IBM scientists will focus on building the first quantum logical bit
IBM announced that the US Intelligence Advanced Research Activities Program (IARPA) will award a multi-year research grant to IBM scientists to advance the building blocks of a universal quantum computer.
In a universal quantum computer, quantum mechanics is used to process huge amounts of data, and to perform calculations in new and extremely advanced ways that are not possible with today's normal computers. This type of leap forward in the field of computing may significantly shorten the amount of time needed to discover life-saving drugs for cancer, to discover new aspects of the field of artificial intelligence through a significant acceleration of the learning capabilities of computers, or to make cloud computing systems resistant to cyber attacks.
Earlier this year, IBM scientists presented fundamental breakthroughs in quantum error detection by combining quantum bits (qubits) with superconductivity in lattices on computer chips, and whose quantum circuit design is the only physical architecture that can be scaled to larger dimensions.
The grant is part of IARPA's Logical Qubits (LogiQ) program which aims to overcome the limitations of current quantum systems by building a logical qubit from multiple imperfect physical qubits. As part of the LogiQ program, the IBM research team will continue to search for the appropriate approach to building a universal quantum computer using superconducting qubits. By encoding the superconducting qubits into logical qubits, true quantum calculations can be performed.
The driving force of the quantum computer is the quantum bit. Many scientists face the challenge of building qubits, but quantum information is particularly fragile, and it requires special techniques to preserve the quantum state. The main hurdles include creating high-quality qubits, and packaging them together, in a way that will allow the assembly of larger structures that can perform complicated calculations while cooperating with each other - which will limit the errors that may arise from heat and electromagnetic radiation.
According to LogiQ, the success of the program will require an approach that combines several disciplines, in order to devise new technical solutions that will deal more successfully with the fragility of the quantum information resulting from problems with the integrity of the system, from errors and environmental defects.
IBM does not disclose additional details about the grant, since they are subject to the completion of negotiations between the parties.
40 תגובות
one
If you read and understand the link, you will understand that they fulfill everything you want, and more.
I didn't see the link, I'll look later.
Or even a simple memory cell like this?
https://he.wikipedia.org/wiki/פליפ_פלופ
I'm talking about something like this:
http://www.allaboutcircuits.com/textbook/digital/chpt-11/synchronous-counters
This is, for example, a synchronous counter realized by normal logic gates (built from a circuit of transistors) that know how to output a constant 0 or 1 at the output, how do you realize something parallel with the help of qubits?
I'm talking about something like this:
http://www.allaboutcircuits.com/textbook/digital/chpt-11/synchronous-counters
This is, for example, a synchronous counter realized by normal logic gates (built from a circuit of transistors) that know how to output a constant 0 or 1 at the output, how do you realize something parallel with the help of qubits?
Or even a simple memory cell like this?
https://he.wikipedia.org/wiki/פליפ_פלופ
one
Please, read and understand:
https://quantiki.org/wiki/basic-concepts-quantum-computation
Miracles,
I did read the article, and I can't say that I understood it 100%, but it quite suited me for the first option:
"In a universal quantum computer, quantum mechanics is used to process huge amounts of data, and to perform calculations in new and extremely advanced ways, which are not possible with today's normal computers."
It pretty much called to me what elbentzo explained in his first option regarding the solution of a very complicated equation by the charging Hindus so that they would fit the equation, and then a measurement that would cause the system to collapse to the correct solution.
I asked, and I'm still trying to understand how it will be possible (in the future) to **practically** use quantum bits, a simple counter that counts the score during the game.
I would love to see even a simple schematic that shows how it would be possible to implement such a simple counter using qubits.
elbentzo, if you are still here I would appreciate your reply.
יוני
Maybe read your link to the end? 🙂
According to this article, Google already uses a quantum computer????
http://www.gadgety.co.il/146172/d-wave-x2-quantum-computer/
one
Didn't you mean that you referred to the second option? This is what the article is talking about here…..
But leave it at that... the result of a quantum calculation is an observation of the system, so there is no longer a state of superposition. That is - each qubit is 0 or 1 …….
The advantage of a quantum computer is speed, but (I already said) that it does not have more computing power. That is, both types of computers will solve the same problems. The quantum computer is faster, and if you want, you can expand on the issue of complexity (BPP vs. BQP). In any case - apparently a quantum computer does not know how to solve NP complete problems in polynomial time.
elbentzo,
thank you for the answer.
You are right, I mainly referred to the first option. So I'm trying to understand, let's say they managed to build a computer as you described in the second option, and I want to build a simple game on it, and as a start I want to realize only the score counter of the game.
Could you briefly describe how, using quantum logic gates that are every moment simultaneously in both states 0 and 1, it is possible to realize a counter that knows how to count a score during the game?
Another question, does such a computer (as you described in the second option) have any advantage over a regular classic computer? (For example if you want to implement a neural network on it, or a game that consumes a lot of resources and memory)
elbentzo
I didn't know the type you described first. Is it like using soap bubbles to solve problems of minimal areas, or like an old analog computer?
Could it be that not everyone here is talking about the same quantum computer?
"Quantum computer" today refers to two different things. One, is a quantum system that can be implemented to perform dedicated computation. A simple classic example - if I want to solve a very complicated equation, I can engineer a system of charges in which the charge distribution satisfies the equation, and then let nature solve the equation for me. I built the set-up in such a way that I was guaranteed that the equations would hold, and then I measured what was happening in the system to know how it was solved. The same can be done with quantum systems.
The second use for a quantum computer is to build systems of qubits - just as a computer works on binary logic gates (0 or 1), you can build quantum logic gates, which have only 2 options. In this case you can do with the computer everything you can do with a normal computer. The only difference will be that instead of some resistor that does or does not pass current (the classic "bit" in the computer) there will be some spin that is in a certain quantum state (the quantum bit). As Nissim said, the translation from the machine to the language will be different but the end result is the same.
one
Please try to understand what I am writing. Let's take it slow. Do you agree with the claim that all computers are equal? This is not my claim, it is the claim of Alonzo-Church and Alan Turing.
Don't lie to me, I'm asking for a yes or no answer. I ask - because the term "computer" has a very precise definition, and I want to make sure that we both understand it in the same way.
Miracles,
When will you learn to have a civilized conversation with people without cursing and blaspheming and going out of your way every time you don't like something? How many times have people already warned you about this and you continue on your own?
I don't know how you deduced from me that I claim there is or I claim that you claim there is a quantum computer today or something close to it. Learn to listen, man.
All I asked was that you show me some kind of theoretical design or basic scheme of a quantum computer that could run ***in the future*** a simple computer game of the type I mentioned, something that could form the basis of such a computer, to be built in the future, when all the problems that are currently being worked on are solved.
one
What are you talking about? I never said that there is a quantum computer today or that there is even anything close to it. Learn to listen, man. The names I mentioned are from graduate studies in computer science. Unlike you, I know what I'm talking about.
Nature will give it after we become more moral and it will be after the next great war that the world will be cleansed of a lot of scum and after that it will be rebuilt
Nature will not give us this thing so easily
From the 80s of the 20th century, for about 30 years, we have been talking about quantum computing. There were problems of noise development in arithmetic operation and implementation: such as quantum entanglement. I don't know at a much higher level than you, but there is plenty of material on IBM's website for download on quantum computation. Now a kind of decision point has arrived. Whether IBM is able to produce replication of quantum bits Qubits, and also prevent the development of cumulative error is a major problem. And IBM thinks it can try. The point is that they implement Qubits in laboratory equipment that takes up an entire hall.
Therefore apparently one is right. Today they are not able to build a computer at the level of the 80's. But there is another factor that one does not take into account. If they succeed in replicating - a computer that works at the speed of light will be obtained. Even if at first it will have the shape of the ENIAC and occupy a closet, eventually it will shrink because electro-optics on chips has advanced a lot. Optical waveguides, amplifiers, multipliers, memory cells - all this can be applied in electro-optics on chips.
Right now you are right (one) - nothing can be run. If that's what you want to say - you have my consent.
But IARPA was convinced that it should stick with IBM, because it removed some significant obstacles in the calculation.
If IBM succeeds in duplicating QUBITS: a. without producing an error at all, b. At reasonable costs - we have already done the rest of the way
on the regular computer. As Witherby said in an interview with Calcalist (the most important electrical engineer in the world) who has now donated $50 million of his fortune to the Technion, there is no telling which of the future technologies will break through first, and it is clear that silicon technology - for a non-quantum computer - has been stuck for about 20 years plus or minus. IBM as a whole excels in developing the best computing systems in the world by far from all others.
Avi,
I don't find much logic in what you wrote, maybe try to rephrase things in a more comprehensible way.
Royal Intelligence After you have built a computer with the same calculation ability as a human, still to solve problems that require the creation of a new intelligence on what exists, the aforementioned computer will need an external, surprising and disruptive random trigger that is eventually added to the existing capabilities
Poor royal intelligence is not something that can be created
Thus all important discoveries came by chance
Miracles,
Instead of philosophizing and throwing names of people you learned about in the philosophy of science degree, something practical. Does anyone today know how to build a quantum computer capable of running even a simple computer game from the 80's like "Space Invaders"?
Come show me a schematic or general theoretical model of a quantum computer capable of running such a simple game.
Meanwhile you are just talking in the air.
one
A classical computer and a quantum computer are equivalent according to the Turing and Church thesis. Are you debating what I just said?
Miracles,
Instead of being arrogant and calling others ridiculous, you should read a little about the subject, a classical computer and a quantum computer are not really equal and are not designed to perform the same type of operations/calculations.
one
Don't be ridiculous, okay? A classical computer and a quantum computer are equivalent - everything that one does, the other also does. The only difference is that a quantum computer is much faster. Give me a quantum computer and tell me what its machine language is - and I can run any existing computer program on it.
There is no quantum computer today, so I cannot describe to you how to actually do what I described. But, as I said, this is not the problem with artificial intelligence.
Anything a normal computer can do. They may not succeed. They are the first to succeed in duplicating the bits without increasing the error - that it will not accumulate. The machines of calculation now seem large. If they succeed in all these, quite complex optical structures can be built in VLSI = Very large scale integration.
one. You're right. At the current level of knowledge, they are struggling with 0 and 1. If they reach the replication of 0 and 1 at a reasonable price, they will be able to do anything.
At working frequencies gallium arsenide and gallium nitride can reach tens of gigahertz (in graphene quantum tubes I understand less, although intuitively it seems to me at least the same frequency) while in silicon and germanium silicon no faster than 3-4 gigahertz. Unlike quantum computing, there is no quantum computing here. The difference for laymen in quantum computing: instead of 0 and 1, there are 0, 1 and 0/1. The problem with these wonderful technologies is the duplication in the amount of billions of gates, and less although it exists - the dispersion of the power which is the sum of the logical gates that burns the component. Neurons - produces the same amount of calculation with much less power. The neurons are more reminiscent of the intuitive mind that thinks fast, compared to the rational mind that thinks slowly and analytically. These two structures of consciousness, according to Professor Daniel Kahneman (Nobel Prize winner for economics), exist simultaneously in our brain.
I read more about quantum computing at IBM. They have taken an important step in demonstrating a reproducible QUBIT that others have not been able to but they are still in a thorough investigation of quantum computing. There the road is still not short, and it is not clear if graphene tubes for example or gallium arsenide and gallium nitride semiconductors will not achieve them in the race to reproduce an increasing number of gates.
Joseph,
I completely agree with most of the things you wrote, the only thing I don't understand is what I asked you and Nissim and you didn't answer me.
I suggest that you read a little about how a quantum computer works and what kind of calculations can be performed on it, as far as I understand you won't be able to play even a simple computer game on such a computer.
Quantum computers can only work in certain subjects, for example certain fields of research or cryptography, as the article says. They have no approach to true artificial intelligence. What's more, the whole idea seems ridiculous to me. The way a computer works is completely different from a human brain. And also the idea of a technological singularity. Already today we are talking about the end of Moore's Law...
Miracles,
Leave you with mathematical sentences where we will talk practically, and let's please stay with the worm and not jump to a human brain. Suppose you have a perfect simulation of all 302 nerve cells of the worm and all the connections between them and the simulation includes all the dynamic changes that occur in the network when the worm learns, and you connected the network to a robot (in the form of a worm) and it behaves one to one just like the real worm!
Now come and explain to me practically without philosophies and without mathematical theorems, how you take the fastest quantum computer in the world (let's say we have already built one) and implement on it the same neural network that you were able to implement on a classical computer.
In it you will explain step by step how you do this in your laboratory.
one
There is a mathematical theorem that says all "computers" are the same, except for speed. If we succeed in modeling the worm using a normal computer, then as soon as there is a quantum computer, we have succeeded there as well - according to the definition of "computer".
What I am saying is that our brain is not equivalent to a computer, and that we are far from building a machine that can think like a human.
In order to depict a person, you need, in my opinion, to build a "person", including his whole body. Consciousness is not software that runs on the brain's hardware.
Playing Jeffery or diagnosing cancer is a dictionary search and does not indicate intelligence or understanding at all.
If you do a survey on Google, you will see that a significant number of researchers - not the majority, believe that within 20-30 years the Ray-Kurzweil singularity will arrive. This means the combination of human and machine abilities (cyborg) or only a machine that will cross the threshold of our consciousness abilities. To this end, centers have been established in the major universities for the study of rationality, and there are scientists who model a brain the size of a rat, and there is a European Union project for the development of computerized consciousness. Look at the robots of the Boston Dynamics company that are planned to serve in the American army - and you will see that there is beginning to be human sensing behavior. I also believe that it is still just a feeling.
Speed in itself is not the only condition for the existence of consciousness (speed of processing) but it is necessary. If the calculation speed of a software does not react in real time to events, there will be no visible consciousness in my opinion. Many scientists claim that the ocean, or the Internet, or civilization is self-aware. I'm still debating. The WATSON software, which is built on theories of machine learning and pattern recognition, wins the Jeffery game in a way that we feel no difference from a human trivia player. The WATSON software is trained to identify certain cancers from images in the way an expert researcher does. This is through software. This year, IBM released a cognitive chip system with 250,000,000 neurons - which perform in the blink of an eye what neuron software performs in 20 minutes. When a quantum computer becomes available to us - and I'm not sure they will succeed in the task
We will have 10 times faster computing power by many powers. The programming languages used to program WATSON or cognitive chips are more complex than conventional programming languages. Pay attention to what I said: you will create in the first step an understanding that does not "determine the difference" between it and a specialist doctor, for example. We are almost there, already today, with regular software called WATSON and with Google's projects in the field of medicine, which does not have deterministic algorithms but machine learning and statistical pattern recognition .
When that software senses the same senses as efficiently as yours, and performs data processing as efficiently as the best doctor - I think it is possible that it will be aware of itself. She will feel 5 senses like you: sight, hearing, smell, taste, touch. Maybe more. She will come to the same conclusions as you. When will she pass the Turing test? If you will not be able to distinguish her from a human and you are not - when will you call her human.
Miracles,
I'm not sure you understood my intention... Let's say we were able to perfectly simulate all 302 nerve cells of the famous worm and the simulation behaves exactly like the real worm, do you miraculously know now how to build an identical neural network using a quantum computer? If so, please describe to me exactly how you will do it.
"Let's say and you have..." correction.
Joseph
One is right - there is no direct connection between a lot of computing power and the realization of artificial intelligence. The world is very, very far from artificial intelligence - today you can't simulate a worm with 300 nerve cells, so a person with 100 billion cells?
Yossi, you are speaking very vaguely. accessible and you now have a powerful quantum computer in your hand, how do you generate artificial intelligence from it? Do you know what neural networks are? Do you know how to implement such a thing in a quantum computer?
If IBM succeeds and I am not sure to develop a quantum computer - to convert the theories into practice, this is another approach towards a computer intelligence that will challenge ours. If the speed of data processing is becoming larger by orders of magnitude, it is possible to produce layers of abstract rules on top of each other that still run fast. In an analogy from the software world, microcode is above assembler, above C, above C++, above aspect oriented programming. It is possible that carrying out a larger amount of processes will produce in the first stage a difference between her and a specialist doctor, for example. In the second stage she might (I don't know) be aware of herself.