World's largest neuromorphic supercomputer is switched on

a reply to: andy06shake

Completely agree, but this is simply the beginning, just an experiment if somewhat complex and costly.

Definitely a mere beginning and certainly a costly experiment!

I foresee that cost as being the true hurdle we have to ever achieving AI. Understand that I do not see AI as being some sort of 'code' that is changeable by the machine itself; AI is much more complex than that. As an example, have you ever seen a newborn kid (goat)? It lies these for a very short moment, then it begins to make an attempt to achieve a certain overall state condition (to stand up). It must learn to do this. The legs especially, but also the overall musculature begins to assume random configurations as the kid actually starts convulsing. Within a second or two, tat convulsing turns to attempts to stand up, as each random convulse is checked against partial success and either rejected as incorrect or accepted as a partial solution and used to improve the next attempt. Within a minute or two, the kid will achieve the goal and stand up.

This is learning. To walk, the kid uses the conditions learned to be favorable to standing to try and achieve the next goal: a step. A similar process ensues, with the difference that the actions previously identified as useless are rejected depending on the extent of their uselessness. This is the analog response I mentioned: there are varying degrees of 'uselessness' and 'usefulness' that are still considered, although not to the extent of learned useful solutions. Soon, these individual goals will lead to an animal capable of running, leaping, and performing some pretty impressive feats of balance.

That's all Pavlovian, based on a built in (aka 'pre-programmed') desire to stand and move about. Pavlov's experiment demonstrated that the goals are instinctive (his dog wanted food instinctively) but that proper responses to fulfill those instinctive goals are learned behavior.

A deeper examination of this phenomena indicates that there is a sort of 'critical mass' that is needed to achieve what we see as learned behavior... that is why I made such mention of a million processors being insufficient. There are millions of sensors in our bodies, scattered throughout, and billions upon billions, likely trillions upon trillions, of potential combinations of these sensory inputs that must all be considered simultaneously in order for a biological learning process to be successful. Even the lowly earthworm will have some of this behavior. The difference between it and the kid lies in the fact that the earthworm's intelligence is in reality mostly instinctive, requiring many less neural pathway possibilities, while the kid's goals are the controlling instinct and thus require many more neural pathway possibilities. The kid will have much greater learning ability and much greater control than the earthworm, at the cost of more sensory input and more neural pathway possibilities leading to a longer time as a child spent learning.

It won't be able to think but simply run code in a very specific manner, just like any other computer.

There is no 'code' that exists in an organic brain; this, I think, is where the confusion over what constitutes AI lies. A digital computer is not intelligent in any sense of the word. It simply runs software. Experiments with self-adjusting code and 'fuzzy logic' have overlooked one vital aspect of the very hardware design itself: the computer core is in no way any less than totally inflexible. It is composed of hardware which cannot physically change any aspect of its arrangement, only run a various number of pre-defined operations which are not negotiable in any way.

Compare that to an organic brain, which is capable of, indeed, expected to be by design, adjusting the very neural pathways that interconnect the simple yet almost unimaginable number of potential processors for any given series of inputs. This flexibility of the organic brain is what allows it to harbor intelligence, not the power of the processors within it.

That is the reason I lean toward an actual hardware-based solution for testing instead of a simulation run on a computer. The simulation might prove fairly accurate, but we cannot be sure due to the primitive level of research to date. We do not even yet know how quickly individual learning processes must occur to be viable, or if there are such restrictions on viability. Thus, how do we know what clock speed is needed, and of we even have the ability to generate clock speeds in the needed range? Consider the difference between analog and digital sound: with a sufficient sampling speed, any difference is negligible to our ears... but drop that sampling frequency and the sound will become 'digitized' in quality and eventually unintelligible.

The problem at this stage is that hardware solutions are outside the realm of practicality. The biological organism creates neurons automatically after conception; we have to create ours by hand. That is our largest hurdle.

TheRedneck

a reply to: verschickter

weak/false AI has arrived and it´s only a matter of time until someone leverages it onto a level to create true artificial intelligence and at that point, I hope if there is a god, he will have mercy on our souls.

I will disagree on that point. As long as we rely on the current method of digital processing through a core, true AI is not possible. We will have created nothing more than a simulation of true AI.

Also, note that in my previous posts I have relied heavily on a pre-programmed but flexible type of instinctive behavior. That is the beginning of all life-forms. Without this instinctive behavior, the impetus to learning is removed and we have nothing more than a big random pathway generator. Someone will have to pre-program this instinctive behavior. There is the dangerous part, and so far we don't even know how to make it dangerous.

Not because we created something that could be considered "life" and playing god but for creating a competetive, extremly fast thinking and decentralized intelligence that could wipe us out in so many ways, we´re not even aware off yet. By intent or not..

As a religious person myself (Christian), I do not see any issue with creating 'life.' The Bible declares we are "created in God's image," which certainly cannot be a physical reference because there is no physicality to God. I tend to agree with the concept that it is the ability and desire to create that makes us a creation in the image of our own Creator, and thus the pursuit of that creation instinct is entirely proper and religiously acceptable, if not religiously required.

TheRedneck

a reply to: verschickter

Plant´s for example do not posses organs or neurons or eyes, yet the breath, can see and they can count, among other things.

Plants also have the ability to change their very growth processes to improve their chances of survival. A tree, for instance, will grow toward a source of light. Over time, with no surrounding shaded areas, this will usually result in the tree growing straight because that is the average position of the light during the summer months where growth occurs. If it is surrounded by other trees on a single side, however, the tree will tend to lean away from those trees because the light it grows toward is no longer directly above, but skewed to the open side.

I consider this a very primitive form of intelligence.

TheRedneck

a reply to: verschickter

I am extremely interested in whatever results you obtain. Please post them!

TheRedneck

a reply to: TheRedneck

Yes, this is called etiolement in plants. More than that, they do not only have to decide when but also how far. It´s also triggered by touching leaves from other plants.

They stretch their nodes instead of building new ones when they undergo etiolement.

Since they can also count (see link above) and communicate, I consider them a primitive form of intelligence, too. They solve problems, take active counter measures at just the right places and it has also been learned that the root system will avoid a rock or before it hits it.

How can it know?

originally posted by: TheRedneck
a reply to: verschickter

I am extremely interested in whatever results you obtain. Please post them!

TheRedneck

I made my promise on ATS that if I get members to help me collecting data, I will post all results and the algorithm exclusive on ATS. I´ve got an overwhelming amount of samples in that thread in my signature and via protonmail. I always keep my promises if the fulfilment is possible.

So far I´ve run into a sort of heisenbug (not really, but it disapears when I look after it) and I treat water. I decided to take a small break from it and do something creative to get a fresh scope on the problem.

It has to do with the much feared red cross in .Net systems. Until I get to the source of that error, I can´t pull the data from the bitmap because the panel I draw on quits into oblivion with missing references in code and that brings the whole system down. ever. damn. time. Not only that I suspect a damaged CPU or RAM because as of recently, the whole machine just quits on me with all monitors searching for signal.

EyesWideOpen is also waiting for a sample to get processed. I did not forget him but currently I need time to get my head clear.

This is why I play around with my avatar currently. Some creative stuff. I have many projects running on parallel all the time. Sometimes I need a short break from all that and those are the periods I am most active on ATS.

a reply to: verschickter

Don't get me wrong, I love what you are doing here. I have some familiarity with Neural Networks, and yes, they can "learn" through constant repetition, and model/select the best paths that are the most efficient and follow the the statistical intent of the designer. By all means, pursue this stuff because it is important. I love the visual representation as well, as it really shows the data in a way that invokes all kinds of interpretation.

People who design computers and system programmers know that the fundamental working mechanism of any CPU, is the interrupt. It is a simile to the way we think, as we are constantly being "interrupted" with what we are thinking about to service all sorts of life events. When whatever interrupted us is satisfied/solved, we continue on with what we were doing.

So look at this computer analogy of what AI really has to be, if we were to build an autonomous, thinking machine.

Computers use a fundamental interrupt mechanism that is based on priorities. Some are just regular, pre-programmed
"timed" interrupts. These task based interrupts are required by the system so that things get done in a synchronized manner. A task pulls the interrupt bus low and the cpu stops what it is doing, places (pushes) a marker (address) of what it was doing on a stack when the interrupt occurs, and then positions the program counter to the address pointed to by the interrupt, which is a service routine. The CPU executes the code stream starting at that address, and then when done, pops the address of where it left off from the stack and continues on.

Next, is the unsolicited interrupt. These are not regular timed interrupts, but instead are random events that alert the CPU to an external condition that needs ASAP attention. It could be say, an interrupt generated by a disk driver when storage gets low. The driver pulls the interrupt bus low with the address of a service routine that will tell the user that they are running out of space... a simple case, and you can imagine how complicated this kind of service can get, depending upon what is going on.

Point being, these interrupts are pre-coded mechanisms that allow the multitasking capabilities of the system to function.

So, imagine an interrupt , not timed, not scheduled, not caused by any computer hardware monitor. An "out of the blue" interrupt where the address presented to the CPU is self generated. At that address is a code stream that the designers never wrote, but the computer has manufactured as a cognizant, unsolicited task (or thought) that it wants serviced.

This is what a true AI computer has to be in order to "think".

It is a paradox as to how a machine could become self aware, and is the reason that a true AI machine probably has to be an interface with a biological peripheral that contains the concept of awareness and is capable of generating unsolicited thoughts which are service routines that it writes itself, and then pulls the interrupt line.

Thanks for reading my diatribe as well

a reply to: TheRedneck

Nearly oversaw that post.

I will disagree on that point. As long as we rely on the current method of digital processing through a core, true AI is not possible. We will have created nothing more than a simulation of true AI.

Note that I wrote weak AI will enable us to leverage itself. By that I don´t mean, weak AI invents or designs strong AI. I mean the advancements in science that ultimately will lead us to true AI, maybe, someday, or not.

Just like the CNC, 3D printer and pick and place robot that I have, enables me to iterate through countless versions instead of planing it to death and never finish. Sometimes it´s wise to throw everything you have at a problem and just OCD through.

We (as a species) have already built a racecar that was blasted over tarmac by a professional while a bunch of neural networks took the sensor input and computed the best possible structural design.

I don´t have a link at hand but you should be able to find it very fast using the terms "AI car sensor race track".


I´m also christian, raised catholic but I since aquired my own model of god. I love my local christian community and most of what it entails (community, fests, helping each other) but I don´t subscribe to the catholic churchs dogmas.

Could not have written it better. However, if we create "life" or true conscious, we in turn are directly, physically responsible. That´s why I strongly disagree in abortions because "child could have a bad life". That´s unacceptable to me. It´s different if there are medical reasons or let alone rape and since I never was in that position I don´t dare to have an opinion on it. I just can´t.

But to get back on topic. If we ever accomplish that, we need to face the reality that we need to adress rights to "it"/"them". And with the potential behind what we brought to life/consciousness, we need to adress this problem before we actually do it.

It´s just to dangerous, not only to ourselfes but the whole environment. We as the current top predator and what we do, we need to get more responsible. We certainly don´t behave like a god would, by far not.

So this all leaves me with mixed feelings. I probably won´t be alive when this happens but my descendats, will.

a reply to: charlyv
Sorry for the mixed up responses. I dance on several weddings at once here and I want to take my time to respectfuly respond in an educated manner to each of you.


I´ve patiently read the interrupt part because I am familar with the system and of course I also have interrupts in place.

In fact, what you describe is what I do. Those unsolicited interrupts might not be de-coupled from the whole system but they emerge from the sieve I used to explain. To be more precise, the patterns that those sieves give, are also analyzed constantly and they will spawn a whole interrupt chain, that sometimes makes the whole framework act a bit hickup-ey.

Again, I don´t say/write that what I do will lead to true thinking, like we know it. I am aware of that border, and to close that thought, I think there is a high possibility with quantuum computers reaching that state.

Imagine a quantuum computer with the full knowledge/ruleset/behavior about quantuum mechanism itself. I can imagine that around that moment, maybe a quantuum computer could be able -provided the quantuum mechanism allow it- to go non-physical.

I´m not an expert in quantuum computing or theory but I know that it isn´t the way it is explained commonly like "doing all calculations at once". How I understood it, quantuum computers can leverage canceling out unreachable solutions like patterns. Sort of like probability of branch A is suddenly 0% and by that other related branches collapse because there can´t be a valid solution in those branches as well.

And that makes it so fast for some problems IMHO. I am sure this is not 100% correct, like I wrote, I´m not an expert in quantuum theory or computing.

a reply to: verschickter

Thank you!

As to your situation... much of that is over my head (I'm a low-level guy and it sounds high-level), but if the problem is video output I'd suspect the GPU or video card. You would know better than I based on the overall symptoms, but that was my first thought. Often a computer's video will go out and it appears that it is the entire computer.

I have many projects running on parallel all the time. Sometimes I need a short break from all that and those are the periods I am most active on ATS.

I know the feeling; I'm that way as well. It's very easy to get so deep into a project you start making mistakes because you're locked into a particular thought process. ATS is also my way of letting my head come back out and get some fresh air.

TheRedneck

a reply to: TheRedneck

I suspect it´s the RAM. Last time I ran a check, I noticed that the CPU fan was not running for two weeks (got interrupted shortly before I finished cleaning it) and nearly fried the CPU while doing the memtest. I considered this was the problem. Now I´m not sure anymore. I need to run a few instances of memtest to check that.

The machine is still executing the OS, I can spawn the "ding" sound when I hit WIN+E and then "ü" it will spawn a ding sound because there is no folder that starts with "ü".

But I can´t tab blindly through and ALT+F4 programs to free some memory, if it´s the RAM. I hope it´s not the card, if so, I might getmyself a TitanX as well. But they are a "bit" pricey and I don´t know if my mainboard supports it fully.

It´s just arduos, right now I got the RAM warning again. If I ignore it long enough, the above will happen. The systemlog will point out three tools using the most ram: VisualStudio, Firefox and Photoshop and a KernelPower failure but this should be me doing the reset.

Thank you for the hints!




@all, it´s getting late and all the typing and thinking in another language took a toll on my stamina. I´ll shut down now and take a walk, go to bed, tomorrow is a fresh day.

I´ll get back to you if someone made a reply.

a reply to: verschickter

I mean the advancements in science that ultimately will lead us to true AI, maybe, someday, or not.

One of the primary disadvantages of a written forum like this is that meanings often get misinterpreted. Yes, I agree that a simulation can be useful for future work. My intent was to emphasize that anything produced by a digital computer, no matter how powerful, is still only a simulation and not a final goal.

So I think we are in agreement.

If we ever accomplish that, we need to face the reality that we need to adress rights to "it"/"them". And with the potential behind what we brought to life/consciousness, we need to adress this problem before we actually do it.

It´s just to dangerous, not only to ourselfes but the whole environment. We as the current top predator and what we do, we need to get more responsible. We certainly don´t behave like a god would, by far not.

That is something I leave for the politicians to figure out. Certainly if we create something that could be considered life, we are then responsible for it. But what rights should it have? Equal to its creators? I am not certain of that; we don't have rights equal to God's, nor should we have IMO.

Any rights granted to our creation would be truly granted, after all. Our rights are inalienable, but that is because they were granted to us by our Creator. Our creations would possess whatever rights we deemed to grant as their creators, in the same vein. Now, depending on what rights we chose to grant, we could be seen as benevolent, loving creators or evil, malevolent creators, and our creations will likely hold differing views on that subject. We will have to make that decision when it comes time.

I must say I find it awe-inspiring to consider that I am now considering issues that God Himself might have considered when creating us... and I also consider it scary. You are quite correct that we often do not act anywhere near the way a god should, at least in my humble opinion. I wonder, was God ever concerned about His ability as our God? I think not, but then I do not know. How could I ever hope to know the mind of God?

Perhaps we will become better than we are and worthy of our titles when that day of creation finally comes. Perhaps something in physics requires such. I hope so.

So this all leaves me with mixed feelings. I probably won´t be alive when this happens but my descendats, will.

I have many projects, some starting as far back as 40 years ago, many much more recent. Each one has a different level of what I would term "publishment-friendliness." Some I guard with almost paranoia; others I am happy to allow outside my sphere of influence. AI is closer to the latter, simply for the reason you mentioned: I will never see it come to fruition. Perhaps my grandchildren will.

So, considering the extremely primitive state of this particular art, I am fairly unconcerned about the danger. While I recognize it does exist, and that at some point will become an issue, the chances for it to occur in my lifetime are about as great as me being hit by an asteroid while lying in an open field after being struck by lightning... twice. I don't worry about that either.

I produced two highly intelligent children with solid moral values. I feel secure they will do something similar. When that day of reckoning finally comes, I have done my part to advance the proper decision, whatever that decision turns out to be. I can do no more. That is the will of MY Creator.

TheRedneck

a reply to: verschickter

If the machine is still executing the OS, the only way it could be the RAM is for the RAM to be returning bad data without flagging it as bad. Possible, but unlikely that the BIOS would not pick that up.

Anyway, good luck with it and thank you for the new thoughts.

TheRedneck

a reply to: TheRedneck

Certainly if we create something that could be considered life, we are then responsible for it. But what rights should it have? Equal to its creators? I am not certain of that; we don't have rights equal to God's, nor should we have IMO.

I think we are now approaching the question, what is God, what defines God

Me personally, I think of God as kind of the spirit that connects all things. Not only life, but everything. In the sense of that God is everywhere and all reaching. Truely now defineable. It´s also a tool for moral, although some atheist keep saying they can act like humans without a god, I think they forget that, a dozen hundred years ago, the world wasn´t as safe as it is today. So religions provided a sense of community, laws and by that, protection. God is also hope.

The term "godless men" has it´s merrit.

I think the more connected we are to nature, the closer we are to what I see as god. I often take long walks, those moments, where you have a fresh breeze, maybe some warm sunrays too but no worries. Just living the moment, calm and peaceful. A truly blissful state.

I don´t judge anyones believe or disbelieve. What I find sad is that some atheists think they are somehow superior to people that believe in god, let alone have faith. They deem people like me as stupid and narrow minded but they don´t see the hypocrisy in it.

Anyways, to get a bit more on topic again, although AI is a topic that almost touches everything to a degree... Maybe you thought of it, too, when you wrote that:

I produced two highly intelligent children with solid moral values. I feel secure they will do something similar. When that day of reckoning finally comes, I have done my part to advance the proper decision, whatever that decision turns out to be. I can do no more. That is the will of MY Creator.

I´ve read in a book somewhere that we parents are seen as Gods. Because our little offspring has the impression that we know how the world works and that we have an answer to everything. Could this be gods hidden lesson? If you are a good parent or a loving human being towards others, I certainly think that what or whoever could be considered God, would like to see that any day, over someone just saying their prayers.

Even Matthew writes it in the Bible I think it was in Chapter 6. I recognize it as a book written by people like you and me, on a similar quest. Trying leave something positive in this world. This is what we do when we raise loving kids and take care of the life around us.

It brings us closer to god than just subscribing to a thought and doing the least amount of work, as in doing nothing bad isn´t enough, you need to do something good everyday.

If everyone would do this, we would be closer to "paradise" while still living.



I really enjoyed this conversation with you and CharlyV. Also, Turbonium, good that we could re-sort to a respectful conversation again.

Best wishes,
verschickter

a reply to: verschickter

I think we are now approaching the question, what is God, what defines God

If we cannot approach that question, it would seem to me that we have no right to become part of it.

I simply see God as the Creator. God is to us as we are to, say, a computer. The computer cannot comprehend what it means to be alive and have intuition and reasoning; these things are (for now) an alien concept to it. All that computer can do is execute instructions... it cannot know where those instructions came from or why it executes them. It simply does because that is what it is designed to do.

In a similar way, we are designed to create... to question, examine, explore, inquire... but what we explore and why, we have no idea. We do not know why we are here, just as a computer has no idea why it was created. If one follows the mathematics behind that logic, we are creation^1, creation to the first power. A computer would be creation^0, creation to the zero power, or 1, representing a lack of creative ability as we understand it. God would then be creation squared, creation^2.

What is creation squared? We cannot know, just as a computer cannot know creation. But I'd bet it's pretty awesome!

I think the more connected we are to nature, the closer we are to what I see as god.

I have similar feelings. I see nature as the expression of God's power, an example of the perfection He is capable of, and a reminder of how compared to HIm, we are so impotent. I have been caught several times just staring at a single blade of grass I plucked from the ground. In that single blade of grass lies a chemical factory that can take solar energy and convert it to chemical energy at an efficiency level far beyond our capacity. At the same time, it is self-repairing and self-replicating, two functions we so far cannot conceive of being possible were it not for that blade of grass, the lowest plant life, something we so take for granted we walk across it and crush it without thought, showing us it is possible.

We do not even understand how our own bodies function at the molecular level. Yet, we survive, thanks to bodies that are remarkably adaptable and resilient.

I don´t judge anyones believe or disbelieve. What I find sad is that some atheists think they are somehow superior to people that believe in god, let alone have faith. They deem people like me as stupid and narrow minded but they don´t see the hypocrisy in it.

Nor do I. I simply can't see the type of omnipotent God mentioned above being capable of being confined to a single religious belief. I do believe in an omnipotent and loving God, but I do not consider my belief incumbent upon anyone other than myself. Heck, I am not always fully capable of following the religious tenets I subscribe to; how hypocritical would it be of me to judge others based on their religious preferences?

I have always said one should be careful about hating anyone... hate Muslims? Don't blame anyone else if Muhammed is sitting at those Pearly Gates when you get there, talking religion with Chairman Mao. We tend to want to think we know who will be there, what will it be like, but the reality is that we don't really know. Everyone who could tell us is... dead! "We see through a class darkly," as the book says.

But I do agree that there is a great deal of hypocrisy in the anti-religious crowd. I actually hate to refer to them as "atheists," because true atheism, a lack of belief in a deity, does not prescribe such attitudes.

I´ve read in a book somewhere that we parents are seen as Gods. Because our little offspring has the impression that we know how the world works and that we have an answer to everything.

That's a new perspective on me, but I see reason in it. The Bible does refer to children quite a bit, and also tends to address parenthood.

As I mentioned before, though, these are all questions that society will have to deal with. Not our society, certainly, but that of our children, grandchildren, great-grandchildren... however long it takes. First things first; we need to understand how to get to the point where the questions we are discussing become something more than moot.

If the world stands, we will get there someday.

TheRedneck

Don't confuse advanced simulations in machinery being closer to the creation of an actual life....


We are no closer to creating life than we were before.


Life has nothing to do with simulations that appear more and more 'intelligent', or more 'real', making us ever closer to creating actual 'life,' one day...


They aren't closer to creating life.


Life is not the creation of an exact replica of any life forms. Or if it looks like we created an entirely new species, either.


They want us to believe that life is closer to being created, but it's all a sham.


All of these terms are used to imply that life, itself, comes one day, with these advances in human science.


Not so.

a reply to: turbonium1

Life the universe, reality, and the whole shebang, may very well be a simulation.

A simulation of a simulation of a simulation etc.

Should the holographic principle turn out to hold any weight.

The terms we use to describe what we assume constitutes life are still being defined.

But this is simply an experiment, it's not designed to simulate anything other than neural network architecture.

originally posted by: andy06shake
a reply to: turbonium1

Life the universe, reality, and the whole shebang, may very well be a simulation.

A simulation of a simulation of a simulation etc.

Should the holographic principle turn out to hold any weight.

The terms we use to describe what we assume constitutes life are still being defined.

But this is simply an experiment, it's not designed to simulate anything other than neural network architecture.

They are trying to simulate neural pathways of actual human brains. It only serves an appearance of more intelligence because it will appear more close to creating actual intelligence, which is a lie, yet they hope all believe is 'almost' real intelligence. Of course, that's why they are trying to simulate HUMAN intelligence. We can easily relate to intelligence if it appears to be much like 'human' intelligence. Simulated 'neural networks' are being created simply to help serve their ultimate goal.


None of this works to create an intelligence, of course. It is done to serve the appearance of creating intelligence, though, and it seems to be working out nicely. The goal is to fool people into believing they have created an actual, independent, form of intelligence. No other goal.

The claims of creating an intelligence will never be verified as true, or false, since it will be hidden behind the greatest excuse ever invented - "documentation on the intelligence created, cannot ever be revealed to the public, on grounds of national security"


It's all true, though. We have no reason at all to make up a story like this, for the public, on such serious, important matters!!

a reply to: turbonium1
Speaking about appearances..
Does the earth appear flat to you and do you believe it to be the case?

a reply to: verschickter

Don't ask about the Moon for Christ sake, as turbonium1 don't believe we went there in 69.