Quantum Information Science - Dr. Gerald Gilbert

that's the engineer that's all isn't it okay oh well we're all set for the next talk so our next speaker is dr. Gerald Gilbert Jerry is from The MITRE corporation and heads their quantum information science group he received his PhD from the University of Texas he has a very distinguished career working with Steven Weinberg from from Texas and also with Stephen Hawking and some of his students he will describe some challenges in quantum information science thanks Jerry thank you very much do I need to attach anything to myself okay so it's a pleasure to be here thank you Margo for inviting me you mentioned quantum jokes earlier they're usually not funny but you know my office is I'm at Princeton University and that's of course in New Jersey so the job that usually isn't so funny is that unlike saying I'm from New Jersey and I have an offer you can't refuse and say I'm from New Jersey I have a presentation you can't understand so we'll see how it goes is this the right thing to push the button okay so you mentioned I'm a theorist not an experimental yeah okay gotcha so I'm gonna give an overview of the field that will hopefully fill in perhaps a few gaps and it's always a good idea in any event to repeat complicated ideas the overall field of quantum information science or QED is is built up out of three parts that are somewhat in their typical interdependent these three parts are quantum computing sensing and communications the three parts are not the same but they share each of them two aspects in common besides the obvious linguistic linguistic fact that the word quantum is in each of their names they each solve a problem set that are otherwise either physically impossible to solve or effectively practically speaking impossible to solve that's one of the two things that they share in common and the other is that they achieve this by us in our design of equipment that does this directly exploiting one or more features of the physics of quantum mechanics in the design of the hardware and in the design of the algorithms in these three fields I could go on at great length about interesting challenging problems that there we already know solutions for that are otherwise not possible to solve in the classical context but I'm just going to give a couple of examples these are probably well known to everybody here but in the case of quantum computing we know that there are problems of a mathematical complexity such that we cannot solve them using classical techniques the most famous so-called killer app that everybody must be aware of I'm sure is the shor algorithm that is used to break cryptographic codes I want to stress just as an aside that the shor algorithm is not known it's not that we don't have a proof that there is no classical method of doing what it what the quantum computer implementing Shor's algorithm does there are quantum algorithms that and solve problems if no classical machine can solve for which we do have such a proof but this famous killer app the shor algorithm isn't thus far like that we don't yet have a proof that it's not possible although it's probably not possible quantum sensing is putting on the best pair of glasses that you can put on so I need to put on reading glasses to read quantum sensing achieves the best possible image resolution that is allowed by the laws of nature as we currently understand them and quantum communications is a bundle of protocols the one that's been mentioned most often thus far today is quantum key distribution or quantum cryptography these two phrases are this are synonyms the proper name given what is actually done is quantum key distribution but it's too late the the wrong name quantum cryptography has taken hold and is done it's not going to undertake so those are two synonyms that mean the same thing and with proper explanation as to what it is that you mean it provides unconditionally secret method of communicating I'm going to go into that in a little more detail it was also stressed in the last talk but I want to talk about it the unifying feature that relates all of this which is the generalization of bit to quantum bit and what underlies that is a connection between physics different parts of physics and information theory so there is a syllogism here which seems almost like it's content list but it's very important and it's not content less information is necessarily encoded in the state of some physical system I mean if we want to communicate with each other or do anything with each other we have to have a physical object at hand to do it we can't just stare at each other so there has to be a physical thing and the behavior of physical systems naturally governed by physical laws determines what's possible but the underlying basis of those physical laws today in the year 20,000 at 2018 is quantum mechanics and we don't know of anything we don't understand anything we humans don't understand anything better than quantum mechanics we don't listen you can't think of anything the history of Bolivia and the seventeen the price of eggs in Kazakhstan there's nothing that we understand better than quantum mechanics there's no experimental deviation from its predictions that have thus far been verified so we therefore it's as if physicists can say I believe it it sounds like a religious statement but I believe quantum mechanics is correct and so that actually describes what physical objects do and if we impress onto the physical object the information that tells us what information can do so you've seen this chart before but the thing on the left is a bit it's in this case a capacitor well my hand is a bit if I hold it like this it's a zero and if I hold it like that it's the one any physical thing that I can flip between two conditions is a good candidate for a bit there's nothing sacred about the number of - number two the founding fathers of information science could have chosen base 17 and they would have had they been sadists but but base 2 is as good as any of it any other in terms of theory and in terms of ease of thinking it's the best one so we are stuck with base - but it's a good thing to be stuck with so anything that can be - valued is a good candidate for a bit a classical bit and that's all there isn't there ain't no more a cup this way or that way is a bit I don't really have to think too much well if I'm Claude Shannon in the year 1942 I don't think too much about weather turning the cup upside down to make it a zero or one does anything special because it doesn't it's just a cup or it's a hole or a not hole in a punch card or a tube that's turned on or off I don't care about the physics of this so here's my bit I have charge or I don't have charge on the capacitor but you've seen this sphere before it's named after felix bloch who was one of the great physicists of the 20th century and you don't need to worry too much about the details of the mathematical appendages hanging off of this sphere but it's a mathematical representation of a physical object that is what is called spin 1/2 dead as marco pointed out earlier there is intrinsically connected to the theory of relativity and that would take too much time to go into but this fear and we're an arrow points on it tells us what is what what what are part of this state of a spin 1/2 physical system and I can have it and the Tyvek it's a vector the characteristic I'm talking about and it points somewhere I can have it point up I can call that direction up if I want and I can identify that as a zero and I can have it point down and I can call there one but I can have it point here and this is a really important quantum information departure from classical information it's not just the the tilt from the pole that tells you that you have some kind in some suitable sense of understanding this and admixture of 0 and 1 it's also where it is in this angle as well going around the vertical axis that phase is not describable in a classical mechanical bit it doesn't it's not part of the classical mechanical bit and this thing can point anywhere and we now have the technological capability to swing that arm around anywhere we want that's something we could do 70 years ago but we can actually take a physical system and make that arrow point somewhere so there's a kind of in a suitable sense a continuous infinity of possible values that this bit can store when we measure it it always collapses into one or the other of two possible outcomes but before we do that measurement it's like this I'm gonna have to condense a lot about a year ago if you look closely it would look like it's less than a year because the date that's on this paper on the right is just from a few weeks ago but it was it this is the third edition of this paper John Prescott my former colleague from when I was at Caltech came up with a brilliant you know name disk noisy intermediate scale quantum technology what he was trying to do in describing this was to capture where we are in the development of this field the key to making any kind of quantum technology is to have a clean error corrected fault tolerant collection of quantum bits and I'm going to talk about whatever correction fault tolerance means in a moment but we're not there yet for any any reasonably large number of these things but we're very close to being in a regime of actual building where we can have of order around a hundred quantum bits and so and they may or may not be perfectly error corrected so that's why he calls this regime noisy intermediate scale the thing that is important for us to understand is that when we reach this value like for example google google has publicly announced they have a 72 quantum bit device you can show theoretically that there are mathematical problems that are already impossible to solve with that number of qubits with any classical computer so we're embarking now on a new era we're in it now of being able to actually develop hardware to solve problems that we don't know we don't know where it will take us well this is what john stressed in his paper from almost a year ago what I want to stress is not so much the quantum computing aspect of this but the quantum sensing and corner communications aspects as it turns out and there's no guarantee that this is going to be true for all of them but most of the existing known quantum communications protocols and many of the existing quantum sensing protocols are different from the known interest in quantum computing protocols the shor algorithm that I mentioned earlier in order to solve a meaningful code to break a code the machine I'm not going to be able to go into it all the reasons why we'll have to have millions of physical quantum bits most likely for based on our current understanding of error correction in order to solve a national-security type code to do quantum cryptography Quan key distribution there are various protocols that have been invented but if you use the simplest one and the simplest one is as good as any of them you only need to control at any one time a unit of time in this discussion I'll call it a bit cell period one qubit we can do that and to do quantum sensing and get meaningful results that are better than the best possible classical results you can already get that if you can control around 10 cubits at a time as opposed to the millions that you need for quantum computing to give a real advantage so so we are already in the NIST era and have been for the last few years for sensing and communications and this graph this is a notional graph don't want you to look at the numbers or the that's why I made these bubbles intentionally kind of fuzzy this is approximately correct the two dotted lines there in compass sort of where we are internationally and how many cubits we can make in a laboratory and there are various ways of parameterizing where we are in the scaling of quantum technology I've chosen one way and any one way that you pick if it only has two axes is insufficient but here I have the number of qubits you can control and then what I mean by control is includes making them error corrected and the number of qubits you can make bang into each other in the vertical axis these are logarithmic axis and we were around like I said Google has about 72 Microsoft has worked about 50 that's approximate range and regime that we are many of you will know that there is a Canadian company called d-wave that has claims that they had and does have 2,000 cubits on their chip and they're using superconducting qubits I'm going to show in a moment that there are different types or flavors of qubit that you can choose you have to pick one to build your device and both Google and d-wave are using superconducting qubits and Google has 72 and d-wave has 2000 so why doesn't do will just fall down and die and so they're not the same kinds of qubits even though they're both superconducting so there's a lot of nuance that goes into discussing this and there's also more to this statement of where we are than simply counting quantum bits there's another aspect of quantum circuits called the depth which has to do with the times time steps associated with an algorithm so this isn't necessarily incomplete picture but with all of its flaws it gives us a rough understanding of where we are we are already capable of doing qkd we are and the boundary of being able to do meaningfully important it's quantum sensing and hopefully soon we'll be able to do one of the sort of two big categories of quantum computing work which is the bubble on the upper left this misc era that john prescott has identified that's what's encompassed in the bulletin deep in the bubble in the upper left the problems that are in that area may not be as romantic and dramatic as breaking the codes of somebody who you want to understand but there are very very important problems in this area that include amongst other things the potential to develop new of medicines and to understand molecular structure in a way that we cannot now I have to be clear about what I'm saying we understand the quantum mechanical equations that describe a molecular structure for anything you might want to think about it's what we what I mean by that is we know what the equations are that we know precisely what the equations are that you need to write down in order to characterize anything but we can't solve the equations except for the very simplest systems analytically and even using classical computers you get to a point quickly when you get to a sufficiently complicated such a system where you can't solve it in any efficient way using a classical machine and that's something we'll be able to overcome and in this era I want to talk about reality here for a second so I have been depressed to discover recently that as I asked people in audiences who remembers Felix the Cat that number keeps on dropping so I will disagree with arthur c clarke and excellent so this is why it's always good to have Jonathan and an audience when you ever use theme and so Felix had this magic bag of tricks and with this magic bag he could solve any problems if there was a mountain in front of him he took out of the bag of hole and he put it on and walked through he put it back in his bag this is not what we have in quantum information science I disagree with arthur c clarke characterization of any sufficiently complicated technology being indistinguishable from magic it's not you just have to take a class in physics and then you see it's not magic and so this is hard-nosed calculable physics physics is different from philosophy or god help us politics you sit down with a piece of paper and a pencil and you calculate and it doesn't matter whether you're in Alaska or Madagascar if you get the same answer you've done physics and maybe see maybe somebody here knows what tonne stovl is from highland good on line so there so there ain't no such thing as a free lunch and that's certainly true in physics in particular and in quantum mechanics especially it's it's science we know what is walk there are boundaries you can't just wave your arm and make something bump even though entanglement at first looks yeah so I'm gonna tell you a little bit of what entanglement cannot do but rather than being vague I want to show you how we actually generate entanglement I say we I'm a theorist so I don't do this but the experiment this is one way of doing it and it is the way that is being done right now in our laboratory at Princeton and you name it wherever people generate entangled photons they use this method it has a name it's called spontaneous parametric down conversion of course this several versions of it I'm not gonna go into any of the details but you have a laser and the laser radiates another laser the first one is it's like a flashlight it's a continuous wave laser the next one is like a machine gun you know bullets of light come out of it that's a the second laser there and a number of optical components are irradiated the light is allowed to pass through specially cut encoded and shaped crystals that do things to the light as the light goes in the most important of which is this this one is interesting if it's cut the right way a beam of energy a beam of light with a certain amount of energy comes in front to the back and two beams come out at a certain angle with half of the energy of the incoming beam then what's done is that we polarize the outgoing light I want to remind you what polarization is so when they look to make them feel when you turn on light like it's coming from the Sun or the light bulbs there's all kinds of vectors that you learn about in electrodynamics there's one called the electric vector which is not the same as the electric field vector or not the same as it's called the electric points moving and it traces out as a picture it's like an artist as the field is moving forward and if it's not coherent like if it's just any old light it traces out a squiggle gibberish as it moves forward if the light has been polarized which means that has been passed to a material that does this interesting phenomenon called polarizing then it doesn't do that gibberish anymore the electric vector tip traces out a picture Generale generically it's an ellipse and if you switch the ellipse this way it turns into a line and that vector goes back and forth along a line that's a linearly polarized field if you squish it the other way it the circle in that circularly polarised light now or generically it's an ellipse so let's talk about vertically polarized so I can do that I can have it just like the polarizing filters that you probably played with as I did as a kid with you you turn them into the light went black and so you can organize this line so that it's horizontal well call some Direction horizontal and then the direction perpendicular to that is vertical we have that polarization applied to the output of the crystal and what it's produced is a cone of light so this picture here is actually not at the rest of this is a cartoon but this picture is actual data taken in this case from our lab the colors are artificial but where the two cone bases overlap the photons there are entangled and that symbol there what would what it what it means to be entangled is that the two parts are not two parts anymore there there are two parts but they're not parts of two different things they're two parts of the same thing and it's not possible to talk about this complex of two photons without talking about both of them together it's probably not it's but to do this correctly you need to write down in detail exactly what I'm talking about but these are impossible to describe separately so there's a horizontal part that is tied to the vertical part and there's a vertical part that is tied to horizontal part the interesting thing is that although it's not really two things anymore it's it started off with two so we think of it as two and you've all heard about the experiment where you put one of these two pieces in in my office at Princeton and then you get the other one to Pittsburgh and so that's a trick to get there to Pittsburgh and keep it entangled but let's overlook the engineering challenge and imagine that somehow it stays entangled and there are four ways to entangle to spin 1/2 particles let's one of the ways is such that if I measure one of these two pointing off to the other ones they could point down so the obvious question that people think at first intelligent people think but they're mistaken is that Paul Revere could have done one of you know up if by land down if by sea you can't do that you can't communicate classical information and I want to give you a quick thought experiment our description of why you can't do that takes a little bit of mathematical precision but I can say words that are in this case actually correct I say it that way because usually a fight as a physicist tell you something it's always incorrect partly because I have to leave out the part I'm leaving out but in this case I can tell you everything in a verbal description of the experiment so John Dowling is in Baton Rouge and he's got one of these two and I'm in Princeton and I've got the other of the two and we agreed to do an experiment but we do this is a scientific process we do a lot of repetitions to get good statistics so I keep on measuring my my spins and I get up down I've done whatever I get and I keep a record we agree on a time in advance and some of our clocks are perfectly synchronized and John looks at his end and he kept keeps his record and then I play a trick on John I don't tell him but I go home and I stopped doing this he doesn't know that because I didn't tell him so he keeps on making his record that's not but I'm not really mean the reason I'm not mean is because if I do this your result is gonna look exactly the same either way the record that you would get if I had actually stuck to my guns and kept on making measurements would have been perfect white noise he would have gotten up down and around and distribution and I would have gotten the same if I in fact left the lab and wasn't doing anything you would be only able to generate the same thing at your end and it would be impossible for us to distinguish these two experimental records in other words I could have dropped dead and it wouldn't have been any different than if you can't not only in other words can you not convey information you can't convey that anything happened at all using entanglement so you might ask well if that's so true you can't do anything then can you do anything in the answers yeah you can do lots of things but you have to do more clever things than saying it's time to buy the stock or time to shoot the gun it's not as simple it's not simple to use entangle you have to come up with a non-trivial protocol but the simple things that you might think you can do you can and I want to say one other things that you cannot do you cannot do faster-than-light communication it is true that whatever it is it happens if I make my measurement in Princeton that the correlated effect that John observes in Baton Rouge does happen instantaneously what I mean by that is that the mathematical formalism describes it doesn't have a slot for this for time usually a lot for time in this case there isn't there's no way to even it's not in corporate union or the better way to say you can take the timecard and throw it away it's not in the expression it happens without the passage of any time at all so doesn't that violate special relativity which tells us that you can't go faster than live and the answer is it doesn't the reason it doesn't is because although Einstein didn't formulate it this way when he first wrote down in German the description of all of this but he would have written it this way if he had been thinking about information theory but you can this special theory of relativity does in fact prevent the propagation faster than the speed of light of an information carrying signal but in this case there's no information so there's no reason that it can't go faster it doesn't violate any of the laws of special relativity very very briefly because I'm not focused here in this talk on foreign computing roughly speaking there are three there's a fourth one I haven't included here flavors of quantum computing you can formally show mathematically that these are the same as each other the d-wave machine which has been much talked about in the press is a type of a restricted version of the second type that the second type that I've but there is called adiabatic quantum computing the new web machine is actually not adiabatic quantum computing but it's a somewhat restricted version of that that's called quantum annealing and in contrast what Google is developing is intended to be a full-fledged at ATO better for computers I want to talk about quantum sensing and I mentioned that it gives you the best possible image resolution so Lord rayel I worked out in the 19th century the formula that tells us the relationship between the resolution that we can get in an image and the important physical parameters lights wavelength and the size of the aperture through which the light is going towards the thing that it is illuminating and it's roughly lambda divided by D if you do the calculation right what you do you know class on this topic it's for good luck there's a number that's about equal to one but not exactly equal to one that you have to calculate but it's approximately lambda divided by D and that's why you have big telescopes because of this factor of D in the denominator because you want this ratio to be as small as possible this ratio tells you in units of wavelength how close to each other can two objects be and the CERN still as distinct from each other so you want that fraction to be as small as possible so for a fixed wavelength you want to make the diameter D of the telescope as big as possible in quantum in imaging that gets modified Lord Rael I didn't know this because quantum mechanics hadn't been discovered when he calculated this but he would have gotten this result had he done it there's an additional factor of n where n is the number of photons per pulse that I radiate the target that you're looking at so let's talk about that the classical result is lambda divided by D if you take into account the fact that quantum mechanics is the right description of things and therefore you don't simply model what's happening as light illuminating a target light waves but quantum mechanical description of light which means that it's both waves and a particle called a photon then there's a number of those photons and the number n per pulse appears in the denominator here but how do I go back well it doesn't matter this is it's on this slide to so then it turns out that the square root of that number is what appears if you do the calculation from scratch you get exactly what real I got including the number that I told you that's for a bit of luck in the front which is why I don't have an equal sign there it's if you then take these photons which you've now counted and you entangle them you get an additional factor of one more square root of n which gives you the final result of a full factor of n so that means that if you have ten photons per pulse that illuminate your target you will see things ten times closer together than you would otherwise be able to see this technique uses a state a type of way of entangling the photons that is called a noon state and you're gonna think he's paid me off but he hasn't dr. Dowling has in fact done the seminal work on this twenty years ago almost now on identifying the significance of noon States it's a particular type of entanglement and so using noon States you can achieve this highest possible image resolution and this is not just theory it's already been demonstrated in many experiments I've only Illustrated it here with one there was an experiment a few years ago done in Japan of course I use the letter Q they engraved it and they image did an up to some experimental errors the improved resolution that they demonstrated was exactly what the theory says it should be these quantum effects are very very real but we don't somehow think that they're intuitive I want to take a moment to talk about that with you why is that so there is a famous American rabbi who's over 90 years old now his name is Abraham torski and he wrote this very deep but simple sounding thing if you've walked on a slippery set of steps and you fall on your behind you know it's gonna hurt theoretically before you do it but there's nothing like actually doing it to feel what it's really like and so we we can talk theoretically about quantum mechanics but why don't we really feel it the same way we feel it if we fall in our behinds as we slip down a step and so there are all kinds of physical parameters that are useful in describing the world around us non physicists and physicists like to know some of the most famous ones you know the mass the speed the angular momentum and so on there's a lesser known physical parameter perhaps two non physicists that is called the action of a physical system for reasons I can't go into it has a special role in this question of determining whether phenomenon is somehow quantum mechanical or is it's acceptable to describe it using the approximation to quantum mechanics that we call classical mechanics and it turns out that the important quantity I hate saying it turns out there's no time to explain in detail why but it turns out that the quantity that determines this is the ratio of the action of the system in question to a specific step and standard amount of that action called Planck's constant and Planck's constant is so what is action action is roughly speak it has units of energy multiplied by time and roughly speaking it sells how much energy is expended by a system during a given amount of times it's a complicated integral over a certain quantity and that's approximately what it is and if you move around open the refrigerator get out of Sandwich it's one way of measuring energy and time is joules times seconds so if you open up your refrigerator or get out of a sandwich or walk or turn on your car or do ordinary things you will be expending of order 10 to the third Joule seconds of action it's not exactly that depends on how much you're doing but that's a sort of a good figure of Merit for how much action characterizes macroscopic activity Planck's constant that I mentioned a moment ago in units of joules times seconds is 10 to the minus 34 of them so the ratio of your human macroscopic action 10 to the 3 to 10 to the minus 34 is 10 to the positive 37 in other words a trillion trillion trillion times 10 the formula that you have to stick this number into is the number of order one multiplied by E the base of the natural logs the exponential function raised to the power minus that ratio so that ratio is I've just told you it's 10 to the it's 10 to the 37 so e to the minus 10 to the 37 is 1 divided by e to the 10th of the 37 which is 1 divided by almost infinity which is almost zero which in that number is a net is a measure of how quantum mechanical the phenomenon that you're talking about it so if you're doing an amount of action that we do in everyday life the corner mechanical effect that you are likely to see is approximately zero that's why we don't float through walls we don't onyl through walls but if you're an electron that same ratio was not is it it's in fact not the huge number that I mentioned it's about one because the reaction of electron is close to Planck's constant so then everything is quantum mechanical but because we are the size we are and the things we touch are the size they are evolution presumably evolved us to do well with the things in which we are embedded so what do us no good to be sensitive to things that never happened to us because of the size and scale that we are so we don't have any intuitive sense of quantum mechanical things other than what we can develop through the study of physics and a lot of exposure to the calculations so I've been doing this every day of my professional life since I got my doctorate so I have fooled myself into believing that I somehow understand this better than non physicists I in some sense I do but nobody really understands these phenomena nobody can because we don't experience them we can only do that through instruments that we build there is a problem with quantum sensing which is that we have to propagate these entangled photons through a realistic atmosphere there's a paper I wrote about co-authored with several other authors about 12 years ago now in which we calculated for the first time exactly what happens when you propagate these entangled noon States through a lossy atmosphere and there are challenges that have to be overcome as a result of the physical phenomena that occur and a lot of work has been done on this to try to get around that discussion took place earlier today about quantum cryptography I want to talk about its unconditional secrecy it is unconditionally secret if you do it correctly except for the part that isn't and professor Oman correctly fingered what it is there isn't which is that first step the zeroth order step which is the need to authenticate so I'm not gonna advocate a stance on whether one shouldn't do qkv but one argument that a people will make is that you have to do authentication always no matter what kind of cryptographic techniques that you use since you have to do that and that is not there is no currently known unconditionally secret method of doing that since you have to do that you have to do that and if you then build on top of that the rest of the qkd protocol that remained there is unconditionally secretive done correctly so at least you've gained that kind of an advantage you know this slide points mainly to the limitation of the need to authenticate and the effect that wasn't stressed so much before which is that qkd in all of its current variations there are about 20 different protocols that have the best four have been discovered is in an intrinsically point-to-point protocol so it's not cleanly and trivially internet izybelle where we can send a message to as many people as we want at the same time all governed by the same security guarantee that's not possible you have to separately do qkd with each recipient I'm just going to show one more Q gating slide real quick this shows the protocol in detail I want to show you how we might do a network so here is a mountain range which is not not geographically accurate I didn't realize that my animation had a sound attached to it I hesitate to press the button that began because there's another vehicle coming Microsoft's fault not mine Allison Bob and they are not connected by a line of sight optical paths and let's assume they don't want to have a in optical fiber stretched out between them so if if they can each see separately in a satellite they can do qkd as follows Alice can initially set up a key case of a with the satellite Bob can do the same with a different key case of B and then if the satellite is a so-called trusted node which is already in violation of the spirit of qkd because you're not supposed to rely on trust but on the principles of physics if you do trust the satellite and somehow know that there's no hacker on it or breaking into it then the satellite that now has the two keys can compute the binary sum the XOR of the two kings and that binary sum is so these are piece to the random sequences of bits so they're information less but you want to keep the key that's of importance to you away from the bit of the enemy but this other key XOR sub a B is information less like any random string and I can send it to Bob from the satellite it doesn't give the enemy any information there's no information in it and it's not the particular information with string case of a that I really care about so that's fine I can do that and then Bob who of course is in possession of key K be an XOR his key against this XOR and it's a property of binary arithmetic that we all know that if I do this double X or I get out the first term which is piece of a so now Bob finally has key sub a which he now shares with Alice and now they can do they can now do use this key as a one-time pad and there can be I'm not going to use this slide ever again network of n Bob's as many Bob's that you like this is not a proper quantum network where the end-to-end features are what they are because of the properties of the physics of quantum mechanics this is a stand-in and this is what the Chinese did publicly announced this a year ago when they did a video call between Beijing and Vienna they did this kind of a what we really need to conclude in order to have viable quantum communications and quantum sensing our quantum repeaters the word quantum repeater is the phrase is a misnomer a classical repeater takes a week so I have a signal that I want to send from Ohio to Auckland New Zealand and as it moves on its way from Ohio it gets weakened and weakened and weakened and then I rectify it and I amplify it with classical repeaters and then send it on and it gets to New Zealand just fine you can't do that with quantum mechanical States because of the no-cloning theorem that was referred to several times already and okay that's good but but it's it's also true that the phrase quantum repeater is a misnomer nothing is actually repeated what is actually done in quantum repeating is that entanglement the distance over which entanglement is shared between separated nodes is increased and increased and increased so a variety of different techniques the together are called Quantum repeating at the end of this process you now have entanglement share between Auckland New Zealand and wherever you started in New York and then you can use that as a resource to compute or do quantum communications or quantum sensing so that's what quantum repeating really is it has the net effect of what would have happened if I had been able to repeat the quantum information but like what I can't so this is what we mean when we say quantum heaters this is the required missing technological piece that is standing on our way of having large-scale internationally distributed quantum networks and I expect that this will be and we already have some quantum repeaters that have been built in the laboratories but we don't have any scalable robust devices that can be put into work in an actual communications network outside of a laboratory I expect that will come very soon so I think I'm done thank you very much [Applause] [Music] [Applause] thanks very much thanks very

Channel: The Artificial Intelligence Channel

Views: 7,540

Rating: 4.9183674 out of 5

Keywords: singularity, ai, artificial intelligence, deep learning, machine learning, deepmind, robots, robotics, self-driving cars, driverless cars, MITRE

Id: pJ3kpCLbP38

Length: 39min 54sec (2394 seconds)

Published: Sun Sep 02 2018