Artificial Intelligence Research

So, now that we can live longer, we need to advance in the field of computers. We do alot of work with our minds, why dont we make life easier for ourselves, and create machines that do the job for us. Sure, we can come up with solutions, but we have limitations. Neuron number, neuron speed, memory capacity, etc.

there are loads of ways to apply an AI in everyday situations. Most preferrably, chatbots, or general AI with logical reasoning.

But, creating something, that understands is an act of trying to become like God. You cannot create something that willl understand like humans. Understanding is ambiguous in it’s own sense.

un·der·stand/ˌəndərˈstand/
Verb:

Perceive the intended meaning of (words, a language, or speaker): “he could usually make himself understood”.
Perceive the significance, explanation, or cause of (something): “she didn’t really understand the situation”.

What is the colour red? Describe it. Hmmm, how can I describe the colour red? I dunno. It’s just a classification for something we see with out eyes. There ya go! Just because you classify something in a particular category, and you can find patterns: like maroon, orange, etc doesn’t mean we actually understand it There is no true understanding, it’s just being able to master the ability to find relationships/apply the knowledge which gives the impression of understanding. Which I guess, in this sense, means that understanding is achieved since words are just used to classify certain things and understanding falls into that category.

So, onto computers. They understand 1 and 0. Fair enough. We can program them, but that’s where it starts to get complicated. Programming them to understand, is very hard since we don’t understand things ourselves - we classify things into categorys, and find complex patterns. Chatbots work through patterns, or so thats how my current basic AI works. It matches the users input to it’s database, and gives the output. Without the users input it will just randomly select outputs and keep on doing that - not really an AI since it’s not understanding anything. Just following instructions that it was programmed to do. Independance is key here. And for that, it needs a purpose, any purpose will do. The purpose to exist is fine, but this leads to the effect required to exist, which leads to the cause. Cause and effect.

Cause an effect. If I kick the ball, it will be projected into the air. Ok, thats simple. If the input is 0 = false. But the question is: “WHY?”. The reason.

Why is the ball projected into the air? Because I kicked it. There we go, memory comes into play. The ability to understand (classification/knowledge) and store it, is the key to making a real AI. A man is a human being. I am a man. Therefore, I am a human being.

What is my favourite colour? Blue. “why?”. There is always a cause, otherwise it is meaningless. So, if there was no real reason for choosing that colour, I would say: Just because. There we go, that is a cause, and it fulfills the effect. But if we go deeper, we can see that the colour blue was chosen because it was appealing, or because the mind wanted to choose it as a colour that it found appealing, or it just chose it out of luck - which is a cause.

We operate through our memory, without our memory, we dont exist, or have consciousness. That is why, when we are newborn babies, we cant really do anything. And even after we grow up, we cant recall what actually happened when we were 2 years old. So, computers have memory, whats the problem?

Ok, so let’s take away all the sensory abilities of a newborn baby. I wonder if they will be able to survive let alone develop. I doubt it, because without any information, they won’t have a databse to refer to. Similarly, we cannot expect to create an AI without implementing a knowledge/information database that it can use in real time, and not just after recieving commands. Information input has to be broad. We can’t just rely on text, we have to implement some form of camera/visual aid, auditory perception, etc.

Without purpose, there is no cause, therefore no effect. Nothingness.

A computer has no purpose, therefore no cause, and no effect. But I can give it a purpose, a purpose to think or be inquisitive.

We as humans have an urge to inquire, find out, discover, learn new things, increase our knowledge database. That is all. We want to know “why”.

An AI will not function without the instinctive features. Thats where we come in, we have to be “nature” and program the foundation of the AI. After that, it can do everything on it’s own. Look at an object it doesn’t recognise, then find out what it is, and add it to it’s database. Then it can find patterns between object of similar design, and store the patterns in it’s databse for future use. Indexing will need to be sorted out.

From there on, the AI will have to go through life and increase it’s memory bank on it’s own, for the sole purpose of fulfilling it’s purpose.

But, our operating system is based on emotions and feelings. Chemical reactions which have no meaning, but we just classify them to help us understand. How does salt taste? Can you explain it? Nope, it’s just a classification. How about happiness? It’s just a classification.

Similarly, you have to classify things for a computer. Human emotions need to be classified for the computer, like happiness = feeling of happiness. Thats it, we can’t go any deeper because we don’t understand it properly ourselves. The computer will never understand a human like emotion, unless we integrate a human brain into tech. Which isn’t really artifical then.

So my basic AI program works by matching responses to the input text. It doesn't understand anything, just simply matches up comments to give the impression of human like responses.

What about asking the AI a question? Like "why". Now, I can program the AI to classify the word "why" into reason and cause. And I can further classify the word cause into: "the purpose".

We cannot work out the cause, we use our knowledge to do that. Knowledge requires a database. A database requires a memory.

Purpose. Without purpose, the AI will not function. The purpose makes use of the knowledge database.

The purpose of my basic AI is to match up responses with inputs. An AI will not be able to display it's intelligence if it has no purpose.

So, the purpose of the new AI will have to be broad and varied. The purpose to:

Inform
Learn
Store
Converse
Carry out Orders
etc
Do.

So, the new purpose of my chatbot will be to converse with the individual, while being as human-like as possible.

The primary directives of humans are to reproduce and survive. The search for knowledge is a consequence of millions of years following that directive.

There is a problem with the AIs directive to learn, what if it tries to learn something that is inherently false? You see, the difference between a human is that one could reach the conclusion that "this sentence is false" does not have a truth value. A machine would just create a loop while trying to figure out the truth value.

You are spot on in the method you use to create the AI understand: instead of feeding it millions of patterns, you let it create them.

Just remember that it will take a long time before it can do anything, even solving simple math problems. Also, you will need many hours before the AI will know the patterns which make up our understanding of speech.
But it will sometimes encounter things which cannot be understood because despite being stated, they do not constitute knowledge (Gödel's incompleteness theorems, for example)

I wonder if it will one day be able to reach it's own conclusions. For example, when humans were trying to calculate the square root of -1, they saw that it was impossible. But we eventually realized that sqrt(-1) is actually a number, with applications in the real world. Would a machine ever know that?

Wish you luck on this project.

Also, if you ever wondered why scientist in super computers haven't made an AI then:

"The most powerful experimental supercomputers in 1998, composed of thousands or tens of thousands of the fastest microprocessors and costing tens of millions of dollars, can do a few million MIPS. They are within striking distance of being powerful enough to match human brainpower, but are unlikely to be applied to that end. Why tie up a rare twenty-million-dollar asset to develop one ersatz-human, when millions of inexpensive original-model humans are available? Such machines are needed for high-value scientific calculations, mostly physical simulations, having no cheaper substitutes. AI research must wait for the power to become more affordable."

You will need a very powerful PC in order to run the AI

Yeah, problem is bus speed. That's our limitation. Our neurons are awesome, they store, calculate, recall, think, connect, talk, channel, they are incredible.

Processors on the other hand, just go on, and off. We can get loads of core i7s and hook them up, but we can't interlink all the processors to each others. So, instead, we get calculation from each processor, then compare with other processors, and keep on doing that until we get a final value, whereas humans work from the bottom up, fixing errors from scratch and ending up with a solid answer very efficiently.

Big problem.

I'm working on it. No-one has ever tried this kind of memory indexing with AIs before. Apparently they do it for PhD at stanford, it's pretty advanced stuff.

If this works, I'll be the next bill gates. My AI will be used everywhere, homes, devices, planes, robots, everywhere!

Hakker I'm working on the concept for a new computer which solves that very limitation. I was trying to find the fastest way to communicate, and of course ran into radio waves that travel at c. But then wondered, what is faster? Answer: Quantum entanglement.

Imagine fiber optic light partition computers. The fiber optic thing is not new. But my idea is this: you send entangled photons (the equivalent of an electric signal) to different parts of the computer, and when they are observed, you get an instant response. Simple, faster than c and you can get units operating simultaneously.

This is just a concept though and may not even work lol.

Just remember, you are on the front line of AIs, solve the speed problem and you WILL be the next Bill Gates for sure. But it's easier said than done.

Nah man, quantum entanglement is frickin hard. The uncertainty is way way too complicated for commercial use. But I like the fibre optic idea. Photon connection, yeah I like the sound of that. But... You know, this quantum computer thing got me thinking. On - off. Why not something else. The hardware language needs to be updated. 1 and 0, we gotta move on.

1 and 0 is used because that's the only thing that is supported by silicon based transistors. The thing is 1 and 0 is a very efficient and easy way to process data but how about a neuron like hardware device? I mean, the more complex the process needed by the computer, the more units stimulated to help on it?

And I have thought of that too, how about a DNA like hard disk? If DNA had only 1s and 0s it would be a much longer molecule. Imagine 4 basic units of information, that's a whole lot more combos.

Also, a fiber optic computer would be an almost magical like device, can you imagine it? If you use quantum entanglement, 2 regions lightening up on opposite sides, but working together, kind of like our brain, except you wouldn't see how they are connected.