Christopher Monroe, Quantum Information Science

well welcome everybody we're we're making a transition here this day is on international engagement and this morning we heard a lot about international engagement in large science which of course occupies a lot of physics but now we're going to make a transition to international engagement in very competitive areas and one of the most important for physics that has been rising very fast of course is quantum information science and so the will will will start with the keynote given by one of the leaders in that area chris monroe so let me tell you a little bit about chris chris is a distinguished University professor at the University of Maryland and he's also a fellow at the joint quantum Institute which is joint because it's between the University of Maryland and NIST and it concentrates in many areas of quantum science including quantum information science chris is a graduate of MIT I got his PhD from the University of Colorado in Boulder one of the great amo physics institutions in the country he's the recipient of many prizes he's won the Rabi prize the shallow prize and the Willis lamb prize and you know that's quite a trifecta in amo physics so it shows he's been doing things that the community really benefits from and really appreciates he's also a starter a founder of a startup company ion q which is in the quantum information space and he's going to talk to us today about quantum information science and its role as a very competitive activity on the international stage so please welcome Chris Monroe [Applause] Thank You Phil in fact I owe Phil a lot more than he realizes I remember a plane trip when we were both in Ann Arbor on the faculty diversity of Michigan we came down here for some lobbying event I forgot what it was but I saw him work work a room full of a few Congress Congress men and women and and their staffers but on the on the flight back to Michigan I also got less scared of using pulse lasers in my own research due to him so I really thank you that really revolutionized a lot of the stuff that I do so the topic today broadly speaking is international you know open science and physics and also competition and indeed this is one of the areas that has its really competition on steroids in this field and part of what I hope you gain from from from my remarks are that we still really don't know what this stuff's good for so in a sense it's very beautiful and in this field really needs to be couched in science nevertheless there are big and small companies involved building things that we don't know what they're going to do so it's it's a rather interesting journey and I've seen both sides of things and I I'm proud to have successfully gone back to my my full-time job at the University so I encourage that startup that I co-founded with jung saem Kim from Duke University about three or four years ago we're about a hundred people now on College Park just about five minutes away from ApS headquarters there and I spent a few hours a week there now and it's it's beautiful seeing all the I mean there are very few quantum trained people at inq maybe about a quarter of them are twenty percent but the engineering aspect of building devices even when we don't know what they're going to do make makes it very fun so this is a picture of 18 I should have counted before about 20 individual atoms they're all individual you terbium atoms each of them is a very good atomic clock actually and we use we use a lot of the features of atomic clocks to to promote the idea of this being pretty much the leading platform to build a quantum computer and I'll talk a little bit broadly about what that is you've probably heard a little bit about that just real quickly the the the spacing between these atoms is only a few microns but we can focus individual lasers obviously we can resolve them optically the atoms are really cold they're about a thousand times smaller than you see this is just the diffraction limit of our optics and it's not perfect as you can also see but this is a real sort of time movie they sit there all day and this is because they're absolutely identical I mean their atomic clocks they define what identical means so in a sense well we don't think of this as being a super easy platform to scale to large really large numbers of qubits it has the fundamental ingredients that no other technology has to scale and that is being able to replicate nearly perfectly so amo physics will I think as soon as the big companies realize this they'll be jumping into this technology in a big way so as as some background in the field I think as physicist most of us are pretty well steeped in the concepts of quantum physics as strange as they are at least we're comfortable using them but information theory is something I had to learn pretty much pretty late in my life and these these two characters should be sort of heroes to to us physicists because they abstracted information into things that could be represented by anything physical but the point the point is if you want to store information you have to use physics there has to be some physical way to do that touring for his part invented this abstract notion of a computer based on a tape that moves across a head some interactions here and this is not a real machine it's an abstract idea but you can go ahead and build it out of many different types of hardware's Shannon on the more mathematical side and his colleagues at Bell Labs invented the concept of a bit the fundamental unit a unit of information 0 and 1 and the an entropy which has close ties dat thermodynamic entropy allowed us to consider information content coding error correction things like this and of course in bits we look for physical systems that can represent the bits and at the time the state of the art was vacuum tubes which were at least they sort of worked they weren't so reliable you could put several thousand of them into machines like these here based on these vacuum tubes again this predates most of us but they weren't so reliable they broke a lot and so scaling up to many millions of them was pretty much a non-starter but of course the beautiful everybody knows this picture from the 40s it's a physics picture of the first solid-state transistor based on germanium a gold germanium Junction here where a current through the gate here controls the current going that may be much larger current going from the source to the to the drain so this is a physics experiment but because it's solid-state it has had the potential three much more reliable and of course we know the rest is history maybe if you all know about Moore's law I would say it's based on the fact that we invented technologies in solid-state physics that allowed us very large-scale integration of VLSI this particular version of Moore's law is only for the last few decades where we still enjoyed this exponential growth factor of a factor of 10 every decade or something like that but we see signs already of it starting to saturate and this is simply because the transistors are just getting too small they're getting so small they're approaching molecular size scales and you know they're interesting effects there it's very hard to control a big macroscopic hunk of something at the microscopic level like that well you've probably seen this wonderful speech of Richard Feynman printed at an ApS meeting in 1959 there's plenty of room at the bottom and he was of course one of the founders of quantum electrodynamics but he liked to tinker as we know and I think he was he was really struck by the the fact that there were solid-state transistors and they could be miniaturized even down to individual molecules and he had this zinger in this paper that really hinted at something more to come here and that is when you have circuits designed out of individual atoms they behave nothing like the you know they behave according to a different law of physics so there should be new opportunities for design this is this was almost 80 years ago and I think by now we we understand that that opportunity is called quantum information science the fact that we we can use different physical law to maybe help the problem of this lack of growth from the classical Moore's law so that that's a little bit optimistic I would say we like I think I said at the outset we're still not sure how this will play out but there are some simple ways of understanding how quantum information systems work quantum computers in particular and these opportunities are that quantum systems can be in states of superposition so bits can be not just indefinite states but they can be in these confused super post states of 0 and 1 and the waiting's between 0 and 1 can be controlled from the outside world and the real strange part of quantum of course is the fact that it's not the wave part that's sort of pedestrian if you're familiar with differential equations the strange part of quantum is the measurement problem where when we look at its superposition it randomly picks one or the other state and that's that's the revolutionary law of quantum I think fine ones referring to here it requires probabilities from the ground floor like no other theory in all of nature it seems to imply that by observing something we affect that something so these are the weird laws of quantum and there's all kinds of philosophy books written about them but like I said as physicists we're at least comfortable with using these laws and the laws are straightforward so to build a quantum computer out of bits the interesting thing that happens is is you put bits together the number of possibilities that you can store in superposition grow exponentially right with one bit there are two states with ten bits there are a thousand states two to the 10th it's about a thousand so every time you add a bit you have a sense doubled the power I put it in quotes of your quantum system so there's an inherent Moore's law there just add one bit and you've doubled the system every year we just have to add one atom that doesn't seem too hard okay so but this probably was realized even back in 1959 but one thing that wasn't realized was howdy how to make this how to make use of all these exponential possibilities because when you measure a big quantum system you get randomness you get noise so I like to say that quantum computing at a very high level is a good news bad news good news story and I told already told you some of the good news we can do parallel computing with three qubits there are eight states and they have eight weightings here and these eight parameters follow a wave equation that's where all the math is and we can get eight answers now eight is pretty trivial we can do that on we can do that calculation on any classical computer of course but with three hundred qubits just three hundred atoms two to the power three hundreds is more than the number of particles in the universe so that's the magic of exponential I mean nobody's wowed here this is more for a non physicist audience but you know getting access to those two to the 300 pieces of data is the subtle thing and of course the bad news is even with just three qubits when you measure the output of this beautiful parallel processor you get only one answer and it's random so it's as though you had no idea what the input was so a one-to-one function is probably not going to work very well in the final computer and most of what we compute today is our one-to-one functions so keep that in mind the final piece of good news I would say was codified by many folks including David Deutsch and I see Charlie Bennet's in the audience including he thought long and hard about entanglement before it was fashionable and these founders in the 80s and 90s showed that there can be quantum interference that can occur and this quantum interference is very hard to draw here I've tried to depict it with these orange dots and these blue things are supposed to be waves and and what what can happen is you can get massive interference where all the answers cancel except one or maybe a few but not exponentially many and then when you measure it and you may have to do it a few times not exponentially many times and when you have three hundred qubits you know there's no way to deal with all of those inputs and the quantum computer the magic is in these gates these operations that do not measure they just reference and then when you measure the output in some algorithms that can depend on all these inputs that's the trick and it's very vague because I think as I hinted earlier they're not so many applications you know that work here well the killer app has been known for almost 25 years now 25 or 30 years and it does hit on the topic of security and it's probably underlies why quantum computing is so important to many governments throughout the world well that application is factoring numbers factoring classically nobody's proven this but we believe that factoring is exponentially hard in the size of the input and factoring numbers while it might seem esoteric is of course the reason why data is secure according to many very simple remarkably simple encryption algorithms including the RSA algorithm Rivest Shamir edelman and the idea is well there's it's beautiful mathematics is about two lines but when you when you want to send a secret you're making public a huge number that nobody can factor because it has a thousand digits and we're relying on the hardness of factoring Peter shor this watershed event in the mid-90s he showed that a quantum computer if built with perfect operations and perfect quantum behavior would be able to factor large numbers in sub exponential time actually polynomial e with with the input now truth be told the polynomial scaling is not great it goes like the cube of the number of bits so if you have a thousand bit number you need about a billion actually there's a pre factor it's a thousand times the number of bits cubed so if you have a thousand bit number you need about a trillion operations that implies that you need a part per trillion errors on each individual operation and that's why none of us maybe there are some folks up in fort meters they're still thinking about this but none of us are really worried about Shor's algorithm happening tomorrow maybe in 30 years who knows who who can predict anything in 30 years but this is a long-term problem and the way it works and again this is a really high level the way it works is when you encode the number you want to factor like 39 into a quantum computer you need at least six bits to do this because two to the sixth is 64 and you store a superposition and you do Shore's prescription and you end up with a quantum wave function that's just a superposition of these two numbers and then you measure it and you get one of the factors it's been a very high level view of how Shor's algorithm works but it points to one of the commonalities of all quantum computations and that is they produce a simple answer from a maze of combinations classically factoring a big number you basically have to test all the prime numbers up to the cube root of the number at hand so yeah I want to pause here because this application is interesting is it's you can prove that a formal computer can do this problem fast now notice that we still can't prove that a classical computer cannot do it fast but it's strongly thought that's that that's the case but this is one of the few examples in quantum computing where we have a hard proof that it will work of course we have to build it it comes down to building machines to do to do something and it's it's one of the most difficult problems you could imagine on upon a computer so there's a good and bad about this it's written the bad is really hard the good is that we have clear proofs that quantum computers are good for something now there's another set of applications that only in the last 10 years or so have started to become interesting and the very good news is that this is a type of problem that is much more widespread than factoring or security its optimization I mean if you had to come up with a problem that has lots of inputs but only one answer you would probably say well optimization problems are like that because there's one optimum answer what's the tallest peak in this two-dimensional two-dimensional landscape well it's right here we can just see that because there's only two variables latitude and longitude but if they give you a function with 10,000 variables finding the optimal value of those variables that minimizes some non linear cost function is really hard in fact as a society we tend to ignore those problems where we invent we invent approximations heuristics so it'sit's it there's beautiful classical computing that attacks these types of problems but there are no proofs behind them they're heuristics now the bad news is applying quantum computers to optimization it's a it's a research field on to its own right now and I don't think anybody believes that a quantum computer can actually solve for the optimum of some complicated function however it may be able to approximate that that optimum better than any classical computer again it'll be a heuristic and maybe classical computers will be better and then it'll be a leapfrog so there's a lot of promise and what's great about optimization is that it hits everything any company in the world that has more than 20,000 employees probably has people thinking about palm computing what it can do for them but again it's a very researchy field a lot of these companies their interest in quantum computing is defensive everybody else is doing so we better do it so you have to you have to really really be careful about that but as an experimentalist I'm an ammo experimentalist I love the state of these affairs because we really need to build these machines and get them in the hands of users because I'm not gonna figure out the application but some user some software developer who might have read the laws of how quantum gates work if they can get access to these machines I think that's really where the growth of the field will come and in terms of optimization of course we don't usually think of molecules doing optimizations but the electrons on those molecules have a combinatorial number of configurations and they do to the laws of quantum mechanics that they follow they automatically optimize it they find their minimum energy and that's called the binding energy of a molecule but a molecule even as simple as as caffeine I think that's caffeine yeah it has about a hundred electrons or so that's still beyond the reach of classical computers to find the ground state we have beautiful heuristics that approximate of course we can measure it in simple molecules like that this one here is a very important molecule it's a it's a Natraj anis that we know that bacteria uses to fix nitrogen to break n2 and ammonia which sounds esoteric but again as a physicist when I learned this I was astounded that many percent of the world's energy is used to do that in a very hot and high-pressure process called the Haber process we don't understand how bacteria is able to do it and we think it has something to do with these funny you know molybdenum and iron core in these in this big organic so understanding the structure of this is interesting and and I'm also amazed that Microsoft has a big and probably the most capable team in the world that's looking at this molecule and trying to map it to a quantum computer and then you know logistics problems there's no shortage of these problems that we tend to ignore or approximate like the Traveling Salesman problem again these problems blow up when you when you add numbers of inputs so now I want to switch over to technology I think I set the stage that we really need to build these things to see what they're good for and my colleague bill Phillips at NIST in cadge a Qi in College Park he's fond of saying that that a quantum computer differs more from your laptop then your laptop differs from an abacus and what he means here is that these two computers really are both touring machines abstractly they follow the same rules of compute ones a lot faster than the other of course but but these two they don't follow the same rules and I think the corollary to this is why should we expect quantum computers to look anything like laptops and I'll even be more pointed why should we expect quantum computers need to be based in silicon or solid-state maybe but that's actually a horrible place to do really clean quantum mechanics if you want to scale every single transistor in your processor is different by almost a factor of two and it doesn't matter because you get to use feedback in quantum you don't have that possibility and you really need standards and that's I think why atoms are going to play a big role I've been talking almost exclusively about computing but the idea of using qubits and quantum information there are some old ideas here but there are roughly three pillars of what we call quantum information science simulation and computing but also communication I didn't point this out but if you really want to get around and quantum decription you can always use quantum mechanics to encrypt to store data in super positions and that way if somebody's eavesdropping you can fundamentally tell if they're eavesdropping that's one kind of killer app I say kind of because if you're a professional you really want to spy on somebody you tend not to go halfway in between sender and receiver you look behind the sender and watch what they're typing or the receiver that's what I'm told by the folks up the road there so so quantum quantum encryption itself it's interesting but it's a little bit of a esoteric thing however quantum communication if you have many nodes and you can share entangled states there's this is active research there may be protocols where you can you can have secure voting you can decide things where you don't trust everybody things like this based on laws of entanglement to me these are sort of tied together at the hardware stage because even with atoms if you want to scale to large huge systems there have to they're gonna have to be modular very much like multi-core processors we're gonna have to communicate quantum bits over a distance it might be across the lab or across the chip it might be might be across the country so these the technologies for these two areas are very similar quantum sensing is something that I would say is in a sense has already arrived we have sensors that have quantum limited noise floors and even below that I think David Reid see this morning talked about how there is quantum mechanics they used in the LIGO and the advanced LIGO source where they're using squeezed states to to get below the shot noise limit and even though it's only the factor of two or three it allows you to look much further into the universe for gravitational wave sources those ideas are closely tied to quantum information there's certainly a linkage there so since Shore and leading into the 2000s this this areas of course exploded certainly theoretically and this is a count it's a few years old now on the number of papers per year in these fields and and they're growing substantially I dug this out and talking with Phil a few weeks ago that there's a famous paper following up on Stein's EPR paper in 35 John Bell had the idea that there's a mathematical identity that you can that you can show that that makes entanglement seem even weirder in a sense that even if there's a completion to quantum mechanics that doesn't have this spooky action at a distance they're still going to be nonlocality and they're they're codified in terms of a bell and equality that he pointed out in this paper and this is the number of citations in his paper over the years and this is a log plot of the same so you could argue whether it's starting to saturate but this is a Moore's law of citations I don't think there's any other paper that has grown over many decades with a constant exponential kind of interesting okay so getting into the laboratory you know there are many quantum systems that we could think of to build these these types of hardware's quantum computers and so forth anything that any lab that needs to use quantum entanglement or many-body physics to study their system is sort of a candidate to think about building devices out of I would say however that the the two at the top here shaded in red are only the ones so far that have been built into systems meaning that these are systems sort of like a Turing machine where we can think about abstracting away the hardware and unleash the power of these devices even though they're very small still two programmers people who know applications and a lot of these other platforms they're still in the research stage some more promising and than others I would say topological qubits in particular is very researchy but it's beautiful condensed matter physics the idea that the qubit can be stored in a different space in the topology of a circuit and there's wonderful research going on in all of these fields but I would say these two are currently being built into systems composed of dozens of qubits which embarrassingly is still still pretty small but dozens is better than one or two and if you're building dozens you're getting an idea how to scale you're learning that the you know the the whole system is more than just a sum of its parts and those are two technologies that are being developed not surprisingly at many companies some very big companies here even even the atoms I'll talk about a little bit are little startup and also Honeywell has gotten on the scene so the the real problem I would guess with with with atomic physics in an industry and that applies to the trapped ions maybe even diamond vacancies in in in sorry envied vacancies in diamond which is an interesting sort of solid-state atom and neutral atoms anything in photonics is that the laboratories to support these they're their laboratories that's the problem there these are not this doesn't look like a system and in fact all of the action is taking place in an you know in a cubic foot right here there's a vacuum chamber and an ion trap and individual atoms are in there almost all of the hardware and those of us in amo physics recognize tables like this most most of us put this slide lock like this up here and say well look at me isn't this amazing I'm really embarrassed by this this is horrible it's very hard to make progress in something like this well the good news is like any system if you know what you want to build you don't need every little mirror to have three screws on it you can you can hardwire things if you know what you want to build so we've been on a journey over the last many years to do that but building a small system in the lab I mean everything starts from from the physics laboratory and building a small system of only a few qubits has been has been very interesting in the last few years it's changed my outlook on doing physics I used I still am an atomic physicist but I'm not learning much about atomic physics anymore I'm learning about high-level physics that we can do by putting together many qubits and controlling them these students they can calibrate the system in the morning and then they can run the system for several hours without knowing they're doing atoms they're doing quantum circuits so that so of these many applications these were mostly done with collaborators who said do you know you can do this with your system oh I had no idea about that and this is the kind of physics that I find fascinating because I'm now in school again learning about all kinds of things one of my favorites is experiment we did a few years ago having to do with the scrambling of quantum information and quantum scrambling you may have heard this term in the last few years it's something that's thought to occur in black holes if you put in from if you put a bit into a black hole it sort of gets scrambled really fast meaning that the information gets lost or does it get lost because we know black holes that and it may be that the Hawking radiation can be correlated with one part of an entangled system an entangled pair that you throw into the back hole and even more recently it's been posited that entanglement could allow us to at least toy with theories of quantum gravity and even wormholes and a wormhole allow is sort of the connection between entanglement and wormholes is really profound and again I'm not even a graduate student in some of these areas but they're you know they're fantastic and folks like maldacena and Lenny Susskind and John Prescott and Pat Haden have put put these things together well another reason I like to point this experiment out is that normally I'll hear some the audience and he will receive the vallée prize this evening and he's one of my favorite collaborators because he calls us up it may take up to a year to be convinced because oh you really should do this you really should do this and this is one of them he said oh you can do you can detect scrambling and I won't go into too many details here but we made a seven qubit circuit and what we're doing again lower Lawyer sites we're testing whether three bit unitary scrambles or not so in a sense where we're doing black hole physics with three qubits I know I should I can't say that with a straight face but the idea of scrambling having a litmus test for scrambling is really hard to do and norm and Ben EO shita from perimeter and other theorists showed that if we implement this circuit unit angle a bunch of qubits put them through some unitary it can be any unitary and depending on certain measurements here if that unitary scrambles this arbitrary input state all the others are zeroed it will be teleported or it will be copied I shouldn't say copy we can't copy if you if you copy the state you have to erase the original at the same time and no-cloning theorem and Charlie Bennett here and who knows very well about that and in a sense is the inventor of the teleportation concept itself and sorry Charlie it's a weird term teleportation because people think of Star Trek but it's a perfect term in terms of sort of the disembodied movement of information from one place to another that's exactly what it is at a single qubit level and if this teleports then you scrambles and I won't get into the details of what we implemented for you it was a variable circuit that we knew would scramble for certain parameter and wouldn't at other in norm showed us the circuit to do we implemented and it kind of worked if you're above 50% the teleportation is you can show that the teleportation can be purified so it's quantum and indeed as this you became more and more scrambling we were able to show the teleportation work better and better there's noise and errors and that's why I didn't go all the way to 100% now recently this circuit was was was pointed out by folks at Google X and Stanford Caltech Princeton and my colleague Brian Swingle at Maryland to to actually extend this by doing more tests and and and making a connection between these these ideas of entanglement and wormholes so you know I wanted to point that out because the physics coming out of those heads is so far removed from my experience that it makes it super fun so I want to point out also that that lab I showed you before without much exaggeration has been shrunk and simplified into a real system in fact this system is much more powerful than that lab I showed you before this will this this has a template of about 32 qubits which is still small 2 to the 32 is about it's about a billion so that's still not big enough but when we get to a hundred qubits that's gonna be too big to be able to model any other way so we're getting there and I'll point out that this this this experiment is at the University it was funded by ARPA and and this merits a few remarks I ARPA is the DARPA of the intelligence community and they're they're there they sort of do research that's of relevance to all of the intelligence agency from NSA CIA and so forth they're very interested in quantum computing for obvious reasons but this is open research and in fact I would say that I our per program in quantum computing is the biggest open program in the world and it has been for many years it affords us the ability to make systems like this and we collaborate with Sandia National Lab that makes the silicon chip that supports these atoms there's 88 erbium atoms floating above this chip about a tenth of a millimeter and other companies we're integrators we we tell them we love your product could you do it this way it's very expensive I our Papazian to do that and we integrated in our systems I'll also point out that I ARPA of their of this largest program the world half of the funding goes abroad to two teams in Europe actually but they're funding anybody they don't care I often I often get a little annoyed at that but I sort of get it I get the bottom line but this should be I would like this to be spread out a little more that that us even the intelligence agencies fund science abroad in a very big way now I mentioned inq at the beginning and this technology that's inside that box it it's the basis for systems we've built down the road at IQ it's about a half a mile away from my lab it's off campus and I think I said there about a hundred employees there now we've built for full step quantum computer systems rather quietly not secretly but but you know we're not we're we're busy heads down building these things and yeah sorry the pictures are so dark here but there's not much to see it looks a lot like this box what you can't see is the software and the FPGA engineering all homegrown a tank you to control these things that's where the entirety of our challenges it's not the quantum bit the quantum bits done will never improve it these are atomic clocks way better than we need them to be it's all about the control of them from the chip itself the electrodes underneath it and the lasers that that push these atoms around this is an example of the control we have and I and Q we have a camera that images each ion as it's being loaded this is a real-time we load from one side of the chip and then we merge them in with atoms they're already there and we wanted 24 exactly 24 and with an 80% probability each time we do get one we can count how many there are there in the zigzag we straighten out the chain with by modifying the electrodes refocus the the lens now ready to go now we can do an algorithm and we have 32 beams that are aimed on each of these ions for this one we only needed 24 it's totally reconfigurable there are no wires here there's lots of wires over here but there are none near the quantum system that's really important if you have wires near your quantum bit you're dead because now you're doing solid-state physics and you can't locate that stuff our wires our wires our laser beams now laser beams are also noisy so we're in the business of making very clean laser beams we're not usually much fiber optics yet but that's in our future we're not integrating optics on the silicon chip that's in the future we have a ton of technology trains that we're gonna bring into this and that's why I say when the big companies get into this technology then you should start to pay attention okay so I'm gonna skip over a few slides really fast and note that I would say 2020 will be an interesting year I don't we probably won't find the killer app of optimization but two very big cloud providers Microsoft Azure in AWS have announced that they're going to offer anybody that subscribes to their services access to hardware on the cloud and Microsoft is using these three types of hardware qci is a start-up in Yale based Rob show coughs company based on superconducting qubits IANA Honeywell are based on trapped ion qubits very briefly you know I've sent eight or nine of my students to Honeywell outside of Boulder and they have a team I think it's up to a hundred or so it's it's a it's a it's talked about competition it's it's great that they're doing this because it sort of validates us we're just a startup and they're a huge corporation they were very secretive for a while and what they're doing for actually very innocent reasons and they're there now much more open they go to conferences so it's it's actually a very nice situation I would say AWS has our system and also a couple of superconducting systems that they can they can put on their cloud so if you don't you know you shouldn't as our CEO tanki says you shouldn't believe any of us when we're talking selling qubits use them test them out even on very small circuits and and I think the year the next year or two it'll be very interesting to see what happens you know maybe somebody will hit a home run on the algorithm side and even though it's a small system most say you know when you scale this to five hundred qubits you know this is gonna be very important but it's it's super critical that people use these machines and computer scientists are apply some of their yes some of their tricks in compiling and so forth I mean we're programming not in a language really it's more at a gate level it's very primitive it's not even assembly language yet so we really need those developers to start to use these things and tell us how to build the next generation okay so I want to step back and and talk a little more maybe internationally about this field I think it's attracted the interest of pretty much every major country with a research budget for many reasons I think I alluded to it's super promising both commercially and for security and and I'll say it again I think the security aspects are you know very long run a little bit overstated I think the more interesting thing is the economic security of some of these potential algorithms how they will help companies optimize things and and of course governments and and departments of Defense are very interested in using them for optimization programs of all different types so this slide I apologize it's a little outdated and you know you can tell where it came from I think it came from the Economist through McKinsey but Peter Knight is here and I would say the UK was really the first nation in the world nationally to put together an initiative and it really made waves across the world that was in what 2014 or 2013 yeah in the EU as a whole the individual countries have some outsized investments Austria Denmark the Netherlands and again these numbers are really outdated China of course has a has a big program in quantum information they have for many years they as like everything there they have a very steep slope and we've heard numbers ranging from a few billion to a few 10 billion on their investment in the coming decade not sure exactly what that number is but it's a very serious effort they have there and the European Union following on the lead from London has a flagship program one of their very few scientific initiatives that are continent-wide on quantum information that will double down more than that on the individual investments from the countries the US only there's a little bit of a late starter in the idea of a quantum initiative part of this is fitting because in the u.s. we have so many different agencies with different missions that it might seem that we're very disorganized from that perspective on the other hand all the missions know what they want to do nevertheless there should be coordination I mean the defense and the intelligence agencies have really taken control of this field in the US since the 90s and they continue to really have the the you know control this field but because I think the interest in the fields moved a little bit from security to Commerce in general algorithms it makes sense that the civilian agencies like do e NSF NIST who has always been there that they start to start to coordinate their their efforts here and so the National quantum initiative because I'm trapped inside the beltway in my workplace I ended up playing a big role in in in helping helping the government formulate the idea of a national quantum initiative and it was passed in a little over a year ago it's an authorization bill and there are dollar figures in that bill but it's only authorization so the agencies are rightfully a little nervous about what that means but I'm happy to say over the last over the innovating time all the agencies are really are really going after this internally and we totally expect appropriations they've already started for the agencies and I'll also say without spending a nickel the National quantum initiative has given impetus for almost every University to spin up their own effort usually from the physics department often it's more broad and I think it should be but these universities are coming up with their own their redoubling their efforts on hiring people in the field students are coming to it so it's already had in effect an immeasurable effect in terms of dollars universities are hiring people I think it's a really good thing that this happened and I have to put the picture of our president signing this in his late December of last year and OSTP has opened an office we'll sort of coordinate the activities of these agencies and I'm happy to say that two of our own actually from atomic physics in condensed matter physics Jake Taylor and Alex Cronin on loan from the NSF are now in OSTP helping to coordinate what's going on so you know I won't talk more about the nqi but it does direct these agencies or authorized these agencies to go in different areas the do-e labs in particular because the OE is used to having big laboratory efforts that are sort of a halfway between a university lab and Industry it makes sense for dle labs to build these things and have some longevity do real engineering which is very hard to do on a university campus and I should don't so I think I've I've said a lot and from for my final slide I borrowed this cover from Technology Review magazine a few years ago the topic was AI I think or blockchain I forget what it was but it might as well have been quantum computing there is there is it's it's impossible to overstate the amount of hype in this field you put the keyword behind everything and and it's just one of those sexy words I guess but I will say more than just hope there are problems out there remember there's two to the 300 exponentiation there are problems out there that will never be solved on classical computers because the universe is not big enough even if every atom were part of a big cosmic computer it wouldn't be powerful enough that doesn't mean a quantum computer can do it but if that problem those types of problems are ever solved they will involve a quantum computer that's for sure so that's not hype that's that's fact but it's also you know it may be that they aren't able to solve it remember that so anyway that's my kind of perspective on that and I think what's best about this field is that we it's it's very researching we don't know what this stuff's going to do so we have to build it and see thanks [Applause] we have plenty of time for questions and I think I'm gonna take some prerogative and ask the first one that's okay so Chris you said that the fields very international your your bubble chart there from McKinsey showed that China is next to the u.s. leading in size of investment so how has that Chinese interest in quantum computing been affecting your work in the sense of the subject of this meeting so my University is that one of the bigger programs in physics in the country you know very much like yours and so it's very competitive for graduate students to come to Maryland I think we get 700 applicants every year and for 35 spots or something we've noticed a little lesser the extent of the decline in Chinese applicants but it is noticeable we see it and you know getting I mean its usual thing I mean you'll say it as well you know getting and getting very you know top-level graduate students we cannot just simply turn off the international cooperation there now as to your point about the Chinese investment internally so I usually I usually go to China every year on my way back from Singapore mainly to see my colleague looming Dhawan who is that my university and and Phil and I know him very well we recruited him to Michigan back in 2004 and he is a quiet uber leader in quantum information when he was in the u.s. he left Michigan like two years ago because he wanted to start an experimental program and he couldn't get funding here and unfortunately now he can't it's very hard for him in practice to travel so the only way I can see him is to go there and the reason I have to see him is that he's he's behind many of the things we do at a fundamental level just anyway so that he's a he's a national treasure it's a shame that to lose him but in terms of the Chinese approach in quantum information as an initiative you know we all read their opening a very big sort of a do a scale lab in half a quantum laboratory led by John way pond and they had a sort of a Time magazine cover type experiment where they beam single photons to a satellite and I think it went into Vienna and so forth and they were able to do this quantum encryption through through satellites very expensive operation but also maybe you know I mean it was very splashy but I'm not sure where that's going in terms of quantum computing I think they're way behind us everything I can see notwithstanding looming Duan zephyr and but that's a research effort and I think in this field because the big companies in the US have gotten involved they're hiring students students are coming into the university because they know that they're going to get many job offers from these big companies and a smattering of startups I think that has been the savior I think for for us in the US but also the research Enterprise it's it is a field that's funded very well from different agencies that have different missions and the NSF now is rightfully grabbing onto this field in a more open way and in more blue skies way as it should so you know I have to you know I hate hate hate to sound you know like it's not an issue we have our eyes on but I'm not all that concerned I think there are very few leaders in this field in China and I'm no I'm being recorded here so I listed I listed them charlie Jeff a question [Laughter] it's an anti de sitter space so that's been I'm sorry if this is a stupid question but you are I don't understand exactly what your qubit is I mean I tend to think of like you know flipping a spin or something but you said it's an atomic clock how do you get a qubit out of a time an atomic clock yeah thank you thank you I should have maybe at the outset mention that so an atomic clock is a two-level system and the energy separation between those two levels divided by Planck's constant is a frequency so if you have a two-level system you can define a frequency and if that's a very stable frequency it can be a standard and these you can call it an effective spin if you want each of these atoms has a nuclear spin and an electron spin and they couple there's a hyperfine interaction so these are two two levels they're hyperfine levels in the ytterbium ion i'veeen and say you terbium the splitting is twelve point six gigahertz and I could list about eight more numbers because we know exactly what it is it's perfectly replicable right it's a microwave transition so the even though the the the true ground state is one of those levels we have an excited state you might say well that's going to relax it takes 10,000 years for that to relax so it's very stable and we can prepare a superposition using NMR techniques basically we use optics to do this optical beat note that's precisely match to the hyperfine splitting so it's a very mature technology to control these types of atomic spins now what I didn't say also these are ions that's why they're forming a crystal and when we when we push on one of them with a laser beam they all know about it and that's how we can hook them up without wires there's there's electrical there's a Coulomb interaction so in a sense that's a wire but there's no solid-state material anywhere except really far away that that can find these electrodes they can find them Ambrose from the American student physics I'm curious what advice you'd give to anyone else from a different sub discipline in the room who would want to go to Congress trying to advocate for the national you know insert the blank here initiative you know how can you just reflect on that process and how you know following the authorizations through following up with the appropriation and you know what advice you give to others yeah I remember Phil Phil called me one time he said happened I mean is so fast well you know when I moved I moved to the area from Michigan in 2007 and I think I'd been here of two years and in the in the capital they just renovated the atrium area on the east side of the Capitol and you know they have the old flag there and so forth and I took we'd come down a couple times a year the whole group to just goof around in the mall and have dinner and the whole research group came into this new atrium and they had congressional staffers that were touring everybody and we had a big group and we had a staffer that was helping us long and and he asked me so where are you from I said no we're just from up the road in College Park well what do you guys do well we we do quantum computing and he said oh yeah I've heard of that yeah we hear that on the hill all the time and nobody does anything about it that's that's what he said that was 2008 so probably through the intelligence agencies interests and the idea that that security could become an issue if we have big quantum computers it's captured the attention of many many on the hill and so they all know about it and again I think that might be I mean it was a good catalyst to get things going I think maybe that's not gonna be the killer out in these fields but they know about it they also know that they know what Moore's law is they know it's ending they know that you know a lot of the the staffers many of them are just science experts they you know that's what they're that's why they're hired by the the member of Congress to advise them on science technology and they know that Intel Corporation for instance you know thirty years ago they knew exactly what their going to do in 10 years they could plan how small the system's gonna get they're not doing that now they have no idea where they're going to be in 20 years on that Moore's Law curve it's getting for the first time ever I think a few years ago it cost more per transistor to build it when they got blow I think about eight or nine nanometers so this is an economic issue if we want to enjoy the economic growth we've had over the last 60 or 70 years we have to be on top of different modes of computing quantum is one of them it's not the only one so I think that's the reason the appetite was there for many years and I think because it's been in the research field for over 20 years the time is right companies are starting to play ball they're starting to make their own investments and and I had a slide with some of our partners that helped us do this you'll recognize lots of big companies here that helped that helped these congressional committees see the importance of this and I'll also say as well that the the the national photonics initiative which is which was stewarded by the optical Society of America and SPIE had a big hand in this they have a professional lobby team and you know these lobbyists I guess I know I'm on camera I'll say it anyway the lobbyists are a little like criminal lawyers you know they have this tinge of well I won't say the word but when you need one they're really good I mean they know how the place works I have a newfound respect for lobbyists I mean they know who's doing what who's interested in this I mean that's a very important they get a bad rap in the press but having access pretty much full unfettered access to a team of professional lobbyists that was key and that was again the National fotox initiative in the osa NP I was a National Academy initiative that was funded many years ago and my Cramer and I kind of took them aside you know you might steer them on to quanto and they totally agreed and bought into it so that's all those things I think made for the you know the right time I'm saying hacker from Stanford and formerly Los Alamos sorry and first of all fascinating talk I'm curious about what you foresee is the interplay between the applications and the machines because if you if you look back at the history of classical computing and particularly supercomputing you know as we called it now is all the way from the ENIAC that you showed or the maniac you know at Los Alamos you didn't show that maniac the ENIAC stretch and then the crane machines it was essentially the nuclear weapons laboratory Los Alamos Lawrence Livermore that drove the development of those computers because they had problems that were big enough that required those computers and then they had people who knew how to use them in other words how to write the software you know Seymour Cray came to Los Alamos in 1976 and he said hey I've got this big machine but I don't know what to do with it yeah you know can you guys help and so we essentially helped the write what it took and then of course it went from there and started to spread out more with the connection machines and massively parallel processing but these laboratories still played a very big role in the application then pull the computers so what do you foresee for this area you know you mentioned optimization and of course encryption and so forth but I can't quite envision as to how this you know this cooperation between the users and the machine people is gonna pull this forward great great question I think it's a real challenge on the one hand because these are not just a new generation of devices were used to even the Cray was based on classical technology it was a Turing machine the rules required to program these things some people just disregard them you know I deal with this at academic engineering departments they say oh you physicist quantum mumbo-jumbo that's never gonna play a role but we need badly need engineers not just to build them but computer scientists too to help us understand how to how to write programs in them like I said we're not even at assembly code language and this has been this has been a challenge NSF has been on top of the game on computer science they've really tried to stimulate academic departments for hiring they even have a program that will pay startup and maybe salaries for a few years for new computer science hires in quantum but you know actually my daughter is a computer scientist major at Maryland they're one of the fastest growing computer science departments in the world they have to hire six or seven new faculty every year because they lose three and the growth is incredible they can't afford to hire a quantum computer scientist is they have to I understand those pressures but it's not just going to be on the academic side like you say national laboratories have problems that are hard and I totally buy what you're saying that we're not going to find a use case without developing machines and learning how to build them these are not a commodity you can't I mean we take your team Lee take 10 megabyte pictures on our cell phones these days because memory is cheap it's a commodity you couldn't imagine that 30 even 30 years ago so qubits are not a commodity gates are not a commodity we need to co.design at the very bottom level at the atomic physics level we need to be able to understand algorithms and that's something I mean some of us are trying to reach up the stack the people in computer science are trying to reach down but it's like an ASIC an application-specific I see that a Cray I don't actually should know more about the Cray machine but I know that my sister is a computer scientist Los Alamos works on turbulence and so forth and she's interested in probabilistic computing which is another form of computing and again they have to co.design I took that term from her and I use it to call on quantum all the time we have to co.design our machines to the applications and we have no idea what those applications are and so it's a big bet to have a startup well actually it's not my money it's venture capital money they're making the big bet that that we're actually going to stumble upon something but it won't be us it's going to be users that tell us how to build the next generation and wire it in just such a way that we can we can you know wring out every ounce of efficiency because we have to they're not a commodity we can't waste anything we can't waste memory and can't make qubits and so forth so I I'm not sure I really answered it very well but I sort of agree this optimization stuff I think everything turns out to be an optimization but you have to really dig into the particular application and see what hardware works what type of a gate set there are different gates we don't use demand the NAND is universal in classical computing is it's so easy to wire together we have a variety of different gates at the quantum world that depend on the hardware but they should depend on the algorithm too so yeah thank you for that I was able to say lots of things I wanted to put in my talk oh go to coffee break and be back in 30 minutes for the panel discussion on international engagement and competitive physics fields [Applause] thank you that's fun

Channel: APS Physics

Views: 1,471

Rating: 5 out of 5

Keywords: international science, physics, international leadership forum, aps

Id: cYQ7TGG3-C8

Length: 60min 11sec (3611 seconds)

Published: Mon Feb 03 2020