adverts.

There are some interesting looking courses on MIT's open courseware site:

6.034 Artificial Intelligence
6.803 / 6.833 The Human Intelligence Enterprise

 

Well, to create anything even remotely as artificially intelligent as us bio-human-computers, would be something that would take us far longer than our evolution did. thats what I think anyway....

 

Longer then 4 billion years!! Are you sure??

 

In light of the question i did hear about a certain instance in AI, and I even have a little article i found on it earlier.

http://in.news.yahoo.com/061221/43/6ah9f.html
Antony to open defence artificial intelligence campus Friday

 

its about time they recognized the power in AI. and building one facility is a start.

 

-Orcot

lol, perhaps not

 

Absolutely there are new developments in AI. Nothing truly awe-inspiring yet, but progress is definitely being made.

How about a robot that dreams?
http://www.cosmosmagazine.com/node/857

 

Contrary to popular belief; robots do not dream of electric sheep.

 

automated cars? trucks? planes???? isn't that a new development in AI--machines that can do things at a human-level????????

 

-mackmack

However... I think we should define the difference between mere automation, and out right Artificial Intelligence.

 

this is going to be off the topic, but you got lots of topics here talking about religion and practically no topics on science.

second, no one seem to beleive anything anymore. it seems like people need 100% proof before any real invention is considered truthful. anything else is considered a hoax. if people are soo skeptical about anything then how are real inventions supposed to be created.

 

I am sure there are many engaged in AI research. While I do not think it will take billions of years, it is likely to be hundreds before we get a AI device with 20% of the capabilities of an adult human.

It will probably take a major break through in computer architecture. Systems using current basic designs with faster CPU’s, more memory, and more disk storage will not get us AI, even with a large number of parallel processors.

If you look into how Deep Blue plays chess, you will realize that it is not close to being an AI device, even though it beat the best human player (Gary Kasparov).

Perhaps the best approach to human level AI is to create a device with the capabilities of a new born human (or a fetus) and let it learn. Designing and manufacturing the equivalent of an adult brain might be the wrong basic idea.

BTW: The human brain is not even remotely similar to a modern day computer. Memory and program functions are not distinct in a human brain. There is nothing analogous to a data base in the human brain. Actually, there does not seem to be separate hardware & software functions. The organization seems to be analogous to a huge number of hyper links that somehow interact to produce consciousness, intelligent behavior, problem solving, memory functions, et cetera.

If you consider the problem of recognizing another human being, you will quickly realize that the human brain is fundamentally different from a computer.

• If you ask a human to identify a person from a photo, you will often get an almost immediate correct response of “I never saw that person.” Such a response seems to be as fast as the correct identification of a close relative or well known celebrity. You will sometimes get a correct identification which takes a noticeably longer time (Examples: Some minor actor in a 20 year old movie; A school mate from 25 years ago; A second cousin you meet briefly every 2-3 years).

The slowest response from a computer will be the null response for data not in the data base.

 

Microsoft's SQL Server 2005 uses Neural Networks in their business intelligence logic for data warehousing. I think a lot of business analysis software is using AI and neural nets for predictive analysis and data mining. I think researchers may have given up on a "concious" machine, but I don't think AI research is dead by any means.

 

This isn't technically AI but may well be a factor in the future:

http://www.nature.com/embor/journal/v4/n1/full/embor719.html

 

The biggest problem with defining what is actually "intelligient" is that we can see and understand what is going on inside the computer.

For instance, there's a program online called ALICEbot (apparently my account won't allow me to link so just google: alicebot). One could argue that this program is artificially intelligient. It reads text sent from a user, calculates a proper response, and (usually) returns a phrase that is relevant to the conversation. Simply put, it's a chat-bot.

However, since we know that it is running on a computer, and performing knowable calculations, it can easily be dismissed as a complicated calculator. Even if programs such as ALICE manage to become advanced enough to seem human, it can still be argued that they are only a simulation of what a human might be like. There's no gauge to determine if someone/thing is intelligient or not, so how can we expect to know whether or not it is artificially intelligient short of spontaneous generation of a program running on a system that can generate output based on given input.

 

Lomion: The Turing Test is an objective method for evaluating AI. So far, no computer has come close to passing it.

 

>> anything else is considered a hoax. if people are soo skeptical about anything >>

Yes the net and science forums have killed science. Who wants to talk science and be crucified ? really pathetic IMO. Yes y'all responsible in your own egotistical way.


About AI, (and I do not intend to respond anymore on this... my new creed)

To achieve AI a new form of mathematics has to be used (and yes I know this mathematics), it is math that allows all knowledge of a subject to be placed in one mathematically analysable number. The math y'all know is the image, it is not real. Real math is unique, as for instance is a prime number.... It is symbolic math, but totally obeys the rules of arithmetic. Computer maths such as concatination, comparisons, boolean logic etc are a total dead end.... no wonder AI is dead for the elite.

AI will remain dead, until you turn a new page in mathematics.