Ladies and gentlemen, it’s a great pleasure to be here. And I would like to thank the organizers of the Qcrypt, I think, including the important friends here. And for inviting me. And also especially I would like to thank professor Pan, he invited me to make this also one lecture in the Micius Salon, which has accumulated a lot of momentum in the popularization of science in China.
And, well, let me also mention two of my oldest friends, the legendary Charles Bennett and the Gilles Brassard. I think that is surely -- I think that, you know, with them here, I'm going to try to give a really good talk. So I think that Quantum Computing has been in the public eyes very very often in the last few years.
And actually, from the activities in recent years, it really seems that quantum computer is on the verge of becoming a reality.
However, quantum mechanics is such a foreign subject. So that outsiders, even for many computer scientists, they still think that quantum computing is a mystery. And they will have a hard time answering simple questions like, “how is quantum computing different from classical computing? ” And “where does quantum computer get its enormous power from?” And so the purpose of my talk, is that I really want to address this two questions to give you a precise answer on the nature of quantum computer and where does quantum computing get its power from? And before going on, let me give away the plot a little bit. I'm going to review who the star is in today's presentation.
And we all know that in classical physics in the 19th century, people thought that all the objects in the world, they are of two types. They either are particles, so you can think of it as a baseball, or a tennis ball. It's very hard, it bounces, it occupies particular positions. And the second kind is called waves.
For example, when you look at the water wave on the lake, or if you -- less obviously -- the light that we get, they are called “light wave” was for a good reason, namely that light was thought as being a typical example of a wave phenomenon.
So there are only these two objects in classical physics, and the physicists were pretty happy about it. They thought they understood everything about nature.
But then comes the 20th century, and scientists, physicists, certainly realize that the world is not what it seems. The classical picture is wrong. Basically, they have found out that if you look at tiny objects, like atoms or electrons or photons, they possess characteristics of both particles and waves. So the particle and the wave nature would manifest themselves, depending on in what context you look at them.
So I think that in the layman's language, that we all know the classical story of 《Jekyll and Hyde》. I’m not sure that in Chinese it’s well known. Jekyll and Hyde, one of them is that is a good guy; the other is a bad guy. They are actually of the same person. And people really found out.
Famously, Einstein, in 1905, wrote a paper, in which he established that light is not just a wave. It actually, under certain circumstances, manifests itself as a particle. So, I think that by now, the physicists, according to quantum theory, that every object in the universe actually has this dual characteristic. So every object, it's kind of having two sides. One is the particle side and the other is the wave side. And the physicists have a fancy name for it.
It's called the wave-particle duality. It’s just a fancy name for saying, that the objects in the world. They are different from what we thought, classically. They have both the wave and particle characteristics. So in quantum theory, this is one of the fundamental properties of a quantum object. It's very, very famous. And so, the wave-particle duality turns out to be actually immensely important in quantum computing.
So in our story, I'm going to show you that exactly the particle-wave duality plays the starting role in making it possible for us to do quantum computing faster than classical computing under certain circumstances. So I have talked a lot about physics, if I’ve convinced you the wave-particle duality, is something that physicists believe and it seems to be true, then you will have no problem following what I’m going to say.
So I absolutely will not use any mathematics in this presentation. And the outlines are these talk, we have three parts。the first part is really the central part of this talk. We are going to show you, why quantum computing is a radically different way of doing computation than classical computers. And we’ll trace the beginning of the concept. And then we're going to show you a particularly famous example, where quantum computers can solve problems much much much faster than classical computers.
So it is in this part, we are going to show how the wave-particle duality plays such a prominent role. And then we’ll give a fairly brief survey of the realization problem of quantum computers, just to give you an idea where we are in the engineering side. And then we'll have some concluding remarks.
So, in 1936, Turing raised the concept of a Turing machine. So that really is the forerunner of the computer as we know it today. After 1936 for many many years, Turing himself and other pioneers of the concept of computation thought, they have solved the problem of computation. They thought they have found the ideal way, and actually the only way, of doing computations in the universe. And for many many years, we also believed that.
And but they're starting from the 1960’s and 70’s, people start -- kind of really innovative people – started thinking about some issues about the nature of computation. So they want to re-examine the concept of computation, for example, how much energy, that does it have to do?
And, along that line, some people also start talking about the possibility of using quantum theory to do computation. And actually, one person who has contributed the tremendously to that is Charles Bennett, who is right here, among those pioneers. And perhaps the most influential work that influence the development of quantum computing is the one by Richard Feynman in 1981.
And he asked a question, actually, two questions that, can quantum physics been simulated efficiently but classical computer? So that's a perfectly reasonable question for a famous computer scientist. It's saying that since we use the computer so much in order to understand and to analyze physics, and the question is that, actually, the classical computer has been really, really successful in analyzing physics. And because the classical physics, they are done by partial differential equations. And the classical computers are very good at doing it.
In many many cases, the classical computers can simulate physical phenomena extremely well. But Feynman is worried about the question of that, if you think about things beyond classical physics, if you go into the realm of quantum physics. so the quantum physics and the quantum equations, they also are some sort of partial differential equations, but they are objects much more complicated than the classical variables.
And so if you want to simulate, along the line of the classical computer simulating classical physics, you’re going to use exponential time, and so on.
So, Feynman asked this question. And I think his conclusion, it’s --that is probably unlikely, because there doesn't seem to be any reasonable way that you can really get a handle on the immense number of variables in quantum equations. And then on a more positive note, he has to ask the question and now that's a very innovative and important question if we give up on Alan Turing, I think that no computer scientist is going to say that, even think about the question --“If you give up Alan Turing can you do better? ”.
But the physicists can because they are not computer scientists. And so Feynman says that, in the typical physics tradition, that they are very pragmatic. They say that, “well, let's see what we can do If we cannot do it in the standard way, are there ways that we can get around this problem, finding something so that we can get answers, get numbers?”
And so Feynman asked the question that if you enlarge the class of computers, so that the components that you use to build a computer is not just the logic or gates, but something that you could use, including molecules and atoms, if you can use those quantum objects, which are known to have very peculiar properties, in particular, the wave-particle duality are, would you be able to build computers that actually can simulate quantum laws? So he asked this question and he did some small example then, and he conjectured that, well, this might be true.
And now, let’s take a look at the difference between the classical computer and a quantum computer and the classical computer, in essence, you have some strings, you have digits, you have bits, and you have a string are as input, and you like to transform it, and you want to do computation, and you go through a classical computer.
And a classical computer is something that can perform digital operations. And now in contrast, a quantum computer is a quantum device that are built up in whatever way, with the quantum material, and you can input, representations of the input bits in the physical space. So in particular, it could be just things that represent the binary bits. But in general, it could be something larger.
But let's just stick to the idea that we are comparing oranges with oranges, so that you have something you'd like to compute. You just represent it in the physical space, and you put it go through the quantum device. And at a certain time, you perform the measurement. You look at the outcome of the object, and just see what number you get. And that's supposed to be your output.
And a critical thing that I think in a way is fun, that the quantum device -- different from the classical computer is that for many many years we thought that the analogs are out because you know the analogs were the favorite of the electrical engineers. You know they have voltages, they have currents. They do signal processing using those things.
But when the computers became popular, we thought that those analog devices should be thrown into the garbage can because that we will never have the need to use them again, because making things digital has the advantage of making things more stable and more controllable, and the analog devices are more unruly.
Okay now, but the quantum device if use things that are built from atoms or quantum materials, then suddenly we're coming back to analog devices. It’s only at the very end, when we decide that the computations are over and decided to look at what we get, then we convert it into the digital thing.
And now, what are those devices in the quantum computer is just manipulates the bits with the Boolean operations. And in the quantum device, the corresponding thing is to manipulate the qubits. So these are objects in a big geometric space. And essentially, you do rotations.
Okay, so if you look at comparing the computing devices in the classical case, you do operations on n bits. And also the space is this 2^n strength, you do mappings back and forth. And now in the quantum space, you're actually looking at the much larger, much richer geometric structure and it’s the complex number, C^2^n. And if you don't like complex numbers, just think of it as Euclidean space, except that dimension that corresponds to n bits would be 2^n: exponentially large.
So, now that's the first indication that we might be able to do something that the classical computers cannot do because now we have such much richer, bigger space. And the rotation, it could be very different from just simple permutations of the bits. Ok, so, that gives us hope.
So now basically, I have defined all the concepts of the classical computer and a quantum computer. And I’m not going to – So the rotations are technically, they are called unitary transformations in the space. But it’s going to make us dizzy, So let’s look at it and forget it. So these are operations.
And now, even though that, in principle, it looks like that we have a larger space to work with. Now, if somebody ask a quantum computer scientist, “what's the secret of quantum computing? why is it? Can you say more precise (what) the quantum computers do that makes it possible to speed up things?”
And I think here is a typical answer that people would keep in saying that. Well, each qubit represents not a single state, but in some way, a particular kind of probabilistic combination of states and the famous Schrödinger’s cat is saying that a cat could be in a state of superposition of a live cat and a dead cat. And now, this is a particular type of state of matter, so that you cannot say it's a live cat or it’s a dead cat.
But you know it has probability to be dead or alive. And it behaves in very peculiar ways. And so, when you have the power of representation to represent both possibilities at the same time, namely, represent the possibility of the cat being alive or the cat being dead. The basic operation, you can represent it in such a way. It is intuitively that gives you some sense that quantum operations have some potential for doing parallelism.
And we know that, in computational problems, a lot of times what you do is that you have many different possible answers and you would like to search through each one of them, to see which one is the right answer that you want.
If you have a parallel computer and suppose you have ideal parallel computers, you can take as many processors as you want and do a parallel search. Then in principle, you can speed it up a lot during this process. So I think that a popular way of saying that the quantum computer is powerful is that the quantum computer can do parallel search.
So there's this quantum parallelism that makes it possible. And I cannot help but to use a professor Pan’s metaphor saying that I'm explaining to the lay person that you know in Chinese fable, there is this monkey king Sun Wukong. And the monkey is possible to make many copies of himself. Sun Wukong can take a hair, and make it into another monkey, so that you can (be like) all these monkeys to search through all the possibilities. And so, in the way, quantum computers sometimes possess this magical ability so that you can do the search fast.
However, we have to keep in mind that this is really just a metaphor.
It’s certainly true that the kind of representation makes it conceivable that there is some possibility of doing a parallel search. However, in almost all the cases, when you look at a typical classical algorithm, and you will not find any way to use the quantum ability to do a parallel search.
So therefore, it's something that is not a real sense for answer by itself. And so what we would like to do is that to give you a real demonstration of the power of the speedup of a quantum algorithm. And also in the process, you are going to see exactly where the parallelism would arise.
1648
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
