"Tying Knots in Quantum Computing," Charles Marcus, Center for Quantum Devices, Niels Bohr Institute

tonight as we are extremely fortunate to have a speaker Charlie Marcus Joey is a professor at the Niels Bohr Institute and the principal investigator of Microsoft quantum lab in Copenhagen there he also directs the Center for quantum quantum devices chari was raised in even though he now lives in copenhagen he was raised in Sonoma California and studied at Stanford and Harvard before he went to Copenhagen in 2011 he also taught at Stanford and Harvard there's a trend there his research has covered a lot of grounds in the quantum physics of condensed matter right now much of his research is focused seeking condensed matter platforms for quantum information as you as you find out and without further ado please join me in welcoming professor Charlie Marcos thank you thank you for the kind introduction thank you for inviting me to give this talk and thank you for hosting me and Aspen this is the toughest audience you can give a talk to in the world why how many people out here have an advanced degree in physics you see you see the situation you see them peppered into the audience waiting for me to say one thing that's not totally accurate so they can bust me either in person or tomorrow so I want to tell you all you people will put your hand up this talk is not for you you're not the intended audience I didn't write this talk with you in mind hold your horses the two pictures on the slide are well the one on the right is the inside of a man a fabrication facility that's used to make semiconductor electronics the one on the left is a famous building if you come and visit Copenhagen if you've been there before that's where Niels Bohr's office is that's the famous Niels Bohr Institute which is sitting just where it was exactly as it was in those days in fact his office is still just how he left it and it's open you can come in and walk around and the story of my talk tonight while I'll try to get to the cutting edge and maybe the part of the cutting edge I'll give a little bit a little bit to the people here with the advanced degrees in physics the most of the stories about the journey the journey from the building on the left to the building on the right and what happened during the last hundred years of the subject and so I hope you find it an interesting journey and we'll end up with the subject of this week's conference in the Center for physics well it's been a remarkable event in technology for the last seventy years or so in fact in just one human lifetime the world has been transformed by something that maybe even its inventors couldn't have anticipated namely the invention of the transistor and here from 1947 to the modern day is is a revolution the revolution I don't need to tell you about it you have it in your pocket it's changed your life lives it's made you anxious about your children's development it's made you anxious about your own development it is revolutionizing every aspect of should't except possibly the microphone the if you look at the size scale here this is spacing between lines of 120 nanometers which puts these lines at something like maybe 50 or 100 atoms wide the technology of nano fabrication has resulted in the ability to make transistors with billions of them per chip and the crazy thing is they all work however there's a finite capacity even with that technology it's not that everything you dream of can happen why not well here's a the inside view of a data center it's a pretty impressive picture those are racks and racks and racks of computer chips each like the one that I showed dense and fast and the endpoint of something you've probably heard of called Moore's Law where things get faster and faster every 18 months that capacity doubles etc but I want to give you the answer to the question of what's the problem that we're trying to solve here by not looking at this fancy beautiful picture but by going down underneath the floor what's that down underneath the floor is nothing but water water water water water flowing through that thing why because those transistors get hot and keeping the system cool keeping the system fed with electricity keeping the carbon footprint of those data centers is not to mention the cost is a limiting factor in the technology that we live with today we can't put enough of those chips together to do the problems that we need to have done in fact even when we do we see let's see if I can get this thing to go to its next slide yep so what you begin to see immediately is that the electric bill just to run these data centers the carbon footprint of these data centers is becoming something inexcusable and unaddressed by this technology so what happens to all of the problems that are much much harder than the problems that these things can solve build more make them bigger make them ten times larger ten times taller and ten times deeper and running on a thousand times more electricity and putting out a thousand times more carbon footprint no that's not the right way to get to the problems that are too hard for us to get to we have to think of other ways of doing it and so I want to take a new tack at understanding a way of computing that doesn't just take what we have and make it bigger and hotter and more expensive and using it more electricity and polluting the atmosphere more I want to take a new look at the problem by going back a hundred years and asking the question what was bothering these guys look at their faces they're bothered by something and so I want to spend a little time asking this is you know the face is right this is Niels Bohr on the left and Albert Einstein on the right and they were bothered by something and this is what they were bothered by if you look at any drop of oil on the surface of water you see colors and probably those of you who are interested enough in science to come to the lecture tonight probably took a science class in high school or college and you probably learn how to figure out in that class what color was what on that sheet do you remember doing that problem you would draw a little line and you would figure out when the wavelengths would add up in some of the war some of the wave would come down to the bottom of the oil and some of them would go to the top of the oil and window away Laith just lined up for the red wavelength that would turn red and when you lined up for the green wavelength that turned green etc you and presumably the teacher didn't bother with the question of when light gets to the surface and half of it goes down and bounces this way and the other half goes down and bounces that way which half what if I'm a little particle of light am I the lucky guy who gets to bounce off the surface or mind the unlucky guy let's go down and under the water and who determines this is part of what was bothering them let's take that same problem and turn it on the side and just really ask about waves here are water waves I take a tub of water and I stick my finger in and I go like this in two places and water waves ripple out I think so far so good I imagine even for the people without advanced degrees you can put your finger in water and watch those waves go and you can even see places over here on the right where when it impinges upon the screen it's got double the amplitude of one one set of ripples but here in between those two it seems to vanish the wave adds up and disappears you can do the same thing with light you can let light shine from two holes and by find it by drawing the the traces of those two lights you can find places where the light would be twice as bright on the screen and you can also find places where the light will be extinguished again so far so good every wave if you've ever seen does that and nobody should be getting a headache about this kind of a phenomenon here's where the headache comes the headache comes when we turn the intensity down or we change it from electrons to photons it doesn't matter and we let the light go through those slits slowly just very slowly just turn down the intensity and keep track of where the light lands on the screen and what you'll see is the following you'll see the pattern of light and dark and light and dark just like you saw from the two light sources that were shining on the screen but that's after it's been built up for a little while when it first starts to build up it looks like these spots well sure it does these are the individual photons of light or these are the individual electrons that are making their way onto the screen but now you'll remember something that maybe will bother you wait a minute when I did it with the light didn't the light come from two different places and now I'm letting the light go through the screen like this so what should I should I draw a picture like this where the light is made up of little balls well that little ball didn't go through both slits the land on the screen if I ask you about let's take this one next to the letter B that piece of light right there and I say where was that piece of light wouldn't hit the screen well that's not a hard question it was right there I could ask you where was it right before I hit the screen and you'd say well well I presume it was on its way to the screen it was just about right where it landed well that wouldn't be right what do you mean that wouldn't be right like an inch before it got to the screen you're telling me it wasn't right about to hit the screen at that point no it wasn't there because when it got back to the part where went through the two slits it had to sample both slits and go in both places that light wasn't anywhere that light wasn't anywhere until it hit the screen in fact the way Bohr would answer the question of where was the light before hit the screen is you're not allowed to ask that question that's not a good question what do you mean asking the position of something whose position hasn't been measured if its position hasn't been measured it doesn't have a position and that is what let it go through the two slits if you didn't let it go through the two slits by having it have no position you wouldn't eventually see the bright dark bright dark that's the problem that's what was bothering for an Einstein and it continued throughout their lives 1935 Einstein's most cited paper in the history of his literature 1935 the Einstein an old man at this point and yet he writes his most well cited paper at this point in which he drills down on this mystery writing a paper called can quantum mechanical description of physical reality be considered complete in which includes at the end of the abstract one is thus led to conclude that the description of reality has given by the wavefunction the waves of quantum probabilities is not complete not wrong but just not the whole story why not what was bothering Einstein in 1935 it wasn't just the two slits it was something much more profound than the two slits it was something that got the name of the einstein-podolsky-rosen paradox I'll give you a abbreviated version of it here go out tonight if the skies are clear look at a star that's all you have to do to be completely freaked out just look at the star why imagine the light is coming from the star the light comes from some chemical reaction on the star that electron falls down and the light goes out like this in a wave now what am i doing with my hands I don't know exactly what I'm doing with my hands I move something with my hands at the what that the light is coming out in waves it's going to toward me it's going toward you it's going toward the back side of the star going a million miles in the other direction the light is coming out like this now you remember that the light has no location it's going out in all directions at the same time then I open my eye it lands on my retina and I know how a retina works a retina has to take the entire photon the chemical reaction won't work unless the whole photon goes into my eye so there it is the photon is there well wait a minute I thought that that photon could have been on the far side of the star a billion light years in the other direction well it could have been in fact there was a probability for it to be there until I chose to go out that night and open my eye and let the light go into my eye and as soon as it hit my eye it was there and it wasn't on the far side of the star so did somebody call out to the far side of the star sorry he put his eye there he caught the photon it can't be on the far side of the star anymore and if that's true how long does it take for the signal to get to the far side of the star to tell the photon that it can't be there anymore because Charlie went out and open his eye and caught it in his eye I made the photon disappear from the far side this is crazy top right the same year 1935 Bohr now also getting old writes a response in a paper with the same title can quantum mechanical description of physical reality be considered complete those are you scientists out there try doing that sometime if your if your nemesis in the scientific world writes a paper that disagrees with you write a paper disagreeing with the same title so you share the references anyway Bohr writes a response to Einstein's paper with the same title 1935 in which he says basically the response to this experiment that for that Einstein called out that is how did the thing vanish on the far side of the star with the message getting over there seemingly instantaneously and and you know you could imagine the father of relativity not liking the idea that the message got to the far side of the star if that's in fact what happened that just sounds like a bad story if you say nothing propagates fast in the speed of light Bohr's responses as follows in fact this new feature of natural philosophy means a radical revision of our attitude is regards physical reality whoa I mean he's not talking about physics anymore he's saying our entire belief structure of how the universe works has been called into question by the idea that yes when I open my eye I caused the probability to on the far side of the star instantaneously don't like it go live in another universe it was so jarring that for decades right up to the present there was a class of scientists who test this idea over and over and over again and every time they do it it comes out the same sorry Einstein quantum study suggests spooky accident action and distance is real so it doesn't matter whether you like it and it doesn't matter whether Einstein liked it all that matters is what happens in the experiment and you have to get used to the idea that you're living in a world that's not quite the world you thought you were living in Bohr is famous for this quote if quantum mechanics hasn't profoundly shocked you you haven't understood it yet so if there's anybody in here who is not in a state of shock right now then I'll go back and do it again let's move to the modern age of electronics and take these same ideas with us namely if I take something that looks almost like a two-slit experiment except it's made out of a semiconductor and electrons can flow through it if I give the electrons two paths to go to and I do some other tricks like make it pretty small and I cool it down to low temperature then I can get the kind of interference pattern that's very much like what is shining on the wall where the electron must have gone via both paths so that you know that this picture of an electron being here and there isn't quite right because it's not really anywhere at the same time it's kind of all over the circuit at the same time how do I know it's all over the circuits all of the time and not just a blue ball because of the interference pattern it must have been at some time in both places at the same time that's the data that tells you about the interpretation what that means is that if I take the electric circuit and run the electron through and say it's in both ways you know electrons have this funny property they recharge in charge can be used on the capacitor plate of a transistor to turn the transistor on and off so now that I've said that the charge really isn't anywhere and it's kind of on this side it's on that side at the same time you must say that it best then turned that transistor on and kind of not turn it on at the same time well that transistor lets electricity flow to some other part of the circuit and not let electricity flow at the same time which then charges up and doesn't charge up another capacitor at the same time it gets a little bit complicated but this is quantum mechanics this is the story of quantum mechanics and it is a story that is not in conflict with any experiment it is how a lot of scientists would define as right the way we say it is that the switch can be on and off at the same time and this wave function with the Greek letter sy defines the state of the system as being composed of both parts together until the measurement occurs and then one of the outcomes obtains that's the rules of the game you can make the game a little bit more complicated by saying that the combination could be 80% up in 20% down doesn't need to be 50/50 in fact the ratio can even carry some kind of complex number that represents phase that gives the interference pattern so you end up drawing not 80% up in 20% down but all the possible vectors on the surface of the sphere and you do the math as if that was the way it worked two years ago the Aspen lecture had a speaker named Scott Aaronson I was here for that lecture and Scott Aaronson is also famous for a particular quote maybe not as famous as Niels Bohr's but Scott is famous for saying quantum mechanics is astonishingly simple and I hope that you found those rules that I just told you very simple it's 80% up 20% down or some complex phase up and down you propagate the whole thing through and when you get done you make measurement and that becomes the outcome of a 0 or a 1 for the state to propagate it through as long as you don't think too much everything's fine like the rules work you get the right answers and that is what Einstein was saying originally which was it's not that the Theory's wrong it just can't be complete and yet from that day to this we've never found a better version of the story we've never found an error with the predictions and it seems as much as anything we accept in our lives as being correct this is correct so now let's ask an interesting question this is a question that maybe a physicists wouldn't ask but an engineer would because their engine any engineers in the audience anyone professional engineers in the audience so tell me if I capture the spirit of the engineer incorrectly I'm not an engineer myself but you could nod or shake your head you might even yourself be thinking yeah that's weird it's kind of interesting is it good for anything is there anything you could do with that do you couldn't do before let's ask that is that a fair characterization yeah he's shaking his head yes so now I can turn on a transistor right here and not turn it on at the same time it lets electricity flow to this transistor turns on that transistor this transistor turns on that transistor in fact it came and go back and turn on the original transistor again the whole thing can propagate all around until finally I measure one of them that determines the state of everything else on the chip how fast instantaneously not even the time for light to propagate across the diameter of the chip the state is determined instantaneously what can we do with that let's ask the question if we built a computer using these rules and take advantage of not measuring until we're late in the game could anything happen doesn't happen with a normal computer in which everything is either transistor on transistor off this is a theoretical question and I'll give you three theoretical answers the first one is in the subject of cryptography and it arose from a famous paper that was published in 1996 on a public archive that read as follows a digital computers generally believed to be an efficient universal computing device but this may not be true in quantum mechanics is taken into consideration that is it may be that there are quantum mechanics problems for which a computer is not an efficient tool this paper considers a particular problem called factoring integers well that's a nice little math problem that everybody in the room can participate in I want all you scientists to shut your head hands off for one second we're gonna play a game for people who are who are moderately good at math but not experts in math a prime number is a number that can be divided only by one in itself and I'm going to tell you that there are two prime numbers that multiply together to give you 15 can you think of the two numbers yeah shout it out not expert shout it out three and five that's the correct answer okay not experts two prime numbers multiply together to give you four thousand six hundred and thirty three okay I'll just tell you the answer is 41 in 113 now what's interesting is you couldn't see it but I could see it from where I'm standing when I flashed the second question up nobody moved a muscle nobody nobody reached for their calculators you would know what to do with your calculator hey go ahead reach for your calculator what do you do with it what are you gonna do to answer not that question to answer that question what are you gonna do with your calculator you don't know you start guessing but your calculators not gonna help you that much or maybe multiply it a little bit faster and I'm telling you that it wasn't because you missed that day in class when they told you how to do that problem that's just a hard problem that's just a problem that computers do it with a little bit more efficiency but they basically just start trying start trying answers so when I tell you that this top number is made up of the product of those two prime numbers that little fact is the basis of Internet security it's the basis of encryption it's the basis of that phone that that the Department of Justice wants unlocked all that stuff is all about that equation because whereas a computer can multiply those two numbers together very easily once it identifies your thumbprint if it has that number I can't do anything about it and it can't figure out what those two loops are just like you couldn't figure out what those two roots are so what that paper in 1996 showed was if there were a quantum computer in which the state determined that state and the state determined that state and everything propagated around the chip that problem suddenly goes from being intractable using the fastest known algorithm with a hundred pcs that a clock speed determined by extrapolating in 2018 once the numbers a thousand bits long it takes the age of the universe to do the problem using the fastest known classical algorithm but if you have a quantum computer running it any speed the problem becomes easy which means all of the encryption all of the government secrets all the internet security is broken by one of these machines that's interesting that's an important problem so it's not surprising that in 2019 finally the US government along with the European Commission in Europe have dedicated in the u.s. its 1.3 billion dollars to fund research to develop a quantum computer because the last thing we want to be is the second country to have one of these it's ok if nobody's got one but we don't want to be the second country so it's an interesting time both technologically and politically here's the second problem I walked into the cafe that was before some of you were there this problem is being discussed here's an interesting this is a picture of a maglev train magnetically levitated train it floats on a carpet of magnetic field it doesn't have any wheels it doesn't make any of those grinding noises because it's flying through the air on a magnetic field the only difficulty is it needs a super conductor to make the currents that generate the magnetic field wouldn't it be fantastic if that super conductor were at room temperature is there a superconductor room-temperature I see heads going in both directions I'll give you my answer I don't know there's an interesting story about the high-temperature superconductor superconductors that conduct better than any metal this one magnesium dye boride has been around for 100 years it's been sitting on chemist shelf this is six sitting there turns out some Joker finally took it off the shelf and put it in low temperature and lo and behold one of the highest temperature superconductors found was sitting on a chemist shelf for the last hundred years how could nobody have known that it's not that complicated molecules magnesium dye boride couldn't somebody have predicted that that was going to be a superconductor not just a superconductor but high temperature superconductor no there's not a computer on earth that could have told you before you started now let's say they did another trick they said well hey I'll tell you I'll give you a hint mr. computer it's magnesium dye boride what's the transition temperature no answer says the computer that's an unsolved problem here's another interesting problem did you know I didn't that we spend over 1% of our energy on earth making fertilizer here's the breakdown in percentages it's an expensive energy consuming hot process that's very inefficient but we need fertilizer without fertilizer a lot of people's lives would be much less happy than they are now now in most plants or most legumes anyway in the roots of those plants there's a bacterium and the bacterium makes fertilizer for the plant it has a symbiotic relationship with the plant and it makes the same fertilizer out of the air it takes nitrogen it fixes it into ammonia and it makes fertilizer it's complicated as you know to take apart the end to molecule that put it together to make a to make to make ammonia it's complicated how does it do it well there's a there's a protein or inside of the bacterium in in the heart of that protein there's a little complex that has in it interestingly an atom of molybdenum molybdenum I mean you might not even know that what is the hell is molybdenum it's also got an iron in there there's this somewhere in there there's a F somewhere I can't see it but anyway and there's a carbon in the middle and there's a there's a you know in that little complex is the thing that rips apart the n2 and puts together the ammonia with the hydrogen's to make the fertilizer and it does it at room temperature and it does it to very low cost how does it work nobody knows you don't believe me look it up I'll look it up for you here from Wikipedia you know Wikipedia doesn't know Wikipedia doesn't know anyway go read about this nitrogen fixation molecule that this protein that's in the roots of every every plant every Legum I'll read you the location of substrate attachment to the complex has not yet been elucidated it is believed that the iron atom closest to the carbon there's the carbon to the inner sister carbon participates in the substrate activation but the terminal molybdenum is also a candidate for nitrogen fixation that's it right there it's that big and nobody knows how it works why not because it's too big for a computer to solve the problem of what happens when nitrogen comes into that molecule why molybdenum well maybe that's just the way plants evolved maybe there's an even more efficient catalyst that we could put in that would make it even ten times more efficient than plants there's no reason to think that plants are you know the best way to do it but we have to understand what that thing is doing in order to understand what to do next to make fertilizer without using 1% of the world's energy so an interesting paper from colleagues in Switzerland in Microsoft and large group of people have studied the particular problem of this it's called a thermo Co complex for fixing nitrogen and ask the question if I wanted to make a quantum computer whose only job was to solve the problem of what are the excited in ground in chemical states of that molecule how big of a quantum computer would I need and their conclusion is that this can be formed in a reasonable time on a small quantum computer it'll turn out you need a hundred logical qubits I'll come back to how you make that later but it's not a hard problem if you have a quantum computer and if you don't have a quantum computer its intractable last example that I want to give and I'll move on quickly to what to do with it this is from the Davos symposium this last year this was this study was published on December so less than a month ago the following article came out how quantum computing could beat climate change this particular article thought it was read to because it's you know clear enough at a meeting of the World Economic Forum's Global futures Council last month a team of experts from across industry academia and Beyond assembled to discuss how quantum computing can help address global challenges as highlighted by the sustainable development goals and climate in particular there are known catalysts for carbon capture but most of them contain expensive precious metals are difficult or expensive to produce or deploy molybdenum this metallo organic is an example of where a metal atom is used in an enzyme or two as a catalyst and we have catalysts that can capture carbon that can remove carbon from the atmosphere but they like that molybdenum involved precious metals given the infinite number of candidate molecules that are available we are right to be optimistic that there is a catalyst or indeed many to be found that will do the job cheaply and easily but finding such a catalyst is a daunting task without the ability to simulate the properties of candidate molecules so there's a chemistry problem folks which is that once the molecule gets a little bit big we're stuck all those data warehouses are not gonna solve that problem but a modest-sized quantum computer will okay well I think I've motivated the problem let me ask somebody who's got a watch on what time is it six 21 2 6 12 and when would you like me to stop talking 6:30 okay that's my goal let's build a quantum computer let's see how people are doing it out there and then I'll tell you a cool idea for the last part of my talk about topology the subject of this week's Aspen conference which is maybe an interesting cut and how to solve this problem well the approach is that you have to solve a certain difficult problem in technology here's the difficult problem you have to be able to control every one of those bits that makes up the quantum computer you have to open this valve close this foul dude like you know like you're like you're operating transistors but you can't measure anything you can't look at it you have to do everything without knowing what anything is ok sounds hard so it's so that you can't even think how do I even write a program to where I don't have to say you would say if you're a 1 ok I can't tack can't do it I can't say if you're a 1 I could say I could say if you're a 1 flip your bit if you're not a 1 don't flip your bit ok good I don't know I don't know I didn't say whether you were one or not I didn't measure you I just said if you're a 1 flip your bit if you're not a 1 don't flip your bit is it possible to write all of computer code using this thing where you don't have to measure anything until the very end yes it is you can write general-purpose computer code without performing measurements only allowing all the states to prop at the same time but in the lab what we have to do is we have to make systems that can be manipulated because you may or may not have to flip your bit and you may have to do it in a nanosecond depending on when whether her bid is up or down but you don't get to know until yeah but you you you we run the algorithm but I can't measure you accidentally I can't even accidentally figure out what state you're in okay so there's a couple of systems that people have been looking at one of them involves atoms and in the energy level of the atom it can be in an excited state or a ground state and whether it's an excited state or a ground state could be the zero or one of the bit you can have current flowing in a superconductor a little superconducting circuit and the current could for instance be flowing clockwise or counterclockwise clockwise could be a one counter clockwise could be a zero you say have they be flowing both at the same time we remember this is quantum mechanics you can't say whether it's flowing clockwise or counter clockwise finally every electron on earth carries a little bit of magnetic north south and it's not much so that if you don't actually actively try to measure and it's kind of hard to measure it you know the universe kind of leaves it alone it just kind of freely spins but if you could use that to orient the zero or one then you tend not to screw with it until you actively put on a magnetic field that flips it so that's a good candidate also so a lot of different labs have lots of different candidates in every lab has their favorite candidate and to be honest it tends to be what they were working on already before quantum computing got invented they say ah mines a good system for that if you're an atom guy like this Dave Wineland Nobel laureate who works with atoms and said you know what I think my atoms are exactly what the doctor ordered here but the spin guys think oh no I think my electrons are exactly what the doctor ordered the superconducting guy's saying haha this is exactly what the doctor ordered my system is here and so far what's happened in the US and in Europe and in China and in the rest of the world is a lot of these approaches are competing with each other nobody has decided incorrectly I would add my own opinion has down selected and said that was stupid don't work on that one because who knows nobody knows it's a brand new problem what's the right system in which to do this well let's let a lot of people try their different problems and see who can get the most bits going and who can have them stay coherent for the longest amount of time that's what's happening now that's a snapshot of 2020 is that there are labs with atoms superconductors spins and they're all kind of competing with each other to make this thing work here's it one of the superconducting ones this little loop right here contains two of what are called Josephson junctions there are two little breaks in the superconductor and you have to send all the signals in and out and they use this superconducting grid to do it this is an electron micrograph it's a very impressive thing but even more impressive is the work that came out of I'm gonna skip more quickly well let me let me tell you what this is here's the issue I sneaked I sneaked a word in and I want to bust myself for sneaking it in without saying what I meant when we were talking about the bits over here and I said well you can't we can't measure and then I said you can't even accidentally measure like I can't accidentally find out whether you're in a 0 or a 1 state let's say you interact slightly differently with that cushion that you're sitting on we're depending on whether you're a 0 or a 1 maybe you're a tiny bit heavier when you're a 1 than you are a 0 the cushion that will no the cushion will know whether you're a 0 or 1 and if the cushion knows it's going to tell the cushion next to it and pretty soon everyone's gonna know and then you're not quantum-mechanical anymore you've been measured not on purpose but accidentally when you thought you were protecting your information the cushion measured you and pretty soon everybody knew that's what happens that's why these machines don't work that well that's why when you put something in a superposition of the two states when you make it a zero in one at the same time you only have a few tens of microseconds before the quantum mechanics goes away it goes away accidentally because the universe accidentally measures it that's important because that's why we don't have these machines right now the ideas are all there the technologies are there we're trying to prevent the cushion from measuring whether you're a zero or one and we don't know how to do it very well so we can only get a few hundreds or a few thousands of operations as fast as we can operate before the quantum mechanics is gone now this was pushed to its logical limit last year 2019 a very important year for quantum computing because Google and a group run by John Martinez at Google went pushed as far as they could with this thing up to 53 qubits interconnected with each other in various kinds of geometries and found that they could just achieve after a few tens of thousands of cycles where all the fidelity had crashed and they'd lost most of the information because it was too long compared but it was long enough to perform a calculation that a classical computer can't perform they outperformed a classical computer what was the problem forget about it you don't want to know was this stupid problem by their own admission it was a problem that nobody cares what the answer to the problem is but it was a problem that a classical machine can solve as fast as their quantum machine could that was a first and that was 2019 next it would be nice to get it to be a little bit bigger next after that it would be nice to have it work on a problem that that that was a that was a practical problem but boy it's important to demonstrate that you can solve a problem fast than a classical computer can solve it and that was 2019 so hats off to John Martinez for doing that for the end of my talk and I probably have about ten minutes left or something like that I'm gonna change the subject to the topic of this week's conference and I want to draw my inspiration from this many thousand-year-old technology of storing information in topology remember the title that he said you need an Incan and didn't weren't expected to know every word topology is a branch of mathematics that is concerned with the shape of objects without any interest in their details if it can be deformed so you know like I lost 20 pounds I'm still me even if I became a sphere I'd still be me but if there's something like a rope that has a knot tied in it that's not the same as a rope without a knot tied in it in fact every one of you knows that a rope with a knot tied in is different than a rope without a knot tied in it because you every day take out your headphone jack cord and it's all wrinkled up and you'd say the same thing to yourself that every other person in the world says which is I wonder if I just pull on the end of it will it come out straight or will it be tied in a knot and the answer is it's either one or the other the answer is not maybe try it sometime next time you take out your headphones just pull on the end of the court it will either be in a knot or it won't they are topologically distinct States it doesn't matter whether you wrinkle it all up it doesn't matter whether you hold a football just pull it and find out it won't change that was the basis restoring numerical information back in Mesoamerica and in fact it was even a system that uses that use the decimal system and they go tie the knot so this way and why do we know that now because the things are still sitting in the ground and they still have NOx tied on them all the parchment is gone everything else is gone but the knots are still there those knots don't come out in fact the first computers were in the business of tying things in knots they were weaving machines here's a picture of the computer actually used punch cards to determine the rows up up and down let's call that the first computer and what was the output of the computer it was weaving it was putting fabric through loops and knots just like the fabric of my shirt and when you pull on it the knot doesn't come out its woven in the memory of the history of which thread went around which thread is embedded in the time history of my cloth and that in fact was the first computer now let's bring this to quantum mechanics because what we said in quantum mechanics was if your state is measured you're not quantum anymore maybe we can hide the state in such a way that the cushion won't measure it the cushion won't know whether it's quantum mechanics said it or not so I drew this picture here and I said if you look at the rope if you measure the rope somewhere locally that no local measurement can see whether this rope has a knot in it you have to kind of step back and see who goes through what in order to see whether this is a knot tie minute therefore any information that's encoded in knots will not be crashed by a measurement that only operates locally which is how most measurements operate they operate on a spot on the material if that spot on the material ways you measures your cubed tells you whether you're a zero or one quantum is gone so hide the information and they're not because the local measurement can't see it here's the only problem the only problem is we need particles that you can tie in knots and they don't exist the memory of those particles doesn't exist well that's ok let's invent them and then tie them in knots so it's kind of a two-step process our friend Mike Friedman from Microsoft was one of the early developers along with his colleagues particularly Alexi kitaev who was one of the inventors of this of a branch of quantum computing called topological quantum computing and he mentions right in the abstract that the chief advantage of this it was called any onyx forget that topic topological quantum computing would be physical error correction the-the-the errors that are encoded aren't susceptible to being measured it was just an idea he's a he's a mathematician I mean he's not even a theoretical physicist he's a step removed from that but it didn't take long for the theoretical physicist physicists to realize that if you had such particles that you could tie in knots they're not worrying at the moment about what those particles are where to get them they're saying if you had a particle that you could tie a knot you could make a computer that only use the tying in knots as the way to process information so you could do an entire general computer just using this knot business and encode information and knots now all we have to do is find the the particles that do it now this is such an important and novel idea the subject of this conference this week of all these physicists out here who are studying materials that have this property that they have a topological character that their excitation can be tied in knots that this was the recipient of the 2016 Nobel Prize in Physics and I'll just read to you about these topologically distinct materials topological insulators topological superconductors topological models are now being talked about these are examples of areas which over the last decade have defined frontline research in condensed matter physics not least because of the hope the topological material will be useful for new generation of electronics and superconductors or in future quantum computers so let's try to make one let's try to tie particles in knots in 2010 10 years after the original ideas of topological quantum computing were first put on the table one particular idea emerged for creating excitations that would have the property that if you wrapped one around the other it would remember and it involved objects that had really never been combined before probably it involved a superconductor a material that conducts electricity without any dissipation it involved a one-dimensional wire by one dimension I mean something in particular I mean that along the length of the wire the electrons can move but transverse to the direction and that narrow directions of the wire they're in their lowest quantum mechanical state just like in a hydrogen atom the hydrogen atom is in its lowest state so imagine something that's in its lowest state that way in its lowest state that way but this way it's free to run so it's like a long skinny hydrogen atom that's a one-dimensional wire and the third ingredient is an external magnetic field you put those three ingredients together the theoreticians tell us that there will be balls at the two ends what happens when you wrap them around each other well only one of two things can happen when you bring them together they will either become an electron or they'll become nothing very simple but rap about topological you can wrap one around each other and change who's got the electron it becomes a game of who's got the electron meaning when you take a new pair and put them together do they become an electron or do they become vacuum that was the idea how do you do it well if you wrap these two around each other and this pair had the electron at the beginning and this pair didn't and then you swap partners well who's got the electron at the end well you can do that by moving moving these things around if I want to wakfu they are I have to step into the aisle bring this one down step around the other side and I've effectively made this change a swap of the who's got the electron game you know I have to say that I had no intuition for this who's got the electron my Arana problem until I was in Estes Park Colorado kind of a kind of a tourist trap kind of town while I'm from Sonoma so I could say something's a tourist trap and inward Asmus so we can say it too but in but there I was in estie's Park and I saw a dude anyone know what that is yeah it says that they're sorry that's a stupid question it's saltwater taffy it's a taffy puller and I was looking at this thing in the window and I would like the way that had matched the window frames also uh and I saw and I saw a pair of Meyer Ana's and another pair of Meyer honest and I was just watching them switch partners that the pair would make a partner and then the next pair would make a partner and they were trading off who's partnered with who and that each branch would then exchange who was the partner and the question is could you take those my Rana pairs at with the ends make a whole computer out of those where each time they swapped not who's got the electron but who would have the electron if you brought the two together and promise you'll never put them together that's the idea of these my Rana computers we know how to make the wires we know theoretically what to look for if those red balls are there what do we look for we look for a peak at zero energy after you put the magnetic field on remember you needed superconductivity the one-dimensional wire in the magnetic field when you have no magnetic field there's nothing at zero you turn on a magnetic field there's a big bump at zero look for the pump look for the bump at zero this was 2011 the year after the the idea came out and it didn't take very long for my buddy Leo Cowen joven here Leo Calvin and his group at Delft University to make the device there's the wire there's the superconductor there's where you put the electrons in and there's this crazy peak right at zero when you turn on the magnetic field up to the right limit calling the paper signatures of Meyer on a fermions in hybrid superconductor semiconductor nano wire devices now at least I dare say the title makes sense very soon after that oh sorry excuse me very soon after that Lehman Rock Anson and his team at Purdue generated another signature the disappearance of a peak another prediction of the existence of these my Ranas and pretty soon 2012 became the year that these first signatures were seen two years after that my lab in Copenhagen developed the next level of technology which is to grow the superconductor in the semiconductor together instead of slapping them together as a sandwich grow them here's the surprising result the semiconductor and the superconductor grow very nicely together so from my lab a grown semiconductor and now the theory turned on its side so it matches the data starts to look awfully like what the red ball prediction looks like we're pretty much getting there now something's interesting about that which is the disk I count Hoeven and I who've been friends and competitors for 30 years now he introduced me on this very stage 10 years ago and the computer didn't work it was an interesting hour anyway we got through it but I was it was then almost spooky when I came in to Aspen yesterday and saw not only is there a coward he'll be building on the corner but that Marcus seems to have invented it and I thought this is a really special town for topological quantum computing and that now the whole thing in Aspen sort of makes some sense so let's make our first qubit the first qubit and I realized I got to stop two minutes he said to me the first qubit is going to look a lot like this taffy puller except it's going to be in a straight line because I'm going to use one wire and I'm going to be opening and closing the doors instead of moving them around in circles to catch to catch the handoff I'm gonna use the semiconductor property they really aren't doors they're drawn here by our theoretician friends Jason Alice Seiya and David Aysen two opening and closing doors but you know that you can open and close a gate on a semiconductor that's what semiconductors do you put a voltage on there and it shuts the door you take the voltage off and it opens the door that's classical semiconductor physics if these things are indeed made out of semiconductors with superconductors on them and what this theoretical paper shows is with some judicious opening and closing of the doors and what I mean by that is the opening and closing of the voltages that open the doors I can I can play who's got the electron when I bring them back together and that that's the idea of the ACE quantum computing the Maya Ranas are these half electrons that either fuse to vacuum or electron at the end the devices themselves start to become rather complicated because not only do I have to have the nano wire with a grown superconductor in the ends right through the measurement but I have to put all those doors on so that I can operate the thing so one little tiny topological object starts to get kind of complicated that thing sits on a chip that chip is on a conventional electronic carrier that carrier sits inside of a big hunk of metal that big hunk of metal sits at the bottom of the refrigerator which then has to take the thing to one tenth of a degree above absolute zero where all of this experiment has to take place why absolute zero because I don't want the cushion to shrink when I have that qubit even though it's topological it's still susceptible to two inadvertent measurement that thing sits inside of a refrigerator which is inside of my laboratory and you know this must be Denmark because where else would you find a laboratory with hardwood floors now things do get more complicated and to build the kind of machines that we need in order to solve that nitrogenase molecule we're gonna have to put a lot of red balls on a lot of sticks and put this whole thing inside of one computer the big breakthrough came and I think that the author of this work is here in the room with us from a couple of groups including Mike man for his laboratory Purdue University is to extend this wire idea to two dimensions now two dimensions is a lot more than just one because once you have something in two dimensions then you can define the wires lithographically not by making one wire and laying it down imagine what a pain in the neck try building a whole computer a little stupid wires instead you have to lift the graphically define all of the wires and when you have a two-dimensional substrate you can pattern one or 100 or 1 million using the same lithographic process that's a breakthrough when you can suddenly pattern wires and so when we see this level coming down hitting zero and making a Maya Rana knot in a wire that's grown as the wire but in a wire that is patterned as a wire the next breakthrough that come through the final thing that I want to say is this we're gonna need a million qubits just in order to get that hundred qubits system why because I need to build in redundancy there's too much measurement that goes on inadvertently but I can get rid of some of that measurement with redundancy but the redundancy factor is pretty daunting like I might need a thousand cubits or ten thousand cubits for every logical qubit that does what it's supposed to do so that overhead factor means my refrigerators start to look a little bit complicated when I have to run all of those wires down to the bottom in fact I hope what you're thinking to yourself is that's no way to run a business that's not gonna get us to a million things down at the bottom of a 0.1 Kelvin refrigerator running everything down through coaxial lines from room temperature the engineer in here must be horrified so one of the challenges which you might find mundane but which is absolutely central to the improvement of the technology oh I wanted to show you what the kind of numbers are so in the back of this paper this one that I showed you before in the back of the paper they give the numbers how many total physical qubits oh look everybody's very happy how many logical qubits logical qubits 100 110 109 no problem oh yeah well when you get down to counting the total number of physical qubits depending on how good each individual qubit is maybe it's only 10 to the 5 or 10 to the sixth or maybe it's 10 to the 11 or maybe it's 10 to the 9 cubits so you better have an idea in mind beyond running those wires down from room temperature to the bottom of the cryostat but actually it's okay because we have chips in our pocket with a billion working transistors the only problem is you got to get them down to the bottom of the refrigerator and have that chip working at milli Kelvin temperatures not coming down from above so here's the first one of a cryogenic control system that functions at the lowest temperatures these will all have to replace the classical control the opening and closing of all of the gates when we're finished this I think is the end of my presentation I've covered a range from the Bohr Einstein debates through to cryogenic control of my rana fermions it's been a wild ride thanks for coming along with me Wow this is one time I feel really a villain to stop the show but there before patty comes up to kill me I think we can take a couple of questions yes our engineer repeat request yeah well if generating funding is useful it's already happened you know I I want to push back on the form of the question in a way which I think you'll accept which is every year something new happens and of course the year of 2018 they weren't asking the question what us they said well why don't we have a part of computer they could do something you know the regular computer can't do it was 2019 now yes what is it going to do something useful well what if it's a little bit useful what if it you know we kind of knew the answer anyway like that factoring problem it's gonna just get better and better every year and I don't think looking back just like with conventional computer technology there is no conversation like when did the laptop happen you remember there were like these gigantic crappy laptops at the beginning and then they got a little bit better and then a few years later for those of us old enough to remember they got a color screen and then they got a little bit lighter and then the battery lasted a little bit longer that's what's gonna happen with quantum computing it's gonna get better every year so that you won't even notice when all of a sudden they're solving problems that they couldn't solve before and so I think that I want to say that this is more about evolution than revolution and that the little games that we're playing right now are part of a long arc that I think will take the next 50 years and at the of that 50 years we'll have something that we will not only not recognize how computers work but we won't be able to distinguish the incremental steps that's my long answer to your question oh well mark saw as an example I work for Microsoft in addition to working the University of Copenhagen and there's a hundred and forty people working in their quantum program and none of them is supported by the US government they're all supported by Microsoft because Microsoft is nice no because Microsoft thinks that this is an important thing for the future of technology Intel IBM Google Airbus all the same thing they're all paying to play oh you mean when will we be able to do something that we can't do with a conventional computer yeah well you can go read that in fact already Volkswagen has purchased optimal solutions to I think truck routing that have been that have been solved on a small of IBM process or something like that now could it have not been done another way no it could have been done another way so it wasn't the Traveling Salesman problem now the interesting thing about the Traveling Salesman problem is that it's an np-complete problem it's in a class of problems I'm not an expert at this but this one I happen to know that's a hard problem that's a problem that's in a class of problems that probably a gate based quantum computer will not solve it's an optimization problem that has many many nearly optimal solutions and I can ask you as an engineer do you want the best route or do you want to darn good one fast probably the latter and so that class of problems where there are in exponentially many solutions and you want a pretty good one not the very best one that's a different kind of problem and there are quantum computer companies called d-wave that are already taking on optimization problems of exactly that ill - including traveling salesmen like problems np-complete heart problems and what's interesting about that problem is it is not known on this very day whether or not a quantum computer will ever outperform a classical computer on optimization problems it is not known I mean not even known to theoretically it's not forget about the question of when we have a machine the theoretical answer to the question of will it in principle outperform a classical machine is not known so we're kind of in our infancy but it's good it's a good time to be alive we get to we get to not know the answer to a lot of things just for coming along with me [Music]

Channel: Aspen Physics

Views: 1,930

Rating: 4.9436622 out of 5

Keywords: ACP, Aspen Center for Physics, quantum computing, topology

Id: lY6Xd-yp9ng

Length: 67min 28sec (4048 seconds)

Published: Tue Jan 21 2020