Researchers at IBM have discovered a new way to stimulate an immune response in the body, using graphene surfaces covered with a thin layer of polymer. The discovery is presented in an article published in the scientific journal Nature Communications.
For the purpose of cancer treatments, researchers have previously developed techniques in which the drug molecule is linked directly to nanomaterials, such as graphene surfaces. The combination between nanomaterials and drug molecules helps doctors to deliver the drug directly to the tumor, where the drug is released into the cancer cells and helps fight the disease.
In an article he previously published in the scientific journal Nature Nanotechnology, the same team of researchers proved that nanomedicines require a combination with coating molecules, such as polyethylene glycol, in order to create biocompatibility and prevent damage to the cells surrounding the tumor. The new study points to the possibility that tiny graphene surfaces coated with a polymer will have another secondary effect, which was not known until now, in the form of stimulating the secretion of cytokine proteins. These proteins act as "warning flags" in the human body, and attract cells of the immune system to the vicinity of the graphene surfaces, such as Helper T Cells, which pass on these signals - and Killer T Cells, which eliminate infected cells.
Simulations performed on the IBM Blue Gene supercomputer show that the polymers attached to the surface of the nanoscale materials increase the initial signaling process carried out in the body. The researchers present a possible explanation, according to which these signs are transmitted within six hours of the detection of the nano drug.
According to Dr. Rohong Zaho, director of the soft materials research field at IBM laboratories, "we actually discovered a new way that allows doctors to activate the body's immune system and put it into action, and this is not an easy task. This discovery may be a huge development in the field of precision medicine. If we focus such nanomaterials, for example, on tumors or cells that have been damaged by a virus - it will be possible, in principle, to stimulate the activity of the immune system in order to attack the cancer and infections, already at their earliest source."
Clinicians may combine treatments based on the immune system, such as nano-therapies that carry traditional drugs, which are already adapted to the graphene surfaces. A combination of the nanoscale treatments and traditional medicines and the response capacity of the body's own immune system may lay the foundation for new ways to fight diseases.
66 תגובות
rival
The idea of the Chinese room is wrong, in my opinion. But - Searle is absolutely right in his conclusion. He lacks a basic understanding of computer science. It describes a person who performs a sequential algorithm that answers the questions. This was really the thinking in the XNUMXs, when it was thought that intelligence could be simulated by a computer program on a Turing machine. The assumption at the time was that our brain has the same computational power as a Turing machine.
Today we know differently. Interestingly, Alan Turing himself also knew this, and it was the subject of his doctoral thesis. The model described there is much closer to our consciousness than a model of a normal computer. This is why we no longer try to write intelligent software, but use neural networks.
But - our brain is not just a neural network. It has two parts. These parts have names like left brain / right brain or linear brain / associative brain and so on. To the second part Turing happened o-machine. I suggest you read his thesis, it is very interesting.
In short - Searle's mistake, in my opinion, is that he ignored the problematic of the image of the same second part, a part that does not perform algorithms.
Alpha-Go is built using this method, and has a linear part and an associative part. He can play any game in the world, but he learns in a way that is fundamentally different from the way a person learns. I want to see Alpha-Go riding a bicycle, or explaining to me that the smell of citrus is more pleasant than the smell of rotten onions.
Miracles,
I read the first link you provided, what is not convincing. This is just another variation on the Chinese room example, as if in our brains the situation is different. If that's what convinces you, good luck.
Miracles,
As for the worm, when the number of neurons in the neural network is so small it is not at all surprising that any small change can have a dramatic effect, I don't think you can infer anything from that about a human brain. Hundreds of thousands of babies born every day whose connections in their brains are completely different from each other, and most of them still have a developing human consciousness, prove in my opinion that your argument is completely wrong.
Thanks for the links, I took a look at them but it seems to me that I will only have time to read them after the holiday.
https://www.technologyreview.com/s/531146/what-it-will-take-for-computers-to-be-conscious/
http://www.rawstory.com/2016/03/a-neuroscientist-explains-why-artificially-intelligent-robots-will-never-have-consciousness-like-humans/
rival
It's interesting, but it doesn't change what I'm asking - what is the basis for thinking that a person's consciousness will develop in such a machine? Experience shows that this is not the case. In the OpenWorm project, they are trying to simulate a worm whose brain has exactly 302 neurons. The way they do it is by imaging each and every neuron. If you take 302 artificial neurons and connect them in a different way than in a worm, you will not get anything resembling a worm. Our brain has 100 trillion connections, and we know that sometimes slight changes in these connections cause huge changes in consciousness. And we also know that there are cases where large parts can be lost without (apparent) change in consciousness. So, we don't understand much…
Here are two short links (your link doesn't talk about our topic):
rival
It's interesting, but it doesn't change what I'm asking - what is the basis for thinking that a person's consciousness will develop in such a machine? Experience shows that this is not the case. In the OpenWorm project, they are trying to simulate a worm whose brain has exactly 302 neurons. The way they do it is by imaging each and every neuron. If you take 302 artificial neurons and connect them in a different way than in a worm, you will not get anything resembling a worm. Our brain has 100 trillion connections, and we know that sometimes slight changes in these connections cause huge changes in consciousness. And we also know that there are cases where large parts can be lost without (apparent) change in consciousness. So, we don't understand much…
Here are two short links (your link doesn't talk about our topic):
http://www.rawstory.com/2016/03/a-neuroscientist-explains-why-artificially-intelligent-robots-will-never-have-consciousness-like-humans/
https://www.technologyreview.com/s/531146/what-it-will-take-for-computers-to-be-conscious/
The link didn't go to the clipboard, but if you copy and paste it into the browser it will bring you to the appropriate page.
Miracles,
Read here on pages 104, 105, 106 -
https://books.google.co.il/books?id=DMqpCwAAQBAJ&pg=PA104&lpg=PA104&dq=%22the+not-so-large+number+of+neurons%22&source=bl&ots=zJF1CCEmP5&sig=hvhtaVo8VhRGwdTH0VXlqXBDGow&hl=en&sa=X&ved=0ahUKEwiYnMKiwJPUAhXCJSwKHRMKB70Q6AEIGjAA#v=onepage&q=%22the%20not-so-large%20number%20of%20neurons%22&f=false
I don't agree with you and I don't think my claim is baseless, but we've talked about it enough times and I don't think we're getting anywhere. Evolution has already created a conscious mind, I see no reason to repeat this process again to gain consciousness one more time. It would be double work and I don't think it's necessary, we already have something that works, it's a shame not to use it.
rival
The questions are indeed philosophical, but that doesn't mean you shouldn't ask them 🙂
Can you direct me to a serious source that claims that the number of neurons of that dolphin is incorrect?
I agree with you that the structure of the body and the environment are very important for the development of consciousness. Take a human baby and raise it in the forest, then its consciousness will be closer to that of a chimpanzee than to that of a modern human.
But, the distance between that and building a machine the size of a baby, that eats baby food, and will develop into an adult is huge in my opinion.
We know very little about a baby's brain, much less than you think. In the rare cases where a baby is put in an MRI scanner, it is a baby with severe neurological problems. And certainly no one is in a hurry to do a CT scan for the baby. Remember that you claim to start from the consciousness of a baby, something about which there is almost no information.
The idea that a machine that has 100 billion neurons and 100 trillion synapses will develop a consciousness, and another consciousness like that of a human, is, in my opinion, without any basis.
I do think that if we develop a type of robot that reproduces and evolves, then it is possible that in the future robots will develop consciousness. I think it can happen very quickly. But I see no reason to think that these robots will like us, especially when they realize that we are competing for the same resources.
Miracles,
1. You have many questions, some of which are philosophical and I don't have answers to everything, I'm saying a simple thing, let's build our neural network as similar as possible to a human brain and see what the result will be! Surprises are one of the most fascinating things in science, don't you think?
2. I know that Wikipedia mentions a certain type of leviathan (a distant relative of a dolphin) whose cerebral cortex supposedly has more neurons than a human's cerebral cortex, but I looked into the matter out of curiosity and it turned out to me that there is a serious criticism of the research from which these conclusions were drawn, both Probably not correct and there is no consensus about them among the scientists.
3. Even if the leviathan you mentioned really has a developed mind like ours it is not enough, it also needs to live in an environment similar to ours, and have a body similar to ours. Imagine if they transferred your brain and your consciousness to the body of a leviathan swimming in the ocean, what exactly could you do with it? Could you as a Leviathan demonstrate your intelligence? Could you use your clumsy fins to build devices? Or something useful? You must understand that this is a bit problematic…
4. It is certainly possible that the new consciousnesses that we will create will decide to rebel against us, you know that many films have been made on the subject, are you familiar with Terminator? So be afraid, be very afraid 🙂
rival
So according to what you say, a dolphin has a higher consciousness than a human. do you think this is true
Do you think that on other planets that have life - human consciousness will always develop there? I understand that you think consciousness is one-dimensional, meaning that it starts with an amoeba, goes through human consciousness, and then continues to a state we don't know. Let me put it another way - I understand that you claim that all two species whose brains are equally developed have the same consciousness?
And another question - what will happen to this consciousness in the machine, when it continues to develop? Her consciousness will surpass ours - so can't a situation happen where she starts performing experiments on us, or she works to take control of us?
rival
So according to what you say, a dolphin has a higher consciousness than a human. do you think this is true
Do you think that on other planets that have life - human consciousness will always develop there? I understand that you think consciousness is one-dimensional, meaning that it starts with an amoeba, goes through human consciousness, and then continues to a state we don't know. Let me put it another way - I understand that you claim that all two species whose brains are equally developed have the same consciousness?
And another question - what will happen to this consciousness in the machine, when it continues to develop? Her consciousness will surpass ours - so can't a situation happen where she starts performing experiments on us, or she works to take control of us?
Miracles,
What makes me think so is a simple look at the natural world, I see that the more developed an animal's brain is and contains more nerve cells, the higher the consciousness that is created in it.
Therefore, I conclude by way of logic that if the neural network that is built is big enough like a person's cerebral cortex then it is also the consciousness that will be created there.
rival
I agree that consciousness will emerge from a complex sapphire machine. But - what makes you think that it will be the consciousness of a person, and not of a dolphin, for example?
rival
The neurons do not suffer... what, or who, suffers, is consciousness.
I don't understand you - don't you think that machine consciousness could suffer? Let me ask you a question - does a wounded dog suffer? Of course you think so, but what is this suffering? After all, in your opinion, consciousness is just software that runs on the brain's hardware. So how can there be suffering? How can there be feelings at all?
Regarding the emergence of consciousness - I agree with that. But there is no reason to think that it would be a human consciousness. Do you think there is a direct relationship between intelligence and the number of neurons? And once we reach the number of neurons a person has, then will a person's consciousness actually emerge?
Miracles,
Okay, but that's what your words implied. So explain to me then, what is the connection between the fact that we don't talk about the moral side and the suffering of a network of neurons, and between an understanding of consciousness? (and the truth is that they do talk about it a lot)
Second thing I have already explained to you in previous discussions, consciousness in my opinion will in any case develop in the network whether or not we understand what it means exactly and how to define consciousness. If we build our neural network properly and allow it to experience the world like a small child or a baby, consciousness will emerge from the network spontaneously without us having to direct it.
rival
I wasn't talking about you 🙂 I'm not attacking anyone else either. I am trying to substantiate my claim, that we do not yet understand what consciousness is. And we are probably not close to developing human consciousness in a computer.
Miracles,
"But what does that have to do with the matter? You don't have anything to say, so you get down on personal lines?'
And isn't that exactly what you're doing when you "wonder" that I'm not talking about the moral aspect of developing consciousness in a computer? What exactly were you trying to imply when you said that? Was it not an attempt to impeach my moral sense?
I, for example, I fed a cat once.
Well… I didn't really feed… I threw a piece of sausage at him… but he ate it!
rival
I am close to vegetarian. I am an active member of several green organizations, such as the WWF, and I have taken care of injured and orphaned animals several times (eg feeding hedgehog puppies every hour for a week...).
But what does that have to do with it? You don't have anything to say so you get down to personal lines?
Miracles,
Why talk about the future? Let's talk about now. Today there are countless neural networks suffering, suffering greatly. Even in the food industry (chickens, cows, sheep...) and in the field of medical experiments (mice, monkeys...) these neural networks suffer both severe physical pain and extremely severe mental distress. Have you, personally, done anything to end this suffering? are you vegetarian Have you ever gone to a demonstration against animal testing? What have you personally done to reduce this suffering? Or is it just not interesting to you?
rival
I probably didn't explain myself well enough. What I'm saying is that those who do believe that we are close to creating human consciousness are the ones who must think about the dangers on the one hand, and the moral aspect on the other. The very fact that those believers do not talk about it, shows that we do not understand what consciousness is (in my opinion).
There is a fascinating science fiction writer who has written exactly on this subject. Here is a link to his short story, which describes what I am saying.
http://home.sandiego.edu/~baber/analytic/Lem1979.html
Miracles,
I don't understand something, if you think that the creation of human consciousness in a computer is something that is so far from us, why do you keep emphasizing here the subject of suffering that may be caused to such a consciousness? And the dangers it could create for humanity? Why does it even bother you if you think such a thing won't happen soon?
rival
You claim that we are close to creating human consciousness in a computer.
I claim that we do not know what consciousness is. Evidence of this is that there are those who think it is moral to create consciousness and then kill it. Evidence for this is that there are those who think that a consciousness will be created that will be "on our side".
What I suggest is to be humble, and stop getting excited. I am definitely in favor of research and development, but, no false promises.
1. First of all notice that you jumped from one subject to another, first you talked about the suffering that may be caused to the neural network, and now you are talking about the dangers to humanity.
2. "Furthermore, one should understand that intelligence is the ability to solve problems with partial information, therefore a smart machine can always make a mistake"
But this is also true of humans.
3. All the things you said are known and many articles have been published on the subject (for example Stephen Hawking's words a year ago: "Artificial intelligence threatens humanity") but what do you propose to do with it practically? Stop developing artificial intelligence? impose restrictions? What is your solution?
rival
I say that you should remember to turn off the gas, and not rely on the fire services...
I am definitely in favor of research, but we need to be aware of the possible outcomes, and I think that the desire to create consciousness is very dangerous.
Beyond that, it should be understood that intelligence is the ability to solve problems with partial information, and therefore a smart machine can always make a mistake.
Yes, it seemed that he was saying crow, but it sounded like he was saying crowd, I saw that the automatic translation there was also confused in identifying the word.
Miracles,
Many have already dealt with these philosophical questions before us, but what do you propose? Should we stop developing artificial intelligence until we understand 100% how the brain works and how to build its simulation on a computer? And then we will build at once a perfect simulation of it that will work without any problems?
rival
He talks about a capuchin monkey and a crow. He is talking about intelligence, not consciousness, and I think they are very different things.
My point is different - I'm talking about the steps before realizing the consciousness of a wise person. Before that - we will have partial consciousness - do you think it will be the consciousness of a mentally retarded person, or the consciousness of a chimpanzee?
Do you think it would be right to unplug such a consciousness? Would she not be aware that she could be killed in such a way? Don't forget we agree that she should have eyes, hands and feet….
It seems to me that I was wrong and he is talking about another small mammal (I couldn't understand the word he says there, miracles?) but the principle is the same - from an evolutionary point of view, only a few seconds passed from the moment a small mammal with a cerebral cortex was created until human consciousness was created.
Again, the segment I was talking about starts at minute 22:20.
Here is the lecture, it is really interesting and worth watching in its entirety:
https://www.youtube.com/watch?v=eX2sb-TKOvI
The segment I mentioned starts at minute 22:20.
Miracles,
I recently saw a lovely lecture by a Frenchman who runs a company that deals with artificial intelligence (I'll give a link as soon as I find it), he says that once we reach the level of consciousness of a rat we will very quickly reach the consciousness/level of intelligence of a person. This is what happened with evolution, as soon as a rat was created, within a few evolutionary seconds man also developed.
rival
I agree with you. In your opinion - what will the last intermediate stages look like, towards full consciousness?
http://hackaday.com/2016/12/11/train-your-robot-to-walk-with-a-neural-network/
Miracles,
Obviously there will be intermediate stages, what a question, every neural network we build today and connect to a toy robot is an intermediate stage on the way there...
rival
I am asking about the stages of technology development. The final stage is known - consciousness of a healthy and intelligent person. I'm asking about the steps before that.
I didn't understand, are you asking about the development stage of this technology? Or about the final version when we want to test it?
rival
Don't you think this machine will be built in stages, meaning there won't be "initial" versions at first?
No miracles, of course it won't happen all at once, awareness will develop in the neural network gradually as it experiences the world and learns to act in it through the body it is connected to.
Just like the consciousness that develops slowly and gradually in the brain of a baby.
rival
Do you think that the consciousness of a highly intelligent person will suddenly emerge, or will there be stages before that?
Miracles,
I'm not talking about transferring an existing consciousness (yours for example) into a robotic body, but about a raw mind that has not yet formed itself that will slowly develop inside a synthetic body. That's why I gave the example of a baby who is slowly learning to control his body.
I don't agree with you that you should start with a baby's body, it can certainly be a body in its final form (of an adult robot) which at first will probably crawl on the floor, but slowly when its neural network starts making connections with its arms and legs and learns how to operate them it will be able to stand and walk and run
I am not ignoring the stages of development, from the very beginning the cognitive brain will be connected to the body and learn to control it, in the same way that DeepMind's neural network learned how to control the controls of dozens of computer games and reached the level of a professional player.
rival
I think you are confusing several things.
If you want to start with a baby's brain, then you will also have to build a baby's body. For example, we know that a critical stage in a baby's development is crawling. The brain changes physically during the first years, and you will also need to simulate this in a good way.
If you think it is possible to take the consciousness of an "adversary", you will need to simulate your whole body well. The brain can fix small things, but there's no reason to think it can get along with a whole new body, including all the senses.
And the most serious in my opinion - you ignore all the stages of development. Detach a person from his senses - and he will lose his sanity in no time. If he lacks organs he may suffer from phantom pains. If the brain "discovers" unfamiliar organs - it will suffer from "xenomelia", a phenomenon in which a person feels that his organs are not his own.
That's why I agree with Deutsch that we don't even understand what the problem is...
Miracles,
I don't underestimate him and I understand that he is a highly respected physicist in his field with important contributions in the field of quantum computing, I just don't think that this expertise allows him to give a professional opinion in the field of neural networks and artificial intelligence.
Regarding your question, I think our brain is flexible enough that it knew how to adapt itself to almost any robotic body you connect to it. I don't see it as "computer simulation", it's not the matrix, you won't live inside a computer, you'll live in the real physical world where you live now, only your brain will be made of cognitive chips instead of a biological brain, and your body won't be biological either, but you'll be able to feel things , see through cameras, etc.
The situation you describe is quite similar to the brain of a newborn baby, it takes time for it to learn to adapt itself to the body it is connected to, and many times it can also be a damaged body for example a body without arms or legs, blindness and other failures.
rival
Our brain is surely a machine - that's exactly what Deutsch proved (even though you underestimated him....). But, our brain is not a computer - I understand that even Alan Turing thought so. To be precise - our brain is not equivalent to a universal Turing machine.
But that's not the point. I ask - how will you feel when you know that you are a simulation inside a computer? And don't forget - we don't start in a state of perfect visualization. You are not a brain inside a robotic body that is exactly the same as your body. And even if so - how do you make sure that there is coordination between the age of the mind and the age of the body?
Miracles,
Why assume? Our brain is exactly the "monstrous" and powerful computer you are talking about, and I feel great living inside it.
PS - the one who usually gets angry and aggressive here is you...
rival
Let's assume that your consciousness has been duplicated and now you "stay" inside some monstrous computer. Without getting angry as usual, try to think - how would you feel?
Correction - with all due respect to Deutsch, I do not think that his field of expertise in physics allows him to have a professional opinion on the subject of consciousness or artificial intelligence. This is the domain of neuroscientists and people who deal with artificial intelligence and neural networks, it is not the domain of those who deal with quantum physics.
Miracles,
With all due respect to Deutsch, I don't think his area of expertise in physics allows him to have a professional opinion on the subject of consciousness. Language is definitely critical for creating a high consciousness like ours, but I've told you that a million times before and I also explained to you that we need to do a few more things besides connecting 100 billion neurons, too bad we're just treading water.
rival
Deutsch knows more than anyone alive today what a computer can do. Do you really want to argue about this?
Do you think we understand what consciousness is? Can you give me a definition? I will ask a simple question - do you think a language is needed for consciousness? Does someone who knows two languages have more consciousness? Does the deaf-mute have no consciousness?
You think that if we connect 100 billion neurons then we will get human consciousness. But, you know we connected 1000 and didn't get the consciousness of a worm.
So what do you rely on?
Alpha-Go is not a neural network that plays Go! It also has a network of neurons...
Yaron,
Here is an article that may interest you in the future:
http://www.tapuz.co.il/blogs/viewentry/1065939
Miracles,
Could you please explain to me how Mr. David Deutsch's specialization in physics and quantum theory allows him to have such a strong opinion on a field in which he has no expertise such as neural networks, brain function and artificial intelligence? Why should anyone give their opinion more weight than the opinion of brain researchers, neural networks and artificial intelligence who think this is possible?
It's funny that he claims that for many decades there has been no progress in the field of artificial intelligence... because since he published the article there has been a development that many define as meteoric in the field of artificial intelligence, mainly thanks to the increasing use of deep neural networks at the same time as the increase in the power of computers. There are many examples from recent years of cases in which artificial intelligence software (based on a neural network) made creative moves that even those who developed the network did not imagine could be made (I came across several such examples, for example in the field of computer games, or in the Go game).
I took the trouble to read the article you linked to, except that he claims it is "impossible" (because only man supposedly has the ability to create new ideas that were not pre-defined by a programmer) he does not give any serious substantiation for his arguments, so he said.
It is better that he continues to deal with physics and quantum mechanics and let the experts in artificial intelligence and brain research do their work.
Rival/Yron
David Deutsch is a physicist who proved that it is possible to simulate any physical process with the help of a computer. He is also one of the forefathers of the quantum computer. So, we can assume he knows something about computers right?
So, please, I suggest you read his take on artificial intelligence. Will you mock him too? 🙂
https://aeon.co/essays/how-close-are-we-to-creating-artificial-intelligence?utm_source=Aeon+Newsletter&utm_campaign=ac75a42213-EMAIL_CAMPAIGN_2017_05_16&utm_medium=email&utm_term=0_411a82e59d-ac75a42213-69476645
Rival: There are neurons that as children have no information in them or have information in them but it does not stay the same, and as adults up to the age of 26 their programming is carried out. The argument is not that fundamental to insist in front of you. That is, during life when it is said that our personality is being shaped, our neurons are programmed in knowledge patterns.
rival. I agree with you. Reality will hit the stubborn.
Christians but dangerous precedents. While I struggle to get a calculation power of 1000 neurons implemented in the software, and the platform at my disposal is a core i7 with an NVIDIA processor (this is not an advertisement), less than that the calculation crashes for me, on a core i5 with an Intel graphics card, more than 200 neurons fails.
At IBM with the Synapse chip they have 300,000 neurons in hardware, and they assemble a computer on thousands of chips.
Like Google. I feel that the difference in calculation power between a private company and establishment and a private person. The gap widened. And it reduces the freedom of the individual to do something. I'm trying to achieve distributed computation on a server with 6 cores and more than a single graphics card.
In the more distant future, 30 years or less, as you estimated, it will be a challenging world to say the least. Apocalyptic like the Terminator or Matrix, or humans will learn to live alongside conscious beings they designed.
Yosef/Yeron,
First of all, there is no such thing as "free neurons" in the brain, all the neurons are in use, only the connections between them change all the time and create new memories (sometimes at the expense of old memories).
Secondly, you don't need to convince me, I am very convinced that it won't take much time (25-30 years at most) until we have computers with a higher level of intelligence than ours and that they also have a consciousness that is not much different from ours. There are those here who do not agree, but I think reality will slap them in the face.
By the way, the number of processors should not be equal to the number of neurons, because due to its high speed, one processor can simulate several neurons and not just one.
Yosef
Whoever says that if we have as many processors as the number of neurons in the brain, then human consciousness will develop - he is trying to sell something (or win grant money...).
This idea has no basis. On the contrary - there is enough evidence that this is not true. The brain of a certain worm has 959 neurons - we know that it takes much more than 959 processors to simulate the brain of that worm.
I'm sorry - but there is no room for debate here...
rival. I went with the commenters here according to their method. With the technology that exists today, using the chip developed by IBM today (Synaps), and the software system, 28,333 chips will be required in order to reach 85 billion. This is a little more than the technology can contain. If it had to contain 4,000 chips in one computer, we would reach 1.2E7 neurons, 7 times less than a human brain according to your method. Those who argue that there are 15 billion neurons in the brain - not fundamental for the purpose of the discussion.
Google understood where the wind was blowing and to my understanding developed and sold technology similar to neural networks. I have to read more to understand.
Those who claim that it is not certain that artificial neural networks simulate neural networks, or that they will be developed thanks, are on the safe side. That is, it has not yet been proven that we will get there. This is true as an example, but scientists are working hard to prove it will happen.
In my view as someone who works with neurons, it will happen. The network will be in my opinion, and it has not yet been proven to those who follow the safe example, self-aware. I work with 200-1000 neurons and achieve performance that surprises me.
There was a series of articles on the site, which we are in the first phase of 3, I do not intend to contradict it.
But when building a military computer for the NSA, and entrusting it with the lives of people and critical systems, it is appropriate not to wait and see if the computer will be aware of itself, and then decide that it is taking command. See the Terminator series with Schwarzenegger.
I will focus on the theoretical field only. There are those who build a brain in a laboratory. I gave examples only to 2 scientists: Professor Michael Elad who recently published with PhD student Verdan Papian, 3 articles that I recommend reading, under the title work locally think globally which explain mathematically, how layered neural networks (convolutional neural networks) work. The same networks, when they are allowed to decode your images in a GOOGLE robot, they do it with greater accuracy than you and me, and in 60 technical tasks where they learned from scratch, they reach higher performance than humans. IBM sells applications for predicting consumption, for deciphering lung cancer images, to the NSA it sells a supercomputer for tracking and eliminating terrorists. At the same time, IBM is developing another supercomputer,
quantum Its size in the first stage of about 900 qubits will be several floors. Each qubit will be cooled by liquid nitrogen, and lasers will perform quantum entanglement between the qubits. Also there, the amount of calculations that the computer will perform at the same time will only be an estimate of 900^2 = 8E270, exceeding any supercomputer in existence today. Besides Michael Elad Worden Papian, there is an Italian scientist Giulio Tononi, whose articles can also be downloaded, and he developed in the last decade a quantitative theory of the formation of consciousness. It makes use of statistical physics and condensed matter physics. In his theoretical work, he shows that when the amount of connections between the nerve centers crosses a certain threshold, a "consciousness" change occurs - a phase transition in physics. The network understands globally and not only in the previous scope, and a new level of consciousness is created. At least Tononi thinks so. Tononi also computationally measured our level of consciousness, and defined the necessary amount of neural connections.
Yaron,
Our brain has 86 billion neurons and the brain uses 100% of them, what you said about 1% is a myth (which by the way is talking about 10%). On the other hand, the part that is considered to be where our high intelligence resides is the cerebral cortex and in it there are "only" 16 billion neurons.
As for such a network developing humanity. I have no intention of trying to convince. There are at least 2 researchers working on the subject. Michael Elad from the Technion also published 3 articles according to an article in Hidaan, with Verdan Papian: work locally, think globally: how multilayer neural networks work. And another physicist Giulio Tononi - developed a quantitative theory of consciousness. I don't think they are on the fringes of science.
Even if 100% of the brain is used, and it has 85 billion neurons and 300,000 neurons in one chip.
Using all parts of the brain, there are still parts that function as read/write memory. That means we get burned during life.
I mean it makes sense to me that some of the neurons will remain free. For example, it is known that by the age of 26 the personality is formed and a major part of the neurons is programmed. And it is known that throughout life, the learning process programs the neurons. The process is called training or learning.
28,333 processors are required. I agree that in this case we still can't. 4,000 we can implement as a supercomputer.
Yaron:
The claim that we use only 10% of the brain (10 times what you said) is only a claim. There is no research behind it.
That means there is a factor of 100 times.
Beyond that, it is an image of a neuron. There is nothing to suggest that an artificial network of this size will develop humanity. But why let the facts confuse
Yaron
Let me correct the numbers for you a bit: the human brain has about 86 billion neurons and it is estimated that the person uses about 100% of them.
But don't let the facts change your opinion!!!
Fix if 300,000 neurons would require 10,000 processors to get 15 billion neurons.
IBM's expert software uses orders of magnitude less, and Google's too.
The human brain has 15 billion neurons and it is estimated that he uses an average of 1% of them. If a single chip has a million neurons, 3,000 processors would be required to obtain equivalent computing power. If there are only 300,000 neurons, 3,333 processors will be required. It is within reach. And Google achieves the cracking of all the information in our mail (no one checks that they are not poking around in the computer's memory. Norton Anti-Virus gives them), we are not talking about text information, with a precision that surpasses that of a person, with only hundreds of neurons. If 1% of 15 billion is required, that's 150 million neurons. It's only 500 processors.
In my opinion, if the new computer contains 4000 processors, it will be smarter than us despite all the articles about artificial intelligence here that we are tens to hundreds of years away from the singular point. I am not open to discussion. From what I see in the academy, they are not as far as they think. This computer was already bought, still in its development process, by NSA for military purposes of tracking and eliminating terrorists. Reminds me of the Terminator series with Arnold Schwarzenegger. Google is developing the exact same thing.
How come there is no comment on things that are not dark matter related. This is a huge breakthrough, which will improve the generation
The new of genetic medicine. An effort of tens of billions of dollars is invested today in drugs that stimulate the immune system.
Here it is an innovation how to improve the stimulation of the immune system, and how to prevent the cells from being attacked.
The entire article can be downloaded at
https://www.nature.com/articles/ncomms14537
Nice of them. It usually costs a lot of money. IBM may have paid NATURE to open the article to the public.
right step. IBM develops 2 supercomputers based on different technologies and sells the products to R&D companies.
Quantum computer - the first one will be the size of a 4-story building, it is expected that miniaturization will come in gradually. Neural network computer. Based on the Synapse chip that contains about a million neurons in hardware. In my academic calculations, I perform calculations with 1000 neurons and everything that is not a core i7, and an NVIDIA GPU card, crashes.
Google realized that they were left behind and are selling the same computer that they copied from IBM.
There is another super important point here. Mega companies: Google, IBM, Intel, Tesla invest in basic R&D. Huge companies of the past were run over and collapsed: HP. Only a company that has made a fortune, and with a huge pool of excellent researchers from academia wins the race. We the public sleep standing up. The money is made through us without us feeling at all. Google gives
Free services: free OFFICE system, free mail, GOOGLE EARTH for developers. the price Their robot knows everything about us.