If our brain was a CPU then how fast would it be? What'll happen when it gets there??

M

Metallica_Band

2[H]4U
Joined
Oct 23, 2002
Messages
3,437
Hey there...I have always wondered...if our brain had a CPU chip instead of a brain, then how fast would it be??? What will happen when PC's can process info faster than the human brain??? Will it ever happen??? How long do you think it will take until it reaches that point??? THANX
C'YA ?:-D
 
A direct comparison between ICs build up out of n transistors and (biological) neural networks is impossible. They're completely different.

First of all, an IC can only perform the functions it was designed to perform when it is provided with input. It can not do anything on its own, nor can it redesign itself.

A neural network isn't just hardware, it's a collection of similar cells which through their interactions create 'software'. Exactly how this happens is still largely unknown at this point, but as we can observe on a daily base, its capabilities are quite amazing.

Furthermore, each cell in a neural network is a completely independent unit. It can be considered to be a small circuit, responding to input in a predetermined manner. The key to the functioning of a neural network is in the interactions between those 'circuits'.

ICs (CPUs) are only good at processing data in a certain manner for many times. They're just a circuit, just like a single neuron, ergo static ICs will never be able to behave like a neural network.
 
It's not comparable, if you ask me. As it has been said, the two are completely different, but even if it wasn't, they would both do better in different ways.

If you went head-to-head with a computer trying to calculate mathematical problems, obviously the computer would own you pretty bad. :p

But, obviously "video" processing, etc. is much faster for humans.

Lke Elledan said, we can create our own "software". A computer may be able to calculate numbers all day, but it doesn't understand what it's doing. Humans can understand what exactly they are doing, and can learn from it.

So, eh... It's a toss-up? :D
 
its very interesting to think about. but just imagine there would be people who would kill others for a high quality chip. i mean u could tell the guy had a good cpu cause he was very smart.

then some hoodlum takes a knife to your skull and takes your expensive brain cpu so he can get it inserted in his head. you would need protection, like a plate helmet. AC+5. wooo hoo!
 
mine would be a lonely Slot A Cyrix processor. I try to overvolt to get peoples attention but i get too depressed to perform well enough for anyone. i have my good days and there are those bad days where someone gets me stuck to an HSF.. i mean we arent even meant for each other and they try to spread this medication like adhesive all over me when i get too hot like feverishy.. it doesnt help much, they think fans would cool me down but i dont think so. then you know it gets worse when a new chip comes along and they stick a probe down your ass
 
Originally posted by RS3RS
It's not comparable, if you ask me. As it has been said, the two are completely different, but even if it wasn't, they would both do better in different ways.

If you went head-to-head with a computer trying to calculate mathematical problems, obviously the computer would own you pretty bad. :p
Click to expand...
Not necessarily. Due to the nature of neural networks it's not impossible that it (or parts of it) can be reconfigured to function as a kind of calculator. There's certainly quite some evidence which supports this theory, like certain individuals (not necessarily 'gifted' people) who are capable of solving certain kinds of mathematical problems within a second.

Don't forget that right now we humans may be quite bad at solving mathematical problems without help from a computer (relatively speaking), but it's not like we ever needed it before, did we? As the initial layout of a biological neural network is determined by the genes of the organism and a tiny bit of chaos theory, a calculator has not (yet) been included into this 'blueprint'.

But, obviously "video" processing, etc. is much faster for humans.
Click to expand...
Parallel, non-synchronous processing is a good thing.

Lke Elledan said, we can create our own "software". A computer may be able to calculate numbers all day, but it doesn't understand what it's doing. Humans can understand what exactly they are doing, and can learn from it.
Click to expand...
Highly simplified, self-awareness is a symptom of a neural network which has advanced suffiently to work beyond certain 'speeds'. This is why substances which negatively affect the operation of the brain
by impairing the transmission or reabsorption of certain types of neurotransmitters can lead to a 'blackout'. An individual who has been drinking too much alcoholic beverages may lose this self-awareness for a while and hence can afterwards not remember anything of what happened, or just some blurred fragments.
 
Originally posted by Metallica_Band
WHOW!!! You're gettin all scientific on me now...LOL...what's that mean in English??? LOL
Click to expand...
Reader's Digest version: an IC (CPU) should be compared to a single neuron, not an entire (biological) neural network (brain), as each neuron is, just like an IC, a single circuit. The working of a neural network is a result of the interactions between those circuits (neurons). The result of those interactions is the closest equivalent to we call 'software' with computers.
 
Well the fact that humans create computers kinda nulls the thought of having a chip smarter than the human brain. The fact that you couldn't even start to understand elledan's post is kinda pathetic though
 
Originally posted by goku238
Well the fact that humans create computers kinda nulls the thought of having a chip smarter than the human brain. The fact that you couldn't even start to understand elledan's post is kinda pathetic though
Click to expand...

Thats not true at all...we make cars faster then people.

What I want to know is how much "memory" can a brain hold? Compared in GBs?
 
Hmm this is an etremely hard question with no answer....We can remember music alot easier then text in some cases.

Is using an abreviation to remember like:
My
Mom
Eats....(or something like that)

To remember planets the Human version of ZIP Compression?
 
i think its more comparable to a form of indexing...zip allows you to store X amount of information in a space that is smaller than X itself. The kind of abrv. you talk about is more like indexing as you have the knowledge in your head but not in the right order...the abrv (its not actually and abrv its something else a neumonic or something like that)...gives you the correct order for the information should so you can recall it in that right order.

Do you see what I am getting at?
 
Originally posted by Jay D.
What I want to know is how much "memory" can a brain hold? Compared in GBs?
Click to expand...
We don't know how memories are stored, let alone in which form, so this is one of those questions which are impossible to answer.

It's likely, however, that the ability of a neural network to store memories is more akin to the way data is stored on a holographic medium than on a HDD, ROM, or similar, in the sense that each piece of data doesn't have a distinct position in which it is stored. If you were to store data on a holographic cube and then smash this cube to pieces, you could still retrieve the data stored on it, as each shard contains the full contents of the data stored on the cube, albeit at a lower resolution.

In effect, this would mean that the capacity of a neural network to store memories is virtually infinite.
 
Originally posted by Jay D.
Hmm this is an etremely hard question with no answer....We can remember music alot easier then text in some cases.

Is using an abreviation to remember like:
My
Mom
Eats....(or something like that)

To remember planets the Human version of ZIP Compression?
Click to expand...
We've less problems remembering the visual, auditory and other characteristics of a planet than the name of this planet. This is because the name, like every feature of a language, is completely arbitrary, i.e. there's no link whatsoever between the structure we refer to as a planet and the name we use to identify this planet.

Why is that the white stuff which comes floating down when it gets cold outside is called 'snow' in English, 'sneeuw' in Dutch, 'Schnee' in German and 'yuki' in Japanese?
 
I don't thinks thats true. I took AP Bio last year and was told that memories are stored in the wrinkles of the brain.

Also Where are you getting this holographic cube thing?
 
Originally posted by Jay D.
I don't thinks thats true. I took AP Bio last year and was told that memories are stored in the wrinkles of the brain.
Click to expand...
hippocampus - anatomy the system of elongated ridges on the floor of the lateral ventricles of the brain, thought to be the centre of emotion, memory, and the autonomic nervous system.

(source: Oxford Dictionary, Tenth Edition)

Anyway, your reasoning is non sequitur. The location of where memories are stored in a neural network says nothing about the way those memories are stored.

Also Where are you getting this holographic cube thing?
Click to expand...
What do you mean?
 
Originally posted by Elledan
hippocampus - anatomy the system of elongated ridges on the floor of the lateral ventricles of the brain, thought to be the centre of emotion, memory, and the autonomic nervous system.

(source: Oxford Dictionary, Tenth Edition)

Anyway, your reasoning is non sequitur. The location of where memories are stored in a neural network says nothing about the way those memories are stored.


What do you mean?
Click to expand...

Well I've never heard of holographic media...
 
Originally posted by Jay D.
What I want to know is how much "memory" can a brain hold? Compared in GBs?
Click to expand...
I can't begin to think how a comparison could be made. For example, how much human "RAM" is needed to remember how to tie your shoes? Tie a tie? Tie a knot? Tie one on? :p Multiply that by many thousands of very little things in our day-to-day lives that we don't even consider, such as the seemingly automatic eye blinking that we do subconciously in order to keep our eyes moist; yet we're also able to make the independent decision, with our conscious mind, to force our eyes to blink. Geez, just the comparative processing power and RAM needed for everything related to the eyes must be immense. Throw in the other senses, plus congnitive thought, plus our bodies' automated functions (eye blink, breathing, etc), plus whatever the brain is doing when we sleep...

Can't forget our built-in SMP. How many equivalent CPUs would it take to accurately mimic human multi-processing?

But then, just to make it more interesting, consider that experts say that we also use only about 10-15% of our built-in capacity . Talk about potential!

:eek: WOW :eek:
 
Originally posted by Bitchnmoan
Can't forget our built-in SMP. How many equivalent CPUs would it take to accurately mimic human multi-processing?
Click to expand...
One IC per neuron plus the software and storage required to make it work.

At least you won't have to synchronize L1 and L2 caches or anything ;)

But then, just to make it more interesting, consider that experts say that we also use only about 10-15% of our built-in capacity . Talk about potential!
Click to expand...
Correction: what those experts said is that at any given time, around 10-15% of the neurons in the human brain are active.

This is actually another example of why a comparison between a CPU and a neural network is utter nonsense. If a CPU is used for 100%, it means that all of its features (ALU, FPU, etc.) are in use. A neural network in which all neurons generate impulses at the same time is about as useful as a program of which all instructions are executed, in parallel, with none of the processing units having any knowledge of what the result of the instruction preceding the instruction they're processing was, or will be.

If such an 'impulse storm' were to occur in the brain of a person, that individual would basically have to be 'reset' (very technical term ;) ), by shutting down all brain activity using a strong electric pulse, and then restoring normal brain activity. If this isn't done, the affected person will die.
 
Originally posted by Jay D.
Hmm this is an etremely hard question with no answer....We can remember music alot easier then text in some cases.

Is using an abreviation to remember like:
My
Mom
Eats....(or something like that)

To remember planets the Human version of ZIP Compression?
Click to expand...

lol, damn right.
 
think of it this way. when someone is speaking to you, you think of a response inside your head by logic. obviously there are only so many things that a person can ask that will pertain to the subject at hand. when s/he is done speaking, you dont know it, but you already have response done or being formulated. a cpu cannot do that, it waits for the whole question before forumalating an answer.
 
Originally posted by Metallica_Band
Hey there...I have always wondered...if our brain had a CPU chip instead of a brain, then how fast would it be??? What will happen when PC's can process info faster than the human brain??? Will it ever happen??? How long do you think it will take until it reaches that point??? THANX
C'YA ?:-D
Click to expand...

I say we give the femals on this forum there won 1000 user ids.

So therefore if we say 13K to someone u cal him a women.



http://genmay.net/showthread.php?threadid=188946&perpage=100&pagenumber=1




And yes, mine would BSOD too.
 
suffice to say....

computers are dumb

they are nothing but mere tools that do nothing but what we tell em.
 
Originally posted by Jay D.
What I want to know is how much "memory" can a brain hold? Compared in GBs?
Click to expand...

Many scientists beleive that the storage is basically indefenite, in that the brain can always find more room to store information, and they think that the brain tends to create more folds in it over time as things are learned and experienced, thus expanding the amount it can hold. I've seen 1 person who estimated that the brain could hold about 250 megabytes of information, but I don't beleive this theory as 1 single DNA "blueprint" in your cells contains almost 1 gigabyte of information all by itself. A better estimate is derived from the estimated 100 billion neurons in the brain multiplied the 1,000 other neurons each one connects to (equaling the capacity of 1 binary KB per neuron) for a grand total of about 97,000 gigabytes of information. 97 terrabytes of brain storage sounds large, but that equation only assumes that all neurons are two-state, in which many are multi-state which would increase the capacity exponentially and therefore allow basically indefinate storage.

Originally posted by Metallica_Band
Hey there...I have always wondered...if our brain had a CPU chip instead of a brain, then how fast would it be??? What will happen when PC's can process info faster than the human brain??? Will it ever happen??? How long do you think it will take until it reaches that point???
Click to expand...

How fast? The brain has a tendency to occilate, or cycle, at about 40 megahertz when it is studying something new, and recalling memories. So you could say about 40 megahertz, though it can't be compared to a old 40 megahertz CPU since the CPU can do 1 peice at a time, whereas the brain can perform millions of things at the same time. The brain is better compared to a gigantic multiprocessor computer, carrying out many things at the same time. Megahertz is a meaningless number to tack onto something unless you know how many things can be done per cycle.

What would happen? We'd play Quake 3 faster. CPU's are already faster in the number of cycles and latency times than the brain and nervous system, tho it will take a long time for processors to ever equal the power the brain is capable of.

It may happen. Though CPU's will always lack what seperates us from them. Abstract thinking, an imagination, the ability to see, speak, hear, recognize, and most importantly no Intel inside sticker.
 
one thing bro, the computer has limits, the brain can form new neural pathways dynamically and we have intelligence, perhaps once we have created true AI, which may or may not be possible, they will be smarter, and they will murder us :confused:
 
Thinkabout it every cell in the body is like a computer. Some cells may even have SMP/ All are controlled by the brain. DNA stores Enormous amounts of information we can't even acess.
 
Originally posted by tdg
Many scientists beleive that the storage is basically indefenite, in that the brain can always find more room to store information, and they think that the brain tends to create more folds in it over time as things are learned and experienced, thus expanding the amount it can hold. I've seen 1 person who estimated that the brain could hold about 250 megabytes of information, but I don't beleive this theory as 1 single DNA "blueprint" in your cells contains almost 1 gigabyte of information all by itself. A better estimate is derived from the estimated 100 billion neurons in the brain multiplied the 1,000 other neurons each one connects to (equaling the capacity of 1 binary KB per neuron) for a grand total of about 97,000 gigabytes of information. 97 terrabytes of brain storage sounds large, but that equation only assumes that all neurons are two-state, in which many are multi-state which would increase the capacity exponentially and therefore allow basically indefinate storage.
Click to expand...
The flaw in this reasoning is the assumption that memories (data) 'stored' in a neural network are statically preserved, i.e. they're stored in a distinct location.

For one thing, neurons aren't "two-state", nor are they "multi-state". They're small analog circuits, meaning that they've a virtually infinite number of states. Beyond that neurons aren't just static. They constantly create and destroy connections with other neurons, a process governed by a set of relatively simple rules.

Please, forget already about comparing neural networks to ICs and memory (RAM) modules. Only someone who is completely uninformed about neuroscience and biochemistry could make a fool out of himself in front of better informed individuals in such a manner.


How fast? The brain has a tendency to occilate, or cycle, at about 40 megahertz when it is studying something new, and recalling memories. So you could say about 40 megahertz, though it can't be compared to a old 40 megahertz CPU since the CPU can do 1 peice at a time, whereas the brain can perform millions of things at the same time. The brain is better compared to a gigantic multiprocessor computer, carrying out many things at the same time. Megahertz is a meaningless number to tack onto something unless you know how many things can be done per cycle.
Click to expand...
In a neural network there are no 'cycles', no 'oscillations'. There are impulses which trigger other impulses, but the overall activity of a neural network can simply not be described in terms used for describing circuits which do employ a clock or similar.

[..], tho it will take a long time for processors to ever equal the power the brain is capable of.
Click to expand...
As I explained earlier, an IC can never "equal the power the brain is capable of", because it is static, just like a single neuron. The 'power' of a neural network comes from the interactions between those static circuits.

It may happen. Though CPU's will always lack what seperates us from them. Abstract thinking, an imagination, the ability to see, speak, hear, recognize, and most importantly no Intel inside sticker.
Click to expand...
You know what? I suddenly got this horrible flashback of cheap 'sci-fi/horror'-series common during the 1960s.
 
Thru some study and research, we are indefinitly provided with huge storage. Immeasurable.

We learn constantly, and remember a whole lot thru day-to-day tasks. Im comparison to a computer, there cannot be one made, its like comparing CISC to a B-wolf cluster. not something along the right lines. apples to oranges.

age and polutants cause problems. killing cells, damaging them and whatnot. it is said a human fed constant knowledge thru birth and nurturing could theoretically use an active 20-30% brain power if never tampered with psychologically or poluted.


but what fun is that? Ever noticed how REALLY smart people are never truly happy or "content?"
 
Originally posted by DeepFreeze
Thru some study and research, we are indefinitly provided with huge storage. Immeasurable.

We learn constantly, and remember a whole lot thru day-to-day tasks. Im comparison to a computer, there cannot be one made, its like comparing CISC to a B-wolf cluster. not something along the right lines. apples to oranges.

age and polutants cause problems. killing cells, damaging them and whatnot. it is said a human fed constant knowledge thru birth and nurturing could theoretically use an active 20-30% brain power if never tampered with psychologically or poluted.


but what fun is that? Ever noticed how REALLY smart people are never truly happy or "content?"
Click to expand...

I don't agree at all! I know some incredibly smart people who are perfectly content :).

guys i have to say this is an incredibly interesting thread. I have a question that I came up with as I was reading a scientific american article explaining the basics of quantum computers. I was wondering if there was any possibility that the brain operates on quantum mechanics principles...a.ka that the electrical impules neurons send to one another take an infinite amount of states at once, until "observed" by the neural network at which point it takes one "correct" state? thanks. elledan this question is mostly pointed to you, as it sounds like you know something about the theory of the human mind. I have a nagging suspicion (though unfounded as I don't have the background to really understand much of this) that the supercomputer AC/III Purple (something like that, being built by IBM, coming in 2004) which is supposed to be able to process of 100 trillion operations a second will not be able to rival the processing power of the human brain, despite the fact that brain is estimated to be able to process roughly that amount. Elledan I understand what you're saying about the brain not being anallogous to an IC, so I guess I sound somewhat ignorant in comparing the two, sorry :).
 
Another interesting contrast is that computers do not, nor do I foresee them ever having, a train of thought. Imagine that you are thinking about upgrading your PC, then you drift to the trip you took to the store with your girlfriend, then think about some date you had with her, then you think about the store next to the restaurant where the two of you had dinner, a music store, then your OWN music... see where this goes? Computers will likely not be able to have random inspiration of the sort that we are used to having. They won't have ideas, only assumptions based on the programming that was fed to them. The programming and processes are too rigid to allow this rambling; imagine that your computer decides to "recall" some Excel spreadsheet that it was working with two days previously while you are doing some HTML programming; maybe you typed in some code that made it remember something from your financial projections: "what the - why is Excel opening? Why is the data changing?!?!?"

My guess is that any real progress in the realm of true AI will be coming from the biological realm, wherein medicine creates an organic brain, not an electronic one.
 
Now that I think about we don't remember alot of the info we know. We use other things we know to be true to peice together our memories.

Thats why a song is easy to remember because we can just keep going with the rythme and guess a word that fits.

Also what would be comparable to having something on the tip of you toung to a computer?
 
Originally posted by Bitchnmoan
Another interesting contrast is that computers do not, nor do I foresee them ever having, a train of thought. Imagine that you are thinking about upgrading your PC, then you drift to the trip you took to the store with your girlfriend, then think about some date you had with her, then you think about the store next to the restaurant where the two of you had dinner, a music store, then your OWN music... see where this goes? Computers will likely not be able to have random inspiration of the sort that we are used to having. They won't have ideas, only assumptions based on the programming that was fed to them. The programming and processes are too rigid to allow this rambling; imagine that your computer decides to "recall" some Excel spreadsheet that it was working with two days previously while you are doing some HTML programming; maybe you typed in some code that made it remember something from your financial projections: "what the - why is Excel opening? Why is the data changing?!?!?"
Click to expand...
"Another interesting contrast is that neurons do not, nor do I foresee them ever having, a train of thought. Imagine that you are thinking about upgrading your PC, then you drift to the trip you took to the store with your girlfriend, then think about some date you had with her, then you think about the store next to the restaurant where the two of you had dinner, a music store, then your OWN music... see where this goes? Neurons will likely not be able to have random inspiration of the sort that we are used to having. They won't have ideas, only assumptions based on the programming that was fed to them. The programming and processes are too rigid to allow this rambling; imagine that your neurons decide to "recall" some thoughts that they were working with two days previously while you are doing some HTML programming; maybe you typed in some code that made it remember something from your financial projections: "what the - why is Excel opening? Why is the data changing?!?!?""

Look at it this way: neurons are the hardware, the connections between them, and the resulting interactions between those neurons is the software.

Neurocyte: the hardware;
Axon and dendrites, the connections they form: software;
Impulses send over the resulting network: software.


My guess is that any real progress in the realm of true AI will be coming from the biological realm, wherein medicine creates an organic brain, not an electronic one.
Click to expand...
I disagree. At this point it's far more difficult to assemble (or convince to self-assemble) structures consisting out of mainly (or only) organic molecules than it is to work with non-organic molecules.
 
Originally posted by pakotlar
I don't agree at all! I know some incredibly smart people who are perfectly content :).
Click to expand...
Yeah, look at me for example ;)

[..]I was wondering if there was any possibility that the brain operates on quantum mechanics principles...a.ka that the electrical impules neurons send to one another take an infinite amount of states at once, until "observed" by the neural network at which point it takes one "correct" state?
Click to expand...
Well, there might be certain quantum mechanical principles involved in the functioning of a neural network, but I really doubt that it's anything as obvious as you describe.
I have a nagging suspicion (though unfounded as I don't have the background to really understand much of this) that the supercomputer AC/III Purple [..] which is supposed to be able to process of 100 trillion operations a second will not be able to rival the processing power of the human brain, despite the fact that brain is estimated to be able to process roughly that amount. Elledan I understand what you're saying about the brain not being anallogous to an IC, so I guess I sound somewhat ignorant in comparing the two, sorry :).
Click to expand...
There really is no way to compare an IC and a neural network. As I said before, they're completely different things. It's like comparing a bolt to a car.
 
You must log in or register to reply here.
Back
Top