Quantum Error Correction with James Wootton

Download MP3
Welcome to another episode of The New Quantum Era Podcast hosted by Kevin Rowney and Sebastian Hassinger. Today, they are joined by James Wootton for a memorable conversation about his wide-ranging interests in a big swath of topics, including the fascinating growing quantum computing hobbyist community and James' passion for education and outreach. James also got into some really interesting material on quantum error correction (and mitigation), and a fascinating bit of math called the threshold theorem.

Kevin Rowney 0:07
Hi, yeah, so I think this, this interview with James Wooten was a kind of a memorable conversation for me because of his his wide ranging interests in a big swath of different topics, including the the fascinating growing quantum computing hobbyist community out there. Also James's passion for education and pedagogy and outreach. That's a mission that I believe that I think Sebastian does as well in the QC domain. And finally, you know, he got the NS to some really interesting deep material on quantum error correction, or error mitigation, and this fascinating bit of math called the threshold theorem. So I, I really enjoyed this one.

Sebastian Hassinger 1:11
Yeah, me too. If you're Kevin, I really liked James, I first became aware of James, as I mentioned this in the conversation, through his postings on quantum games, which I found fascinating the idea of using quantum computers for gaming, you know, gaming has been game development has been sort of an early adopter of almost every new technology that comes along. And I think it's because the spirit of play is a really effective way to both provide an entry point for as you said, he's really interested in education and brought bringing people into the community, but also experiment to experimenting with, you know, novel approaches, and really not taking the current assumptions as givens and experimenting with it around the edges of what people think is, is possible or feasible or correct. Areas of Focus. And I think that's the other fascinating thing about James is that the the interesting interplay between that gaming and pedagogical area of interest with his sort of professional area of passion, which is quantum error correction, which is extremely esoteric, complicated problem to solve with, with know, a bunch of potential solutions, but no clear winner just like qubits themselves. So we get into that a little bit. And I think that was really interesting introduction to a topic that I'm sure we're going to revisit with with future episodes as well. So I really enjoyed it too. Good stuff.

Unknown Speaker 2:38
Well, without further ado, here we go.

Sebastian Hassinger 3:16
Great, so today, we're gonna be talking with James Wooten. And James, you want to start by introducing yourself.

James Wootten 3:23
I'm James Horton, I work for IBM Quantum. And I have the background of having a PhD in the field of working on it in research for many years. So mostly, my research has been on quantum error correction. But also, I'm very interested in sort of education and outreach. And in fact, even before I joined IBM Quantum, I was involved in some outreach projects. And that's kind of why I gravitated towards IBM Quantum, because they were the ones with the device is on the cloud at that time, which I thought was a really exciting thing. So yeah, I guess I

Sebastian Hassinger 4:00
think that's that's how I think I knew you before I knew of you before you joined IBM, because of explorations of sort of game development using using the public IBM Cloud quantum systems. That was probably like, I think you made some medium posts or something about that very early on, right.

James Wootten 4:19
Yeah. So well, like that. The long story is I was in a Rosa, which is the birthplace of quantum mechanics where Schrodinger came up with the wave equation. And I had this idea to make a game based on my research to be the heart of a, an outreach project. So that's what I was doing in 2016. And that's the kind of mindset I was in where the IBM put devices on the cloud. So I straight away thinking, well, I should make a game that runs on these. And then not so much because it's, you know, an example of quantum advantage and everyone's going to be falling to use this in the game industry. But yeah, as you said, it could be the heart of a blog, you could have a blog say I made a very simple game. Right? It's how I made that, therefore demystify using quantum computers, though.

Sebastian Hassinger 5:06
Yeah, you recently tweeted about, I think was in response to somebody else, pointing out that the early days of classical computing were largely about appealing to hobbyists, right? Like Apple's first computer was essentially for hobbyists, and Microsoft's first products were for hobbyists, micro basic, basic, in particular. And you responded, I think saying you expected hobbyists to be or the field to be more accessible by hobbyists, I guess, is the way you would put it, right.

James Wootten 5:36
Yeah. So even going back to the 50s, and 60s, it wasn't so much a hobbyists, I suppose, because you had to have a certain status to have access to a computer, but the way people were using it was very much like that, yeah, hacking culture that we have. So later on, though, people were very interested to push these devices as far as they could. And so I expect that there should be a passion of for people to do the same thing with quantum computing as well,

Kevin Rowney 6:12
that always feels like a vital part of any sort of like rising tech movement is if there's a subculture of people who are really actively playing and, you know, really involved in creative way with pushing the edge on even economically irrational outcomes. But still, they're just delightful intellectually and socially. Yeah. So Right. It's such an area in quantum computing for, you know, hobbyists to get into I mean, if you're already mathematically inclined, it's not a it's not an easy walk in the park to get up on this material. So I mean, it's, it's interesting to contemplate how maybe there's other routes of entry on pedagogy for for this for this material.

James Wootten 6:51
Yeah, that's certainly the kind of theory that we've been doing in quantum computing. And in the theory behind making the Mueller making the devices, and also using the devices is quite abstract. And it's not something that a hobbyist can easily pick up. Yeah, but the people who are making the early computers say we're using methodology, which is not what we do now will be used by junior. So if we can get people to get doing some of the minimal things that are interesting, then they might be able to define new ways into the field ones that we with more experienced, never thought about, and therefore find ways to use it, which are more accessible. So I think there are certain elements like a bell test. So this is a test of the uniqueness of quantum mechanics proving that Einstein was wrong about what he thought about quantum mechanics, this is something that I think is actually quite easy to explain, and very easy to reproduce. So if you get someone proving Einstein wrong, then you've already got them to a fairly good point. And maybe they'll find their way from there. So you're

Kevin Rowney 8:02
referring to the spooky action at a distance kind of controversy that against them as well. But yeah, really. So that can be a very powerful lesson. And so how do you? How would you invoke that for somebody who's not maybe into the mathematics and the underlying physics, what would be a way to sort of get them to grok? The intuitive understand,

James Wootten 8:23
so I often try to avoid the the using the notion of superposition, because I think that just kind of conjures up the idea of randomness to most people. And also, it's really, it's referring to something that's going on in the linear algebra. If you're not teaching people, the linear algebra, you should not be using the term superposition.

Unknown Speaker 8:45
Usually those those words, yeah. So

James Wootten 8:49
I have a way of kind of explaining how a qubit works in which I say it's like a bit but with a bit, or you can do is ask it, whether it's zero, or one. But with a qubit. There's multiple different sort of ways that you can access that information. It's almost like there's different functions you can call to ask it, whether it's zero or one. And depending on which you choose, you'll get a different result. And then that's a starting point as you go into notions of the uncertainty principle, and also the the the sorts of correlations. So I see in the qiskit textbook, we have this interactive visual game thing. I wouldn't really call it a game because the game has to be more fun than this. This is unashamedly pedagogical. But it's game inspired. And that I did it so of course, I'm going to say it's good, but I think I did a fairly good job of, of getting people to the point where they can perform a vowel test and and giving them the intuition behind it.

Sebastian Hassinger 9:53
Yeah. So you had mentioned that I think the the Open Source Textbook that you try to write the open thing sections of that without relying on on as you say linear algebra or other advanced math or physics concepts, right? You just sort of tried to take a conceptual approach. Yeah, I think

James Wootten 10:09
that usually the first thing that people do in a textbook on quantum computing is say, Sy equals a zero plus b one, here is a simple equation. And most people are going to look at that and think, What if that's a simple equation, close the book. So I think it's better if we start off, building a little bit of intuition, and then bringing in the maths, I think it's important to bring in the maths. But I think it's better when the maths is describing an intuition that people already have. And that's the starting point.

Sebastian Hassinger 10:52
The math SIRs are snuck in after you have a conceptual basis to understand why

Kevin Rowney 10:57
it's so refreshing to hear you speak this way. Because it'd be I think a lot of people who actually are in math and science culture have, what they often suspect is true is that many of the people operating at a high level, actually do have an underlying kind of like non rigorous, intuitive metaphor to understand the subject of study. But they don't just need to explain that. That intuitive metaphor, they just go through the formality of the definitions and rigor and proof. And so it's, for a lot of people, it just feels I'm sure, refreshing to hear you. I don't know announced this kind of perspective on the material.

James Wootten 11:31
I think a lot of people probably have have done quite simple mechanics. So they maybe know how to multiply velocities and times and things together to do the trajectory of a ball. And even that probably seemed quite hard at the time. But at that point, you knew what happened when you were born? Right. So you had to base your intuition on starters. Yeah, that's just throwing both spinning qubits. I think it's also more good to have a interest.

Sebastian Hassinger 12:07
That's a really good point, because your point about superposition before I feel like in a way, it also gives rise to the sort of fallacy of like trying all the possible answers at once kind of the bad explanation for how qubits work. So you've done a number of sort of game jam game hacking type of events, what are some of the things that you've seen that that were sort of like super inventive ways to use quantum computing and in sort of a game context?

James Wootten 12:40
Yeah, so you have different types of game that you see in Game Jam, some of them, people make a game without really thinking too much about the quantum and then just use a quantum computer with a random number generator, right. And some of them are more inventive in the way they sort of approach the pedagogy. So their idea is to try and help people understand quantum computing. So this is often someone who's learning quantum computing while they're making it. And they're also finding a way within the game to teach other people and let people explore what they learn and get some idea about the simple basics of measuring and different bases or superposition or entanglement. So I think the sophisticated games are usually ones that aim to be a little bit educational. And what we haven't really seen so much as people really trying to take in the challenge of using a device, I'm going to use a device and I'm going to use it in the best way possible. I haven't seen too much of that. But educational ones are more sophisticated.

Sebastian Hassinger 13:49
Yeah, in fact, the the world quantum data is coming up in I think mid April, if I'm not mistaken. There's there's a couple of of quantum educational games that are coming out of a group in the University of Chicago, that look sort of, to your point is sort of the sophisticated game that is trying to teach a quantum principle rather than using a quantum computer to do so. But you've done I mean, like you have the quantum emoji generator, right. That's one of your projects.

James Wootten 14:20
Yeah, or with this another thing, which is just you don't need too much quantum computing, to to get experimenting. Because one thing you can do with a qubit is just use it as a bit. And so you can write a bit string on it, and that bit string could represent by the time I had 16, a 16 qubit device to play with. So you can put 16 bits which is enough for two ASCII characters, right? So you can write an emoticon but you've also got superposition. So how about just superposing two emoticons for the fun so they're they're similar think that even people without too much knowledge and who have already watched, read the first couple of seconds of the qiskit section so the kids can textbook can start playing around.

Sebastian Hassinger 15:09
Right? Right. Right.

Kevin Rowney 15:12
I've also seen up on Steam, but I haven't downloaded yet this interesting thing called Quantum chess. So have you have you heard of this? This this game, it sounds sounds interesting as a powerful way to maybe illustrate some of the pedagogical ideas and still have some fun.

James Wootten 15:28
So this is something that's been around for a few years. If you only know from Steam, you should check out the YouTube video in which Paul Rhoads plays Stephen Hawking. This is the kind of support but

Sebastian Hassinger 15:43
I think it's the pinnacle of the sort of celebrity involvement in quantum computing.

James Wootten 15:51
So So yeah, so way of using superposition and entanglement in a game. So there's, there's others. But yeah, I think that's the one that's got

Kevin Rowney 16:01
it is just a YouTube video, accurately present the the underlying rules of quantum chess I mean, it sounds like it's a really, it's a game of substance with real rules, and, you know, consequences for your movements. It's not just some imaginary story, right? Yeah.

James Wootten 16:18
I think you have to play the game and look into the physics behind the game as though it's a starting point in which you can understand quantum computing. I think if, if people would just watch the video or play the game, casually, they might just seem like random. The Cisco remembers the

Unknown Speaker 16:34
point. Yeah, right. Yeah. Right. Yeah. But there are other games

James Wootten 16:37
that are more explicitly based on putting down quantum gates. So there's another big one coming out at the moment, which is quantum Odyssey, where you're really putting down quantum gates and you're, you're seeing a time evolution of the quantum computation. So that's something where even if you played it casually, you have, you're kind of forced to understand what's going on, because that's how you progress.

Sebastian Hassinger 17:03
There's also that great Minecraft clone called Mind test level that Jim Weaver created, which is called Quantum blocks, I think is what he called it, right? Where you're actually solving qiskit blocks, yeah, you're actually solving puzzles, by constructing gates into a circuit to produce the desired result, which I actually helped Jim, to present or whatever, do a seminar with a bunch of middle schoolers in San Francisco a few years ago with that level. And it was remarkable to see 6070 middle schoolers sort of rip through those puzzles, because they just saw it as well. This is Minecraft. I know how to do this. There wasn't a moment where they might like, this is not intuitive. I don't understand how this works is like Minecraft. Yep. Got it. Yeah, it was great. It's

James Wootten 17:56
an amazing effect when you package these things up as as puzzles for kids, because we had this. When I was at University of Basel I, I collaborated with IBM to make Hello quanta, which was another game based on using quantum gates. And you give it to an adult and they'll stumble. But there was one day when these kids were coming in on what we're supposed to do Tuesday, we had to teach them about quantum, they would come in for 15 minutes, I'd give them the app, it all just complete the app. Right, they'd go off. No problem at all. Yes.

Sebastian Hassinger 18:32
Yeah. Yeah. I mean, it's moments like that, that makes me sort of, I think, you know, understand sort of how you bridge sort of that aspect of learning quantum with your fascination with quantum error correction, which is quantum error correction, to my mind is one of the most most complicated, most challenging aspects of quantum information theory in a lot of ways because it's the, you know, getting reliable results out of hardware that's operating at a macroscopic quantum level is so incredibly challenging. And the the maths behind that are like, unbelievably complicated.

Kevin Rowney 19:14
For starters, just in classical it's, it's no nonsense math to begin with. That's Basecamp you ascend from rare I mean,

Sebastian Hassinger 19:21
yeah, and it's, you know, the starting point is no cloning theorem. Right. So, exactly is

Unknown Speaker 19:25
like good luck.

Sebastian Hassinger 19:29
Yeah, but I mean, it does feel like like starting super early with with building a quantum intuition might be, you know, might be the way that we end up with with getting breakthrough approaches to quantum error error correction. Is that is that sort of in the back of your mind? Yeah.

James Wootten 19:46
I think like a lot of techniques for quantum quantum error correction are actually sort of sledgehammer to crack and that just to keep it simple, we can we, we can make it a lot simpler for us to, to conceive have, if we if we burn a whole load of physical qubits just to make one nice protected logical qubit. But error mitigation, as you as you go from the real sledgehammer to crack an end of quantum error correction towards error mitigation, and requires a lot more subtle effects and better understanding of the kind of dynamics that goes on in there. So

Kevin Rowney 20:26
that's a critical, crucial distinction. Maybe our audience is not well versed in the difference between quantum error correction and quantum error mitigation. Would would you feel comfortable doing a little dive into the distinguishing characteristics of those two subfields?

James Wootten 20:42
Yeah, so Well, I would say that, when we design algorithms, we assume that our qubits are perfect, right. And when we make qubits, they are not perfect. So there has to be something. At least in the past, there had to be a proof of principle that we could go from one to the other, we could use these imperfect qubits to make perfect qubits. And so there was a whole theoretical structure framework of quantum error correction that was derived in order to firstly show that it could be done but also show how it can be done. But error mitigation is more where you're you're taking that imperfect qubits and still trying to use them in your algorithms. But you've got to kind of find more of maybe quick and dirty ways to manage the amount of noise it's in there or find ways to mitigate for it or extrapolate from the noisy result that you get the end, what the real result will have been. So it's not quite so theoretical and clean and proof of principle and mathematical. As quantum error correction, there's a lot of maths in it, as well, but it's more just trying to work with its work around the errors, whereas quantum error correction is just completely annihilate them. So you don't have to think and write

Kevin Rowney 22:05
really, really useful distinction. Thank you. And I've read I don't know if it's true, please comment is that is some people at least think that quantum error mitigation is probably a shorter term strategy, and that there's a lack of clarity about whether or not QAM will scale when we get to large amounts of quantum volume. And there's lots of theory on quantum error correction. There'll be relevant once we get to big scale, but we're not at that big scale yet because of engineering. So it seems like we're there's two separate eras of where the the best strategy will be relevant is that, does that match up with your understanding? Well,

James Wootten 22:43
as certainly as you go into the far future when we've got like billions of qubits, qubits, and we could ever need, and we can very just very much just use the techniques of quantum error correction rates are quite expensive in terms of resource requests. Yeah, right. Now, the resource requirements are so big, we can hardly really even make one error corrected qubits with all of the resources we have. So we were for years mitigation.

Kevin Rowney 23:11
They're very fact right, Jeremy, I think many people from the outside of the quantum computing movement who are just getting involved with it, they don't understand that that very basic fact about a huge limit on the space. I mean, stuff like Shor's algorithm is like a remote science fiction future until we get past that gigantic threshold. You know, there's a long way to go before we're ready for the more serious applications.

James Wootten 23:33
Yeah, sounds like between these two extremes. It's just I don't know whether it's going to be sort of discrete, or it's going to be continuous. So are we going to find progressively better error mitigation, which just leads us into quantum error correction? And there's not really going to be a distinction, we're gonna find a path between them? Or is it going to be we struggle with error mitigation for as long as we have to? And then we just, we're glad to be past that error, correct.

Sebastian Hassinger 24:07
Right, right. Yeah. I guess. I mean, I wonder, when I read about mitigation strategies, it feels like some of them are engineering techniques, right? It's better fab, it's better control systems. And it's tricks, you know, with, you know, calibration or, or, you know, readout techniques, and that sort of thing. So that feels like, you know, sort of the, the continuous evolution of the way we make qubits. But then there's the, the information theory side of that, where it's, you know, novel approaches to using the noisy qubits in in an unexpected way that as you say, May may sustain a continuing like an ongoing, novel approach to computation. is So I mean, I don't know if we talked about Even as yachtie A couple of weeks ago, and he and Savar case have just done this, what they call stereoscopic fingerprinting of the noise profile of of specific quantum systems. And I thought that was a very interesting sort of, you know, creative approach to trying to understand holistically what the the noise is on a given system and potentially provide some strategies for mitigation. Yeah.

Unknown Speaker 25:25
Good point. Yeah.

James Wootten 25:28
Yeah, noise is a really interesting subject, just finding out what's going on in these devices, it brings you to another area that hobbyists might be interested in, which is the ongoing science behind even trying to build them. But, uh, yeah, so with the error mitigation. Yeah, and I'm not sure whether you'd label things like in a circuit you asked for a control or not. And there's already been a whole lot of work and making that controlled, not the best possible job I could ever be. But then on top of that, you can you can have tricks like, well, I want to run this circuit with these control knots or whatever. But I'm actually going to run two versions of it. And they're going to be slightly different. And I'm going to use some techniques to extrapolate the real noise. So that's the real result. So that's more than the error mitigation. But even then, at some point, when you compile your circuit, someone else is going to be using that technique. So I suppose there's, yeah, there's where error mitigation start is also slightly, right. Ambiguous?

Sebastian Hassinger 26:33
Yes. Yeah.

Kevin Rowney 26:35
I mean, I'm very new to my studies on this topic. But I'm just speculating. Could it be that lots of the statistical tradecraft with respect to, you know analyzing data and attempting to interpret its meaning in the presence of acknowledged errors? And the signal is perhaps a foundational part of QAM? Is that accurate and true?

James Wootten 26:58
Yeah, I think that has a role to play, especially with readout error mitigation is one of the biggest sources of error in the quantum computers is when you ask your qubit, I use zero or one as a as a big source of noise. And but it's also exactly where it interfaces with the classical, where all of superpositions go away. So you can actually use quite classical techniques. At that point, in order to try and choose what the real noise would have been. When things are within the actual quantum thus, then I suppose it becomes a bit more ambiguous whether you can get established classical techniques to work in or certainly for mitigation.

Sebastian Hassinger 27:43
Right, right. What are the the sort of major themes that are the most interesting to you these days that, you know, you, you tweet on a daily basis like today on the archive and quantum error correction? Are there particular sort of trends that you see that are interesting at this point?

James Wootten 28:02
Yeah, so as well, probably about a third of them are about error mitigation rather than error correction. And maybe, maybe a sixth, just making up fractions here are very, very abstract yet, I think, maybe a decade or a decade and a half ago, everything was at the very abstract and people want to do the fancy maps, they come up with fancy new error correction stuff, just to for the love of the mathematics. So you'd be proposing things that needed four dimensional space in order to exist just to see what the limits were. But most of them in quantum error correction are now quite more of a practical level, they are looking towards the hardware, they might still be a little bit theoretical, because running large scale quantum error correction codes is still a little bit far away. But they're looking towards the hardware and have a more practical outlook, which I think is good.

Kevin Rowney 29:08
It seems like one crucial characteristic of the quantum error correction puzzle is this central theorem, which you know, it's not actually it doesn't seem to masse, it's the the threshold theorem. I hope you agree that that is a important part for people new to the space to sort of understand it. If you agree that's important. Would you attempt to explain it to us and the audience?

James Wootten 29:35
Yeah. So well, of thresholds within quantum error correction codes is something I can explain quite easily, I think. And that's the heart of the threshold theorem, really. So within the current primary creating code, you have many noisy qubits, and you're using them to build a structure which one noise this qubit can live and it kind of lives in I think coded not just in one single, noisy qubit, but in more in collective properties of a whole bunch of them. And then you keep measuring those noisy qubits to see traces of the noise that's happening. And so this gives you hints about what's going on. And it's basically a puzzle you can solve, trying to figure out where the noise was, and therefore how to mitigate it. And so what you look to do is make your code ever bigger, because the bigger you make it, the more kind of detailed information you get. And the less likely it is that the hints you're getting, are going to fool you in some way. Because then a very simple, very small instance, there, you might get a weird little bit of noise, which looks like it's something but it's actually something else. And, and that causes you to mitigate for the errors in the wrong way. Where when things are bigger, then you have more information to try and work out what's going on. But also, when it's bigger, you're using more noisy qubits, there's raw noise going on. Everything is essentially noisier. So there's a trade off there between how much information you're getting, how much detail you're getting, about the noise. But also, there's, there's more space for things to go wrong is, is more likely, there's something disastrous is going to happen. And so really, which of these trade off wins, depends on how strong your noise is, if your noise is very weak. If you're hardly ever, ever seeing noise events, then the fact that bigger gives you more information, and makes it harder for disastrous things to happen is, is going to protect you.

Kevin Rowney 31:50
But adding new qubits is a net gain rather than introducing Yeah, more noise,

James Wootten 31:56
it makes you more protected, yeah, less likely for something bad to go wrong. But if the noise is too high, then the fat just adding more stuff means you've got more entropy, there's more, you're going to have a lot more chance that something horrible is going to happen. And so as you scale up your code, you really find that if you're looking at the probability that something horrible is going to go wrong, it becomes a step function, where if where if your noise is over a threshold, then it's just something bad is going to happen. Just give up. And if it's below the threshold, then you're You're fine, you're gonna find a way to mitigate for that error,

Kevin Rowney 32:37
you can keep adding more qubits and get to better and better accuracy.

James Wootten 32:42
Yeah. So if you then take this because this here, I'm just really talking about one encoded qubit sitting there doing nothing other than feeling noise, of course, you also need to do gates, you need to take these encoded qubits and interacting with each other and do everything you need for an algorithm. So once you build all of that in as well, you get then the notion of a threshold generalized to the entirety of quantum computing, but still, it's as long as you're below a certain noise, then everything is going to be fine. And that was a important thing to prove back in the day, because it showed that there was a finite value of noise that we can cope with. And then the first proofs, that value was, like 10 to the minus 20, or something stupid, because they were just really trying to prove that it was finite. Right. Since then, we've we've got a lot higher. So So noise levels now are approaching what we believe the threshold to be. Right? Yeah, but that's kind of

Sebastian Hassinger 33:49
the freshmen who, who were who were the some of the the researchers involved in that initial theoretical proof that it was finite.

James Wootten 33:59
So yeah, those those names don't come to mind. I know. So the person who has it has the same name as another person who is famous. So it's a doubly famous name at the forefront of my head.

Kevin Rowney 34:17
Yeah. And so what is that? What is the current math result in terms of the threshold of of error occurs that you need to get to to to get over the

James Wootten 34:28
hump? Yeah. Well, that we're looking at sort of the percent level and depends on on code and decoders and

Unknown Speaker 34:36
delete one or two, one or 2%. Yeah, yeah. Yeah.

James Wootten 34:39
But it's the sort of percent level not the 10 to the minus something's right.

Kevin Rowney 34:43
Yeah, that seems unachievable. Yeah, by any remotely feasible current engineering. Yeah. That's really interesting. Well, I said that it does provide a little ray of hope around the the progress on on the engineering. If the theory says that you could get it down to 1% it says That's likely some of the people on the engineering side see a route towards getting to that that outcome?

James Wootten 35:06
Yeah. So we seem to be on route for it. And also, there have been experimental results already where people have built minimal examples of quantum error correcting codes and showing that they have a beneficial effect. So that's what happens when the noise is low enough. Right. Right.

Sebastian Hassinger 35:24
Yeah, I think Ken brown point that I think Ken brown actually led a group that that implemented a bacon shore on on a trapped ion system. I think it may have been sent Diaz research system, possibly if I'm not mistaken, but but can also in conversation with Ken, I remember him saying, every year, they say, it's not possible. And every year we get a little bit closer. So yeah, fun. He's been at it for awhile.

James Wootten 35:54
So IBM Quantum, there was an experiment, which I think went out last year. And also, more recently, the Val rush group, eth, to me, did a surface code, right, which, so usually, it's always some sort of code that's easy to fit on the devices people have like, right bacon shortcode, or the particular instance, that was used in IBM Quantum, but to see a 17, qubit, surface code, that kind of thing that asphere just loves to see, because it's kind of simple. That was very nice.

Sebastian Hassinger 36:30
Is there a and can you describe what you mean by surface code? Because I've heard that term thrown around. But I, I confess, I don't fully understand what that means.

James Wootten 36:40
Yeah, well, I've been doing research based on the surface code now for a decade and a half. So can I give you a short answer?

Sebastian Hassinger 36:52
Well, less than a decade and a half long, maybe.

Unknown Speaker 36:57
So we had the time. Yeah.

James Wootten 37:00
Yeah. Well, initial examples of quantum error correcting codes, didn't really think about the practicalities of how your qubits were able to talk to each other. But there are certain codes now, which are usually within a class called topological codes, where your qubits are imagined to be sitting on a two dimensional surface, and they only have to talk to their neighbors. And so they're a way to do quantum error correction within those constraints. And the surface code is a classic example of this, you typically think of your qubits sitting on a square lattice, and they're talking to their neighbors. And so half of your qubits are just sitting there talking to each of their neighbors and seeing if they've seen an error, and then measure them to see if there's an error around. Yeah, so that's the basic idea.

Sebastian Hassinger 37:56
So you're using entanglement to to reflect whether the the actual the data qubit is is has been affected by an error? Or if the has been affected by an error? The error would show up in one of these entangled Ancilla qubits. Is that right?

James Wootten 38:15
Yeah, well, how much is entanglement needed? It's gonna, so you can, you can think of a version of the surface code, which only looks for bit flip errors. And that requires no entanglement. Oh, I see. And there's a version in which it only looks for phase flip errors, and they need a lot of superposition. But it also doesn't need entanglement. But if you want to do both of those at the same time, you can, it's perfect, they're perfectly they they work perfectly well with each other. But as soon as you do that, then the only way that the system can support taking all of that information at the same time is to be in an entangled state.

Kevin Rowney 38:55
Right? And I guess for our audience, that's I think a crucial point is the the two aspects of sources of error within quantum computation, right, is the bit flip and phase error. But both have to be accounted for that's very distinct from classical computing. Are there any other types of errors that somehow quantum error correction has to struggle with? Or are those the two main pieces?

James Wootten 39:19
Yes and no? Well, okay, so there are a huge amount of this is another thing that I could probably go on too long about, I think I like to think of quantum computing, it's kind of a mixture of continuous and discrete computing. So the wave particle duality of quantum becomes continuous discrete duality. And so, we have continuous errors, which are a thing that is awful in in analog computers, but we have error correction kind of discretize is so so everything turns into A bit flip errors and phase for parents. Even if it's something very discrete and weird quantum evolution, it'll turn into fitflop errors and phase for parents, unless it's a loss error, which is the kind of the other type, and that is your qubit disappears. Right? That's

Unknown Speaker 40:17
your dooby, dooby decoherence. And that particular. So

James Wootten 40:22
it depends on what your qubit is. So one way is, so I'm kind of talking your qubit physically disappears is one way. So you've got photonic qubits, they might actually right fly off in the wrong direction. Kind of things that we have IBM Quantum there, we isolate the lowest two levels of an harmonic oscillator. But there are other levels. And if it decides it wants to go up into them, then it's not a qubit. Effectively. It's kind of gone.

Sebastian Hassinger 40:54
Yeah. Interesting.

Unknown Speaker 40:56
That's interesting. We're kind of what do you call that particular error?

James Wootten 41:00
Or the data leakage? air leakage? Thank

Unknown Speaker 41:02
you. Okay.

James Wootten 41:04
So leaks out of the space is supposed to be Wow.

Kevin Rowney 41:09
This is what I love this detail. And last, last, the most interesting topics we come across the speaker wants to go on for a long time. So don't don't hesitate. But I'm glad you're able to be concise. So thank you. I saw recently i You probably are watching this way more carefully than I am. I think it was a Nature paper, announcing some new qubit architecture, I think I think it was on just raw silicon chips, where they were able to do enough proof of concept around you're running that qubit in that architecture. But if the error threshold that people were concerned are around the threshold theorem is Am I remembering that correctly?

James Wootten 41:52
Yes. So there's been a lot of research recently in silicon spin qubits, and you want to show that you can do a controlled not, that is good enough to be below the threshold. So silicon spin qubits are quite far behind a lot of other approaches at the moment in that when you're demonstrating a two qubit gate, you've probably only got two qubits. So it's not quite the same as when you've got 127 qubits. And you're showing like, you've got two good two qubit gates on there. But they that is showing that you can get a good two qubit gates on two qubits is the stepping stone to more. Right. So actually, some of the work I do is looking into alternative qubit architectures. And so we're also looking at Silicon spin qubits ourselves. So yeah, that was, that was a big day in silicon spin qubits that

Sebastian Hassinger 42:50
are pretty cool. I guess that's that's where superconducting qubits were in 97, or 98, or something like that. 99, but two qubits.

James Wootten 43:02
Well, I transmit and I think is much more recent than that. Uh, yeah, that's true. That's true. This difference between transplants and silicon Spanky, which is huge. Right, the moment I transplants being a kind of superconducting qubit that is, but uh, yeah, maybe 20 years down the line. And I think

Sebastian Hassinger 43:21
that's what keeps the field really interesting, though. So much. Oh, it's right. Yeah.

James Wootten 43:28
Yeah. So for me, I don't mind where my qubits come from, as long as they present good, interesting problems from an error correction perspective. So that's why I'm happy to work on transplants and on silicones pinkcupid.

Sebastian Hassinger 43:41
Absolutely. Absolutely. Interesting. It's great. Excellent, James. Well, I recognize we're nearing the end of our time. Is there anything related to any of the topics we discussed that you want to touch on before we close out?

James Wootten 43:58
Now? No, I think I there's nothing sticking in my mind. I think I wish I said, blah, blah, blah.

Sebastian Hassinger 44:09
I wanted to give you an opportunity to interject one last blah, blah, blah, if you want it. Yeah. That's great. Well, I really do want to thank you. I think that the contributions you've made to the teaching materials, the learning materials, and just the the ideas around making quantum computing accessible to hobbyists and to, to, you know, newcomers, I think is all extremely valuable for the community overall. And it's, it's, it's It's great stuff and I've enjoyed a lot of the stuff you've done. So thank you very much.

Kevin Rowney 44:43
Yeah, we really appreciate it. Very fun conversation. great blend of topics right though. Yeah, the pedagogy and the gaming add this this deeper theory on on EC. So really, thank you so much for your time.

James Wootten 44:52
Yeah, thanks. It's good to talk to you

Sebastian Hassinger 45:37
All right, that was fun. I really like how James talks about almost building a hacker community or bringing the hacker community into the quantum technologies community because you know, as a sign of vitality, and sort of harkening back to the, the early days of Silicon Valley and the Internet and the.com, boom, I think that there's a lot of merit to bringing in those really are attracting this really creative, intelligent sort of problem solvers into a new space. And that, you know, feeds into his sort of interest in education and pedagogy, which I think are extremely important topics, it doesn't really make sense to build a new technology if you don't actually have people who know how to use it. So education is key key. So

Kevin Rowney 46:23
agree, I think I also just another theme that just keeps coming up in in these interviews, right, that I think is noteworthy is all this stuff around you new ways to think about types of analysis of error and accuracy. It's such as T one time to T two times and bit flip errors and phase errors. I mean, that's, that's, that's just the base camp. There's so much more out there. And so I mean, new types of of errors and frameworks of analysis that you can use it to control for, I think, is a rich topic of huge researchers out there, but also a great way to see with clarity, what's really going on with the current crop of vendors. So that was pretty cool to hear all that dialogue about so called quantum computing leakage errors. Why are we so grateful? Thank you, James. If you're listening to this recording for your time, but we have so you, you the audience enjoyed this much as we did this guided tour of these two facets of QC Reacher's research and education. And thank you for your time. Thank you. Okay, that's it for this episode of The New quantum era, a podcast by Sebastian Hassinger and Kevin rolling. Our cool theme music was composed and played by Omar Costa Homido. production work is done by our wonderful team over at pot phi. If you're at all like us and enjoy this rich, deep and interesting topic, please subscribe to our podcast on whichever platform you may stream from. And even consider if you'd like what you've heard today, reviewing us on iTunes and or mentioning us on your preferred social media platforms. We're just trying to get the word out on this fascinating topic and would really appreciate your help spreading the word and building community. Thank you so much for your time.

Creators and Guests

Sebastian HassingeršŸŒ»
Host
Sebastian HassingeršŸŒ»
Business development #QuantumComputing @AWScloud Opinions mine, he/him.
Quantum Error Correction with James Wootton
Broadcast by