 first talk in 10th Paa at Char 2017. Einstein once described it as a spooky action at a distance. Now we're building communications networks with them. We're building communication networks that could provide a fundamentally secure way to communicate where the privacy would be built at a physical level. And this sounds really awesome and interesting. So here to present an invitation to help her us build a quantum internet is Stephanie Werner. Please give her a very warm round of applause. Well, thanks for the nice introduction and thanks all for coming here after dinner or before dinner or after drinks or before drinks. So before I start with talking about the quantum internet which I guess from the first slide might indeed seem a bit spooky, I want to ask you a question. Namely, I want to ask you what happened on the 29th of October, 1969. Any takers? Star of the Internet, very good. We have experts, we are of course at Charon. So in 1969, the internet I guess as it's known today was slightly smaller. And in fact, I guess it was this time it was called ARPANET and it consists of only four nodes connecting a few universities in the US. Namely, UCLA in Los Angeles, Santa Barbara in Stanford and the University of Utah. So like on this day, these guys in fact made the first transmission over the, I guess this early packet switch network. You can see that they're working very hard. They're at 9, 10, 30 in the evening. They tried to kind of communicate over this network. And I guess as the story goes, this communication was not so easy at this point. So they in fact wanted to call each other on the phone in order to verify that the transmission worked correctly. So Charles Klein was down there in Los Angeles and he called up to Stanford to verify whether the kind of packets actually arrived. So they call each other on the phone and they type an L and they ask, did you see the L? And they get an answer and he says, yes, we see the L. It's excellent, we've just kind of transmitted the first message over what is now with the internet. So they typed the O and they asked on the phone, did you see the O? And reply comes, yes, excellent, we did see the O. So they typed the next letter and in fact, the system crashed. So at this point in time, like sending messages over the classical internet was not so easy. Of course we know that it works something like this and so this is kind of our what's that group at UTEC and we can now quite confidently kind of send messages over the classical internet, even pictures at any point. At day, right. So what I want to talk now is, I guess, let me say taking this to a next level, namely instead of kind of talking about the classical internet which I guess you're all familiar with, I want to talk about the quantum internet. So this is some effort actually that we have in Delft, of course there's other people working around here. So this is I guess the effort that requires of course a lot of people to be involved. So at UTEC kind of one of our goals, in fact is to build the quantum computer and one of our goals is to realize the quantum internet. So these are some of my colleagues, so Ronald and Tim who kind of work in the lab, they kind of make hardware and David who's one of my theory colleagues and with myself and we also have an engineering team at TNO which is very essential and in order to realize this. So we're currently working quite hard to kind of achieve this 2020 goal of actually setting up a demonstration network here in the Netherlands. Incidentally, I guess we're aiming it to have four nodes as per convenience and connecting Delft where of course we are the Hague which is recently closed and Leiden and Amsterdam. So we are of course not alone in the work, so we are part of the European Quantum Internet Alliance which kind of contains some of our collaborators who kind of help us realize this goal. So of course you might now be asking the question so what is a quantum internet and kind of what makes it different from a classical internet? So I'm painting here sort of a cartoon of a quantum internet that just like in the classical network you can imagine that kind of the quantum internet connects computers but of course in this case not classical computers but actually quantum computers by kind of optical communication. So kind of the aim of a quantum internet is to enable quantum processors to kind of exchange qubits potentially over long distances. In kind of exchanging qubits, they might also sort of create entanglement between them as per this purple line. So the ultimate, I guess, dream or the big ground vision of a quantum internet is to enable quantum communication or the generation of entanglement between any two points on earth. So you're probably getting a bit worried about this given that you've sort of read about the quantum computer and you're thinking, oh yeah they're building a quantum computer and now also the quantum internet and she's telling me that like making a quantum internet is like attaching quantum computers to optical communications. First I need the quantum computer and then I also need to connect it, it's usually complicated. So kind of the nice thing maybe about quantum communication is that it derives its power, I'm gonna say a bit more about this from quantum entanglement. Of course quantum computers also do that but the quantum computer kind of becomes interesting as long as soon as kind of you have a number of qubits or quantum bits, then you can no longer simulate efficiently on a classical supercomputer because you know why you get a quantum computer if I can run it on a supercomputer. So in contrast to that kind of all applications of a quantum internet derive their power from entanglement and already two qubits can be entangled with each other. So it actually turns out that in order to do something useful with this we don't need very large quantum computers meaning very many qubits but actually relatively modest size quantum computers are already sufficient. So let me maybe say a little bit on why one might actually want to do such a thing. Apart from the fact that of course it might be cool to generate entanglement between any two points on earth. So there are of course some applications and I guess probably the most well known which maybe here I don't need to explain in such great detail is the fact that we can use it to communicate securely and I'll say a bit more about this in a second. But even though of course everybody usually thinks about secure communication as in fact the introduction already highlighted there's actually very many other things that one can do using quantum communication. And yes, I know that we are at Char so probably secure communication is kind of important but there's other interesting applications actually which are maybe a little bit less well known. So one of them for example is the fact that quantum communication allows you to synchronize clocks much more accurately than can be done using classical communication. In fact if you want to get a bit crazy it can also be used to extend the baseline of telescopes. So essentially how this works is that you have two telescopes which might be very far apart and you can think that kind of one particle of light is being actually teleported using entanglement to the other telescope where you can do an interference experiment. So this allows kind of the extension of baselines that have over much larger distances. Of course we can use it to kind of test various aspects of physics for example the gravitational effects if you care about such things. Let's just like a quantum computer can lead to exponential savings in time like it's much faster. A quantum internet can in principle lead to exponential savings in communication. So there's certain tasks which I need to communicate exponentially fewer quantum bits than I need to communicate classical bits. Now of course as I'm gonna tell you shortly communicating quantum bits is pretty hard. So you might say yes there's some exponential savings but it's like super hard to send these qubits around. So of course exponential is like a dramatic scaling. So even though I might need kind of exponentially more classical bits it might still be worth sending qubits. But if you're not convinced by this and like I said there's many more applications actually that are not mentioned here that are related to synchronization and coordinate nation tasks and distributed systems for example. And if you're not convinced one can even use it to cheat certain games. So people have proposed that two players who want to play a card game like bridge actually even though they don't communicate they can have entanglement they can kind of cheat this game. So maybe there's some potential for the future like you know Char in 2030 and hacking using quantum entanglement. Possibly. So of course you know like just like classically you can use quantum communication or kind of communication to actually access computers also to link computers. You can also use quantum communication to either sort of wire up small quantum computers into a large quantum computing cluster to be more powerful or you can use it in fact to access and quantum mainframe. So I don't want to kind of make a statement of course that quantum computers will only exist in the basement but you can think that in the at least somewhat near term future quantum computers I mean will exist not everywhere sorry to disappoint you in your house but they will probably exist in very few places like our lab in the near term future going back to some area known as mainframe. So you can imagine I have a quantum computer in the basement you only have a very simple quantum terminal on your desk and nevertheless you want to run some kind of interesting computation on that mainframe. So it turns out that in fact it is possible to kind of compute something on a quantum mainframe in the basement in such a way that you don't need to tell it at all what you're computing. So it's possible in fact with information and theoretical security to perform a communication on a quantum device and without informing this quantum device what you're actually doing. So if we kind of want to access quantum computers securely in this way in the future and then we also kind of want to send qubits around. So that's all very well you know there's some applications but the question or that might be on your mind is like why can I not do these things using classical communication? Or rather kind of what makes qubits so special that kind of we can achieve all these kind of tasks while kind of classical communication is less powerful. So of course you've probably seen all these talks right you're probably waiting for me to say that kind of qubits can be both zero and one at the same time like they can be in some superposition a particle can be both left and right at the same time maybe a thought but there's a lot of videos out there that will tell you about this. So it is true and if qubits are quite special they can be both zero and one at the same time but what really kind of gives quantum communication their power are essentially two features of quantum entanglement. So I want to tell you about these two features of quantum entanglement which kind of intuitively at least explain or give you some feeling for what kind of quantum communication is good for and why. So let's imagine that we are kind of have Alice and Bob we like to talk about Alice and Bob all the time and let's suppose that they actually have already two qubits kind of which are depict by these boxes and they are actually entangled with each other. So for example, such two entangled qubits can be created by Alice having two qubits entangling them and sending one over to Bob over the quantum internet. Okay, so now they have two entangled qubits and entanglement has a very nice feature so this is feature number one, namely that even though it does not allow faster than light communication, it allows something which is called maximal correlation. So let me explain what that means. So let's imagine that Alice and Bob are possibly far apart to make this maybe a bit dramatic. Let's imagine that Alice is here on earth and Bob is on Mars and they have these entangled qubits. Alice has one and Bob has one. So let's imagine that they make a measurement on this qubit. For example, they measure and they want to know is it say red or green? So measurement that has two possible outcomes and the cool thing is that even though this measurement outcome does not exist before, so these measurement outcomes are actually randomly generated, if Alice measures and she observes that the particle is red, then Bob, even though he may be arbitrarily far away, also observe this particle to be red. And similarly, if Alice sees that it's green, then Bob will also see that it's green. So remember that I said that these are kind of randomly generated, so they couldn't know this ahead of time, okay, this does not yet exist. Nevertheless, whatever outcome Alice gets, Bob gets exactly the same outcome. And this in fact is true for any measurement that they might make. So not just if they ask if it's like red or green, but for example, if they were to ask, is it like pointing up or down? If Alice sees that it's up, Bob will always also see that it's up. So this is feature number one of entanglement, okay, which is called maximal correlation. And that's actually already pretty cool because it's sort of at the heart of why entanglement is extremely good for tasks that require coordination, like synchronizing clocks, for example, or also sort of synchronizing things in distributed systems, for example, like agreeing on a bit and various other tasks where coordination is kind of the essence. So having entanglement kind of allows you to coordinate actions much better than can be done classically. So let me now go to feature number two, which is actually pretty cool. So not only does entanglement allow maximum correlation, but entanglement is also what we call inherently private. So let me explain what this means. So if you have Alice and Bob, you might have asked the question, well, you know, Alice has a qubit entangled with Bob's qubit. Maybe, I don't know, there can be a third qubit held by someone else that is just as entangled with these qubits as Alice and Bob, okay? So maybe it could be that his qubit is just as entangled with Alice's qubit as is Bob's qubit, meaning that they could all be sort of like maximally coordinated at the same time, for example. So the cool thing about entanglement, actually, is that the answer to this is no. So it is not possible, in fact, for entanglement to be shared. Like if Alice and Bob have two qubits and they're maximally entangled with each other, it is kind of physically impossible for anything else in the universe to have any share of that entanglement. It's pretty cool, yeah, I agree. So this, in fact, follows from all kinds of physical principles. For example, if this was not true, you could kind of send information faster than light and physics would actually look very different. So this is kind of a fundamental feature of quantum entanglement that does not exist in the classical world. So note that this feature of entanglement actually already also tells you that you cannot just copy qubits. If you could copy qubits, you might imagine that Bob can make a copy of his qubit, give it to his friend, I don't know, turn it to a dark side. And we would have three qubits which are all equally entangled with each other. So entanglement cannot be shared. This is known as the monogamy of entanglement. If two qubits are maximally entangled, nothing else can be entangled. So this has some very kind of profound consequences if you think about it. Particularly, it means that if we could somehow test whether our qubits are entangled with each other, we know actually that we are inherently private and nothing else can be entangled without a qubit. So the question is, is it sort of possible to make a test that checks whether we are sort of producing entangled qubits or whether maybe there's kind of a third party that has some share of this entanglement? So the answer to this in fact is yes. So this has been kind of discovered actually by John Bell already quite a long time ago, somewhere in 1963. And he observed that kind of if you have entanglement, you can observe certain statistics, certain correlations between different qubits that cannot be generated physically. There's no explanation for this. So this means that we have kind of a means to verify essentially whether we are producing entanglement. So before I continue with this, let me maybe kind of say that by sort of checking or kind of producing such entangled qubits and verifying that they're entangled, and we can actually also see whether kind of these qubits can be described classically or not. So kind of testing for this entanglement apart from the fact that we'll have very cool applications in the quantum internet as I explained shortly, also tells us something quite fundamental about nature. Namely, it tells us that there's certain sort of correlations between distance systems, say Alice and Bob, that cannot be explained by classical information. It tells us that these kind of measurement outcomes cannot be sort of classically described by pieces of paper attached to these particles. And I just see the answer when I measure it. So such tests can actually be performed. We have performed such a test in 2005. So this is our campus, in fact. So this is kind of where our offices are. And this is 1.3 kilometers away on the other side of campus. So this, in fact, will be done. So let me now say a little bit kind of about the applications of the quantum internet. And I actually want to talk only about the most famous ones, given that it's also about SAR, and you probably also care about security. So let me now explain a little bit how we can exploit these two features of entanglement, the maximum correlation and inherent privacy, to communicate information securely. So I'm quite sure that everybody here knows about encryption, but I nevertheless want to be absolutely clear what kind of situation we are considering here, okay? So the situation is as follows. Let's imagine that we have Alice, who wants to, even after she kind of leaves Char, sort of, I don't know, communicate with Bob. So we want to transmit a classical message, and we want to encrypt this classical message. So the kind of concern here is that there's someone listening into this communication. So they want their communication to be private from some kind of eavesdropper who's trying to intercept. And I guess you all know that one way to kind of go around this is, of course, to use encryption. And encryption, of course, requires some kind of key. That's many encryption schemes. And in a classical kind of symmetric encryption scheme, for example, Alice has two key, Bob has the same key, and they can use this, I guess, just like in a safe, put the message in the safe, send it across. And if I have the right key, I can open it and recover the message. I'm sure you all know how this kind of works. But the question is now that you might sort of ask yourself, is where does this key come from? Or maybe more specifically, how much key do we actually need? So this, some people have actually thought about a long time ago. So there's this really excellent person, Claude Shannon. There's a very beautiful article in fact in New York, which I highly recommend. And he has established actually already quite some time ago now, that in order to be absolutely secure, meaning that kind of the eavesdropper genuinely gains no information. It's not like he's not powerful enough or might need a long time, but really there's no way to kind of get the message. You need a key just as long as the message. It also means that if you send several messages, like one big message, you need kind of more and more key as you keep sending messages. So this of course a little bit annoying because it means that even if Alice and Bob might say exchange some key already here at Shawan, sooner or later they really get to like each other, they want to send more and more pictures to each other, they will run out of key. So it turns out that in fact classically, it is impossible without all kinds of computational assumptions which may or may not be true, to produce key or in fact to produce more key. So even if Alice and Bob kind of exchange kind of a small amount of key here at Shawan and later they run out, there's just no way for them to make more of it. So there's maybe a question that you might be asking is why classically, is it not possible to kind of turn a small key into a larger key or make any key at all? So classically, information can be copied. This means that if Alice sends a message over to Bob and the eavesdropper can intercept it, in fact the eavesdropper can kind of read anything on this communication channel, the eavesdropper knows exactly as much as Bob. So eavesdropper knows just as much as Bob does, but Bob is of course supposed to learn the key. There must be a way for Bob to learn the key, otherwise Alice and Bob cannot change information later. So if there's a procedure for Bob to kind of learn the key, of course also the eavesdropper who has exactly as much information is also able in principle to learn the key. So kind of this is very intuitive why in fact, it is not possible classically to produce more key. So the cool thing is actually that using quantum communication, it is possible. And like I said, I want to kind of explain to you how these two features of entanglement actually make that possible. So you might be surprised that actually the theoretical ideas of quantum key distribution go back a long way in fact to the 70s, where of course people could not send qubits, like this was a totally absolutely theoretical idea at that point in time. But nevertheless, if it was observed actually already by Stephen Wiesner, that this in principle might be possible. So of course we're a little bit further along, so people have proposed schemes. And I want to give you an intuition actually about this beautiful way of looking at quantum key distribution and put forward by Arthur Eckert in 99 that exploits precisely these features of entanglement. So QKD quantum key distribution from entanglement works as follows. I'm gonna give you like a cartoon of this protocol, of course to make this work exactly is a bit more complicated. So I already told you that kind of entanglement is very useful. So let's think about how we're going to produce use entanglement to make key. So kind of a cartoon version of this protocol works as follows. Alice produces two qubits which are entangled with each other and sends one of them over to Bob. Of course somewhere on the way maybe the e-stoppers intercepting it, we don't know. So they're gonna do this very, very many times. Remember that we know that Alice and Bob can check whether their qubits are entangled with each other. So what they're going to do is they're going to send very many qubits across and some of these qubits they're going to test and they're going to ask, are these qubits still entangled? Yes or no? In fact, like I said, things are a bit more complicated. In fact, one can test precisely how entangled they are. But Alice and Bob are going to perform this test. So if the test succeeds, then they're going to measure their entangled qubits to produce a key. So I've given you these two properties of entanglement and I wanted to now use them to explain why such a protocol might actually work. So can someone remind me what is feature one of entanglement? Yeah, so they're inherently private, exactly. So this was one of the features. What is the other feature? Correlation, okay, very good. So note that any kind of key exchange system before we even start caring about security, it certainly must make sure that Alice and Bob have the same key. Otherwise, they cannot communicate. So the scheme should be correct, which means that they should have the same key. So let's remember that entanglement kind of achieves maximum correlation. This means that if they measure the entangled pair, they get a random outcome, either red or green, but they both get the same. So if we were to use this outcome as a key, certainly we've achieved task number one of any key distribution scheme, namely that they actually have the same key. So I was reminded, in fact, that entanglement is also inherently private. So this means that if Alice and Bob actually have entangled qubits, then we know that nothing else in the universe, in particular any eavesdropper that kind of tried to intercept this entanglement, can have any part of this entanglement, meaning that also the key is completely private, because nothing else can have share of the entanglement. So this is kind of how these two features of entanglement lead to security of quantum e-deceivation. Of course, I mean, things are not so easy in practice. In practice, there will be some error on the channel, there will be somewhat entangled. So they have to kind of do some error correction and private implication. But this is really the sort of essence or the intuition of why quantum key distribution works. And we again see here why also like classically this cannot be achieved. So this, of course, you can achieve. You can always achieve maximum correlation. I could, for example, just send the bits to Bob. Everybody knows them, Bob knows them, eavesdropper also knows them. But kind of entanglement ensures that this is also private. So I've been telling you that with entanglement, one can do all these cool things. So you're possibly asking actually, what is the state of the art? And I actually wanted to start with something very recent, which has done by a Chinese group, which is in space, which has generated entanglement of the distance of 1,203 kilometers via satellite. So this is an amazing achievement of engineering. So you can imagine that just like in a classical internet, we will have sort of ground-based networks to connect nearby areas, say, Netherlands. And we will have satellites to sort of bridge very long distances or longer distances than on Earth. So this, in fact, can be done. It is relatively slow. So entanglement is generated at the rate of one hertz, roughly one per second, over such a large distance. So on the ground, in fact, communication is relatively mature at short distances. In fact, if you want to do quantum key distribution and you only care about short distances, you could buy one of these boxes, like made by my IDKontique, Huawei or Toshiba. You can install it and you can plug in some fiber and generate keys. So the kind of grant challenge of quantum communication is to, in fact, bridge these long distances. So if you Google around on the internet, you may also learn that there are sort of ground-based networks when they might kind of indicate, you might think that they're already very large. For example, from Beijing to Shanghai, I guess it's not so visible on the slide. But these currently do not yet allow the end-to-end QB transmission. So these are called trusted repeater networks, just with the trivial extension of what I just described. So let's say that Alice and Bob are kind of too far away to kind of use one of these, I guess currently commercially available QKD systems, but they can ask the help of their friend. So Alice could kind of make a key with her friend. Bob could make a key with this guy. And of course, as long as they trust him, they can then kind of do end-to-end communication. So this, of course, is some nice engineering and kind of allows you to test certain deployment. But of course, this is not what one wants to achieve so in fact, true kind of quantum communication from end-to-end does not need the trust of an intermediary person. Because I guess we all know how that ended. So the question is sort of like, given that we can do quantum communication at short distances, how can we go beyond that? And of course, you might ask the question is, why is long distance quantum communication actually difficult? And there's essentially two reasons for this, one of which which I've already alluded to, namely that it is fundamentally impossible to kind of copy arbitrary qubits. So this is kind of a problem because it kind of means that we also cannot try again. So repetition is ruled out as a means to correct errors. Currently I'm wearing a microphone, usually I don't talk so loud, but if you don't hear me, I can just repeat the information, I can resend it and I can overcome losses. This is not possible using quantum communication. If I want to send a qubit and I lose the qubit, there's no way that I can sort of have copied it and resend it. So the second reason, so this is kind of a fundamental reason, the second one is a technological reason, in the sense that it's very difficult, or I guess currently still very difficult, to manipulate many qubits at the same time. So if you're kind of following the news on quantum computing, like the sort of largest general purpose, like for any kind of computer is only 17 qubits. So this also means that we cannot do large-scale error correction. So it is in fact possible, just like in classical communication, that even though I cannot copy qubits, I can sort of dilute them, I can sort of add some redundancy, such that if some of these qubits are lost or destroyed, I can nevertheless recover the original qubits into error correction. So of course doing error correction means that I need to manipulate the entire error-correcting code, I need to perform such an error-correcting encoding and I need to decode. So of course it's desirable to sort of avoid large-scale, the need for large-scale quantum communication to bridge large distances. So how can one nevertheless kind of hope to bridge large distances? So there's a very neat trick to this, which is based on something called content teleportation. I'm sure many of you have heard about this before. But in this context gives a very cool feature to actually transmit qubits. So how can we actually transmit qubits? So one of them of course is to by direct transmission. I have a quantum channel, for example, a telecom fiber, and I send my qubit using a photon. I send the particle of light. Another way to send qubits is by using teleportation. So let's for the moment imagine that Alice and Bob already had entanglement. For some reason they've already managed to have two qubits that are entangled with each other. And we're also going to allow them to communicate classically. For example, over the standard internet. So quantum teleportation allows us to use this entanglement to now send the qubit. So how does this work? Alice is going to perform a measurement on say the qubit that she wants to send and the entangled qubit. She's going to kind of get some kind of measurement outcome which is going to send to Bob. So Bob still has the qubit, like from his end of the entanglement. And based on the measurement outcome, he can apply correction information and recover the qubit. So you might say, you know, this is all very great. Why do I care about that? Because first of course, where does the entanglement come from? First I need to make the entanglement and I can teleport or what have I gained? So the cool thing about first making entanglement is actually two for it. First of all, note that if I produce entanglement, I know exactly what quantum state I want to do. I can build a machine that tries in producing its entanglement many, many times. This also means that kind of if I send kind of entangled qubits, I can now try several times. So repetition is again kind of allowed as it means the correct error because I know exactly what I want to do. So I have a machine that produces entanglement all the time. Once I succeed, I can teleport the qubit. So there's one cool feature but the other cool feature can actually be used to bridge long distances and this is what is called a quantum repeater. So let me give you a little bit of a cartoon here. So every one of my kind of blue circles is a little quantum computer, okay? So this is the Alice's. This is the repeater station in the middle and this one is the one of Bob. And let's now imagine that kind of using these colored balls up here. Alice and this in the middle repeater station already have made entangled qubits. For example, because Alice is very close to the repeater station, she prepares two qubits, sends one of them to the repeater station. So let's do this again, okay? So Bob and the repeater station are also quite close to each other. So Bob can, or the repeater station can make two qubits that are entangled and send one of them onto Bob. So now we're in the situation of Bob. So Alice has entanglement with repeater, repeater has entanglement with Bob. So note that teleportation is able to teleport any qubit. In particular, it can be used to teleport one of these qubits from the repeater to Bob. So this means that I'm kind of consuming the entanglement here, but the entanglement that Alice has with the repeater is now transferred to entanglement between Alice and Bob. So now suddenly I have entanglement over longer distances. Now I have entanglement over distances where they could not directly transmit qubits. All right, that's great. But now we know that we have entanglement over large distance. And now we can use this entanglement to teleport the qubit across. So this is kind of the basic idea of quantum repeaters. If you first make entanglement and then you transmit qubits by teleportation. Okay, so like I said, again, this is nice because one can try again. And also the operation is actually reasonably simple. So if I can go back to my cartoon of a kind of quantum network, which really is like a cartoon, apart from these sort of quantum computers at the end, one also needs to realize these quantum repeaters, kind of in these intermediary stations that allow us to generate entanglement over long distances. So let me now say a little bit kind of about the state of the art. So there exist kind of several systems. We have one which is called envy centers in diamond. So these indeed are indeed diamonds. They're of course not natural diamonds, they are artificial diamonds. And you can think that they're kind of small quantum repeaters. So in fact, like a diamond, if it was completely pure, I guess would have no qubits, but one introduces effects into this. In fact, they're called color centers. It's what gives diamonds their sparkle. And in each of these kind of defects, and there are several qubits that can be addressed. So they have roughly six qubits and in principle one could, I guess, more. One cannot store these qubits forever, maybe right now. And one can generate entanglement between such two sort of diamonds over some distance. So this is kind of the system that we're using in the universe. There's a few more. For example, one of them with our collaborator with Tracy in the Innsbook, and there's also Chris Monroe in the U.S. There's also a few groups in the U.S. at MIT in Harvard which also have stars in the sky. So what does this look like? Okay, so that's what it looks like. There's like a diamond. It is kind of mounted on this chip, which you can see here. It was actually cooled down before Kelvin. And indeed in this diamond, there's an electron spin, and in the vicinity there's several kind of nuclear spins which are qubits. So I guess I don't want to say too much about this. But because my talk, I guess, was also called, I guess, invitation to consider quantum internet, I wanted to talk a little bit now about kind of some of the sort of classical problems which are actually not something someone should do with the quantum hardware itself, that would need to be solved in fact in order to realize the quantum internet. In fact, also to do meaningful things for the quantum internet. So I guess you all know that kind of the classical internet also does not work just based on hardware. So if I were to describe a protocol that says, you know, I don't know, Alice sends a message to Bob, then that's great. But like how does the message actually go to Bob? Somewhere in your lab, I guess it was like on your computer, there's a piece of hardware, but if you don't run anything on your computer, it will do exactly nothing. So the question is sort of like how kind of, do these kind of applications, how are they actually sort of realized on this underlying hardware? And I should say that actually this is something that is currently quite open. Currently trying, in fact, we've cooked up all kinds of sort of intermediary protocols that all talk about this. And I wanted to kind of give you only one kind of example, in fact, that already illustrates that in fact, also the kind of classical network stack that some of you all know and love kind of needs some extensions here. So let me remind you kind of what kind of IP packet looks like, of course it says like, where's the packet going to? Where does it come from? But in particular, it also has this thing down here, namely the data. So it's sort of convenient that I can here send the data together with the classical control information over the same wire. That's kind of convenient, it arrives all nicely together. So this is not possible here in this kind of quantum hardware because the qubits live inside this device and the kind of classical control information is sent over kind of a classical channel next to the quantum channel. So the first thing, in fact, that we need is some kind of layer that combines kind of these qubits with this kind of classical control information. Simple sort of first step. So it kind of defines such a thing. If you remember now how I've said we're going to actually transmit qubits, namely by teleporting them, you immediately see that kind of we also need sort of a protocol to keep track of this entanglement. In fact, I can already generate entanglement over very long distances. Maybe I swap in the background. There's some background process in my network that continuously produces entanglement, but I need to keep track of it. If you wanted to kind of go a little bit further up, you can think about kind of how do we actually route qubits around. Yes, I can send them over the kind of from one end to the other. But if you think about that we can send qubits by teleportation, it's sort of obvious that in fact I could maybe, if that's convenient, I always communicate several over a particular place long distance away. I could generate such long distance entanglement ahead of time. So kind of routing can be quite dynamic actually if I can produce entanglement. So these are all sort of questions which I guess we are currently thinking about, which I find them quite exciting and interesting, which are kind of currently in fact unsolved, but which of course are needed if you want to kind of do quantum communication over large distances in a meaningful way to actually do something interesting. So in order to do this, I guess we are still writing some documentation, so it's easier to kind of understand this. We have actually, I guess this is mainly made in Manj Ladler for you or for the community, to kind of, we have a simulator in fact, which will simulate qubits for you. You can install this on your computer, this on my computer and your computer, somewhere in the background these kind of will talk to each other and pretend to have qubits. So you can basically have it right, sort of a classical client that can communicate through this sort of back simulation and also for Bob. And you can kind of ask it to create qubits, measure qubits, send instructions to the qubits, send the qubits around to the other node. So kind of our simulator can be programmed directly in Python or Twisted, this is of course not exactly how we're going to program this. So we're aiming to actually once we release it completely publicly and to put the same interface as we would use on our 2020 demonstration effect. So I will announce this on Twitter, once we kind of have the pre-data if you're interested. And of course you can now explore sort of like actually what do I need to do on this kind of classical layer of communication to kind of control and keep track of the entanglement here, to send qubits around meaningfully and to actually realize the application. And here in the end of my talk, I also have a very shameless estimate for the somewhat younger people around here. So we do have kind of a master program at Q-Tech for excellent students. And so if you want to feel challenged and join us in this endeavor, you can talk to me. You might ask actually, what is Q-Tech? So like I said, we are a collaboration between TU Delft and TNO. So I guess as you know, TNO is kind of applied research. We have what is called the National Icon of Inevitance. So we have all kinds of support in order to realize these things. And so our goal is to realize the code computer and obviously your code in there. And we also have some partnerships down here. So if you wanna do more about this, kind of stay tuned for the release of this monochrome. I've taught a class on edX last fall, which will actually run again this fall. So if you want, you can follow it online. It will go online on the 14th of November. And for more information, see our website. Thank you very much. Thank you very much, Stephanie. We still, we have a few 10 minutes left for Q&A. We have 10 minutes for Q&A. So if you would line up at the two microphones, I will call you up. So first question, please. Wait a second. Try again. But we can't hear you on the stream, sorry. Okay. So the first one, the one in the front is not working. So please line up at the second one. It seems that the security of the system depends on being able to only entangle two qubits. Is there some physical property that prevents you from entangling more than two in such a way that would allow a Alice to share the qubit with the dark side? Yeah, so like I've just explained, so like it is in fact physically impossible for three qubits to be equally entangled as two. In fact, if this would be possible, one in fact could send information faster than light, clone qubits, actually do all kinds of things, which are quite magical. And in fact, there is a test, namely this belt test, that will verify, that will only kind of succeed if two qubits are kind of sufficiently entangled that I can rule out the entanglement with the third qubit. So like basically you do this many times, you kind of check on a kind of subset whether they're entangled, meaning that I rule out that a third qubit is equally entangled with that. So this also means that like if the e-stopper, for example, who's in fact allowed to have a quantum computer and arbitrary large quantum memory, in fact allowed to control your entire rest of the universe, has sort of tried to intercept part of this entanglement, one will detect that. In fact, I should, I've talked here about always like the extreme points, like you know, maximal entanglement, no entanglement. From this test one can essentially measure a number, and from this one can make in fact a very quantitative estimate of exactly how entangled this other qubit can be, and how much information the e-stopper can have gained. I have a question, like six qubits, you said actually two questions, but six qubits, is that equal to like six bits of information you can send at one time, or would it be more information? So like at one time, like of course you can send kind of one qubit at each other, like these qubits are actually mostly useful as a memory and a local processor. So like you can of course keep sending more qubits, but it means that you can store six of them at any one time. How fast is it gonna grow in the current future? Cause now you're at six qubits, is it going quickly, or are we kind of stagnated at six? So I mean, there's a question of priorities here, of course kind of in the quantum computing domain, like you want to have more and more qubits, in fact like in for example, superconducting qubits, you can already have 17. So their qubits are actually growing quite fast. So note that the kind of qubits that you want, or the kind of hardware for computation and communication is a little bit different. So for computation you want to do very many gates very quickly, and of course you might also want a large number of qubits, but you also don't need to store them for very long. So if you want to do a communication task, it's less important that you have many qubits, because a lot of these tasks actually need only one. But you also need a system where maybe the gate speeds are not so important, but you want to store for a longer time. Namely, because you might have a protocol that requires Alice and Bob to sort of exchange classical messages, so you need to wait until that message arrived before you can kind of continue. So this is actually why here, the objective is not to kind of go to very many qubits, but rather to increase the storage time. Yeah. Next question, please. Hello. So basically as I understand quantum things, there are some strange type of ways. And when you tell me that the quantum error correction requires multiple qubits, I think multiple waves, and I think a Nyquist probing theorem, that you need more and more resolution than more error correction you do. Isn't this prohibitive if you want to send large packets with error correction? Yeah, so not that actually like here, like one is really only using some extremely small scale error correction. Because the majority of the errors is actually not kind of corrected by errors, but by repetition. So, but let me transfer your question to the computing domain, where actually one wants to kind of use large scale error correcting codes. So there of course, like it is quite challenging to kind of build large scale error correcting codes. To be honest, like from a totally practical perspective, I would actually be less worried about sort of the theoretical existence of large scale error correcting codes, kind of I think they can probably be realized. But again, like here, just like in the communication domain, there's a lot of need in the computing domain for the development of say classical control and algorithms to actually perform that error correction. So one of the reasons in fact why it's difficult to do say large scale error correction is not necessarily just the number of qubits, but it's the fact that you need an error correction algorithm that very quickly decides what to do and kind of which qubits to correct. So this is in fact one of the main challenges, which is not so much quantum, but it's actually classical. Another question. Excellent. Hello, you mentioned that you can use quantum internet in clocks synchronization, how does it work? Because I know that now the constraint is because of the latency and do we reduce latency to zero and we get better synchronization or we have some out of the box other solution. So actually we can talk after that, it's a bit complicated to explain it, but so let me explain something slightly simpler, okay? So let's actually imagine that I want to say measure, say a relative phase shift between two things that happen at the same time. So let's say that like, I don't know, something is turning here or something is turning over there and I wanted to know sort of how much faster is this one as opposed to that one. So in fact, using entanglement, one can in fact accumulate this phase drift into the entangled state and then read it out from there. So this is actually not just used for clocks synchronization, there's also say applications where people use this for sensing. For example, to kind of, there's proposals to use entanglement to detect sort of field changes in the magnetic field on earth, I guess to find things in the ground, which may be a useful application, but this works basically by transferring the phase. So this is quite similar. If you want to know exactly how one uses this synchronized clocks, then that's more complicated but we can talk about that. Excellent, another one, awesome, yeah. How fast can you communicate between the qubits? How fast can we communicate? That is a good question. I should say that currently our goal also for this 2020 demo is not to kind of go as far, like speed is here not the objective. Currently we want to kind of make it possible to communicate long distances and I guess also connect these quantum processors. Like so kind of QKD systems at like short distances can go extremely high, they can do megahertz, but we are not doing that. We are in the regime of Hertz, okay? So we're communicating qubits per second. Okay, but, sorry, my basic understanding was that if you change one of the qubits, the other changes too, how fast? Okay, so let me kind of answer this question. So there's a difference between communication and correlation, okay? So like this happens instantaneously. So if Alice is red, Bob does not need to wait until he also sees red. And if Alice is green, instantaneously Bob will also see green, okay? In fact, it does not matter who measures first. And they will always get the same outcome. So this kind of correlation is instantaneous, which is also why entanglement is very cool for kind of coordination tasks because you have sort of instantaneous coordination. So the question is now kind of why does this, why is this different from communication? In particular, you might ask, you know, why does this not allow me to send information faster than light, okay? So remember that I said that kind of this outcome is random, okay? So like they will either see both red or they will both see green. In fact, this is random. They cannot kind of determine this ahead of time. But this also means that they cannot use this entanglement to communicate, you know, like they get a random bit which is totally correlated, which is very cool, like for making key and kind of synchronizing other kind of tasks. But they cannot kind of use this bit to communicate. If they want to teleport something, maybe you remember my slide, then we needed to send this kind of measurement outcome to the other side. And you can think that in fact, before I do that, the sort of Bob only has an encrypted QB, he does not know how to interpret this until this measurement outcome arrives and he can apply that correction information. So this of course takes time so fast and light is impossible. Great. I said Roger. Good, one quick question and one quick answer, please. Okay, yes. I actually had three quick questions. No. How big is the test environment going to be? Is it for what you can currently do or what you expect to do? Is there still a DDoS problem in networks so that- Is there a problem I didn't hear? A denial of service problem with quantum networks. And why didn't you design the Dutch network to look like the pre-internet version instead of a square, make it a triangle with a point? Yeah, good question. We have constraints. We have people who give us space to put places and we also have fiber from KPN. So we are still a bit constrained. We cannot put it everywhere. Yeah, so, I mean, of course, like we are aiming to sort of, we would like, of course, to have a larger network. But, you know, like even together with TNO, like we are a research institute doing something that no one has ever done before, okay? So currently we are trying to make something that kind of works. And, but in order to sort of like make this widespread, you would need some much more industrialized process than what we are doing now. So, I don't know, like some long-term future. Okay, so thank you very much. Once more, please give a big round of applause for Stephanie.