Human Level Artificial Intelligence (part2)

Human Level Artificial Intelligence (part2)

These are machines that can think like human beings. 3PO, R2D2, the Transformers, and data from star trek are just of the robots that fall into the HLAI category. In Artificial Intelligence, the ultimate goal to achieve is to make a machine that can think like a human. To instill intelligence into artifacts that can think and reason and to understand language like a human being is one of the original goals of scientists. After almost 50 years of research into this field no scientists has ever come close to building a HLAI program, the best intelligent program any computer scientist has ever created can't even achieve insect intelligence.



http://www.humanlevelartificialintelligence.com

human
level
artificial
intelligence
.com


Part of the problem is that human intelligence comes from the human brain. Scientists have been studying the brain to find out what it does. For the most part they accumulated incalculable data on the subject matter but not enough to understand the details of how the brain store, modify and retrieve information. No one in medicine can tell what the conscious really is. So, if they don't know what the conscious is then how the hell is anyone suppose to write the computer codes to human intelligence? To design a computer software the programmer must fully understand what he/she is building. Especially when we are dealing with something as complicated as the human brain.

There are hundreds of fields in Artificial Intelligence and all of them are aiming at "strong AI". Strong AI is a form of computer that can learn from past experiences -- it uses a bootstrapping process to retain old information so that it can use the old information to learn new information (to sum up strong AI it really means human intelligence). I have studied all the different fields and i can tell you from the data i gathered that all the fields in AI are leading to a dead end. They will never acheive intelligence at a human level. Neural Networks are a waste of time, language parsers are a waste of time, genetic programming is a waste of time, predicate calculus is a waste of time, planning programs are a wasted of time, bayes network is a waste of time, and so forth.

I have written and published three books on human level artificial intelligence. If you want to learn more about the technology go to my website. It will give you a basic idea about the technology. If you are still skeptical then purchase my books to get a detail idea about the invention. It would be preferable if you had a background on computer science or electrical engineering to understand the books but i try to write the books so that non-computer scientists can understand the technology too.

Not to mislead people, when i say i can build a machine with human intelligence i mean i can build a machine that is smarter than a baby but not as smart as a child. I would estimate that this machine will have intelligence similar to an 8 year old child. The robot can learn and understand language, follow orders, reason in a simple way such as addition and subtraction, and answer simple questions. It isn't capable of doing calculus or complex reasoning. One form of complex reasoning is a joke. The meaning of a joke comes from a very elaborate form of reasoning. If you told the robot a joke it won't understand the reason for the joke.

Another thing i would like to note is that a machine is only intelligent base on how fast the CPU is running and how much disk space is available. The best mainframes today aren't fast enough to run even baby intelligence. Even if someone was smart enough to build a human level artificial intelligence program as smart as an adult the hardware has to exist in order to run the program. This hardware won't be around for maybe 100-150 years. There is also the possiblity of a limit to how fast a computer can run or how much disk space is available for a computer. This means that robots can be intelligent only based on the laws of science.

Besides the human robot the components that make up the HLAI can be used to build:

1. cures for cancer, HIV, and all human diseases
2. technology to prolong human life
3. build automated cars/planes/trucks
4. protect America from terrorism

 

Log in or Sign up to hide all adverts.

If human level natural intelligence can't do any of that what makes you think that human level artificial intelligence would be able to?

-Dale

 

Log in or Sign up to hide all adverts.

SPAM. And a [post=55809]repost[/post].

 

Log in or Sign up to hide all adverts.

this is not a repost. i have built a new website and i have included additional information about HLAI. the website address is different from the free website in the first part.

i am presenting updated website and information about the technology.

 

The human conscious is something that took a lifetime to create. if you are 21 years old that conscious was created in 21 years. The conscious is made up of multiple things that work together and i try to outline all of them in my books.

one form of a conscious are lessons that is learned from society and we apply this lesson to solving problems. for example, if i wanted to cross the street, the instructions pop up:
"look left, look right, and make sure there are no cars before crossing the street".

This lesson of crossing the street isn't one lesson learned but many lessons that were "averaged out". the conscious is like the internet, it has webpages added and taken out every second. It keeps changing and modifying itself. The collective of the knowledge of the conscious works the same way. it forgets things and learns new things every second.

what exactly is stored in the conscious -- what kind of data is being stored in memory?
if you look at the internet there are many things being stored such as webpages, html codes, java codes, pictures, text, movies, etc. for the human level artificial intelligence program i store things called objects.

storing, modifying and retrieving objects is the most important foundation for my computer program. this part took in a full year just to design. it learns from past experiences and use that past experiences to learn new things. for example, human beings need to first learn their ABC's before learning how to read. Or if you are doing math you have to learn addition and subtraction first before learning calculus.

most of the important ideas are in my webpage. but the detailed things are stored in the books. when i say detail i mean detail. these are not algorithmns that i'm presenting-- they are specific computer codes

 

I guarantee that no human has data stored in "objects" in their mind. There are no lists, there are no first in first out, first in last out data retrieval programs, there are no streams of 1's and 0's flying across the neurons of our respective neural nets.

The problem with most amateur AI research is just this. PCs and Macs were not designed for sentient thought. The code and mechanics running them are optimized for computing numbers, not pattern recognition.

You figure out how to create true emotions in a computer and you will have an AI. Without emotions, the computer, no matter how powerful the code, has no motivation, and without motivation, will do nothing. So, you end up with a AI that sees no reason to be.

- KitNyx

 

AND!...the primary purpose of AI research is NOT the development of Strong AI...I am not sure what you are referring to, because very little more than speculation is directed at the deveopment of Strong AI (Which is the actual term used in the field for sentient AI's). Granted, there are many of us in the field who research it and TRY to come up with original solutions to mimic the abilities of the brain in either software or hardware (or even wetware, in the case of working with artificial neurons), however, we do this research in our own time after we complete our REAL jobs of programming AI applications. Some examples of these? An intuitive program in MS Vista that "learns" from the applications you usually open at said time of day and pre-loads them for you. Or, perhaps the programming of streetlights. These are the AI applications that most of us work on.

IF...you are lucky or unlucky to get on a Strong AI research project, you can expect to have it being paid for by the DARPA or if you are extremely lucky, some sort of private fund, either way...any process you make is the property of said funding agency.

Quote from your webpage,"language is very important because it brings order to chaos." Are you suggesting that thought is not possible without language? That is a powerful statement since language had to originally come from some where, and in isolated cases, it continues to be created in the lack of.

- KitNyx

 

I apologize for being so skeptical, do not take it as a personal attack on you, but you have hinted at having the knowledge or experience, or DIY to design and build a true strong AI, and as far as I am concerned, since you made the claim, the burden of proof falls upon you. So, I would appreciate a response, or responses on my comments...afterall, if you do have what you say you do, then this is very exciting...

In response to the paragraph above (quoting the post which initiated this thread), I have a few questions.

1. I take it by your veiled causic disregard of the accomplishments (or lack there of) of "scientists" that you do not like them much, but also, by the way you did not say "other scientists", I understand that you are not one of those "scientists". Am I correct?

2. So, is the data incalculable, or have they not collected enough to calculate whatever it is that they should be calculating? I ask because your sentence is contradictory.

3. Can anyone anywhere give us a definitive definition of consciousness? Can you? If so, I would be VERY interested in hearing it.

4. Very few of those working on strong AI are trying to recreate the human mind. They may model their efforts after the human mind, but they are not (as far as I know) trying to recreate humans. Knowing this, how is evidence of being unable to recreate the human mind proof that they cannot develop strong AI?

5. You state that we cannot create a strong AI because no one understands the human mind totally. Yet, you claim to have the know how to "program" one...does this mean you have a complete understanding of the human mind?

If you can answer those 5 questions, I will know to whom I am speaking and whether you deserve any more of my time. Also, I tried to preview your books...the self published ones that are linked to above. Are the contents so classified that you cannot even show the first couple of pages of text, or have you not written them yet?

- KitNyx

 

"Quote from your webpage,"language is very important because it brings order to chaos." Are you suggesting that thought is not possible without language? That is a powerful statement since language had to originally come from some where, and in isolated cases, it continues to be created in the lack of."

language is important to intelligence. without it we can't build on complex reasoning. language builds on itself to create complex thought. i'm not saying we can't be intleligent without langugage but we will lack the intelligence that we haave beciase of langauge.

and as for my books. i wrote the books becuase no one will believe me. i need proof. i claim in the book and i quote "i can build a robot with human intellignece". a robot that is comparable to an 8 year old child. if you llook at an 8 year old child they can answer questions, follolw orders, think for themselves etc. thats an impressive technology in itself.

this is not a working theory its the real computer codes and the real deal. funding on the other hand is something that is hard to find. you can't just go up to people and say blah blah and get funding. everything must be proven by scientist before any funding takes place. i'min the process of doing that right now.

 

Nothing need be proven by anybody. All you truly need to get funding is an intriquing story and a persuasive speaker. And of course an audience that has something to gain. *shrug*

- KitNyx

 

people have to understand that a human brain was created by someone. Those instructions there can't possible be created by random happenstance. I can prove that the brain was created by someone by creating a computer program that strings letters together randomly. If this program can string letters together and form a book. When i say book i mean a book that took a human to write, then that is proof that random things can create something complex. I can run this program forever and i guarantee that this program will not create a book that people can read. complexity builds on itself, the more complex something is the more intelligence is needed to build it.

the conscious is defined in my website. The computer codes to build the conscious is defined in my books. if you know how to build a conscious then that means you understand it.

The instructions that our creator put into the human brain is his/her interpretations and is the artwork of the designer. When i say that there are somethings that can't be replicated from the human brain, i don't mean that we can't replicate it, i mean that the system is too complex and we can't reverse engineer all the small things in that system.

 

Uh, that's not how it works, actually...

Just some points I've figured out:
1) Neuron cell is at the same time a RAM unit and a single processor unit (cell?)
2) To be self-conscious (be able to learn without new experience), you have to be able to see your thoughts. That is to say, theories must be "seen", entered as "new" sense data, then compared to other knowledge.
3) Think, hard, about how much data we assimilate daily. Or as children. A machine could never have this much data compressed in so short a time.
4) It would need to be able to imitate. This seems a very strong diffrence between machines and people. To see and compare.

 

"A machine could never have this much data compressed in so short a time." The point of my storage part of the program is to optimize storage and to compact data. This took me one long year just to come up with. the brain doesn't store all the data that is coming in, only the most important things. This is why we forget.

another thing is that the data that is stored in memory needs a very fast search algorithmn to find data. no search algorithmn today can do this. I made a search algorithmn so fast it almost comes close to "constant complexity". constant complexity is a search algorithmn that can find one data entry in infinite amount of data in "one try". in other words nothing goes faster than constant complexity. My computer program comes very close to this.

 

If you read the physics forums a lot you know there are various people making claims of perpetual motion devices and infinite energy, etc. I usually do not comment on these, as it is not my field. Computational sciences are however my field. What you claim here is equivalent. In fact is is an extraordinary claim, requiring extraordinary proof (since I already have a proof I can present which says it is impossible). Please do us and the world a favor, post your algorithm. I don't care what language or system.

-AntonK

 

Hash table?

 

hash table?? ah, hash table and binary trees have independent data. all the existing algorithmns is not used. Since describing everything is so long, all i'm gotta say is that all the data are interconnected (global)not like a neural net. this organizes the data so that retrieving it is fast. for those that don't know what constant complexity is, it is a search program that finds a data in a given amount of time. for example, if i was looking for one webpage on the internet with 8 billion webpages, constant complexity means i got the webpage in one try. if i was searching for one webpage on the internet with 90 trillion webpages i found that webpage in one try. This works for infinite possiblities. This is good for database entries because it can get the data fast as the data entries get larger.

 

Wow...the computer version of Chomsky's Universal Grammar...Uggg. I have to agree with AntonK; talk is cheap, show us some algorithms or code.

- KitNyx

 

i just updated my website to include pseudo codes from the AI program.

http://www.humanlevelartificialintelligence.com

 

I see absolutely nothing of value in that.

-AntonK

 

Mackmack, if you have truly thought through some of the problems of AI, it is my passion, I would like to explore some of your thought pathways.... nothing that is sacred to you, just general ideas.

>> a very sophisticated form of intelligence that will allow it to survive in a world with fixed rules >>

Yes I believe a ROM exists in LIFE.... it allows basic abilities, including the logic of the brain. The imperatives/directives of the ROM are IMO, outside our evaluation, so therefore only fixed artificial intelligence can be achieved by human beings.

However a very smart computer can be constructed, and an expert of experts would be very welcome to our civilisation, maybe to become civilised, but I can not see true (LIFE) AI as being possible, unless the ROM can be deciphered.
For good AI, a computer needs to be hooked up to sensors, so it can evaluate the real world, .....an imperative if it was to extend its internal world to correspond with reality.


>> How do we store data in memory?
How can we represent the meaning of language on a computer?
How do we make decisions?
How does forget have to do with data in memory?
How will the data look in memory?
How can we change the data in memory?
How can we use language to change data in memory?
How can we learn things? >>

all good points, but the biggest problem with data is referencing it... so that the package received contains all the data in one image...almost holistically

It appears the neurones in the brain are continually firing, producing an electric field structure which can be read by all brain membrane sensors.... as a hologram

To achieve that criteria would require a new type of computer ????????


I work in the concepts of
trees, forests
top hats on components
lists of options, ? random/ordered
and use mathematics exclusively, so a result contains all the information known on the subject, packaged in one number and a key to unlock it.