What is quantum error correction, and why is it so vital for future quantum computers? Take a listen to Episode 2 of insideQuantum to find out!
This week we’re featuring Dr Philippe Faist, a postdoctoral researcher at Freie Universität Berlin. Dr Faist did his undergraduate degree and PhD at ETH Zurich, followed by a postdoctoral position at Caltech, and is currently in his second postdoctoral position here in Berlin.
For more information and a full audio transcript, see our website insidequantum.org.
🟢 Steven Thomson (00:06): Hi there and welcome to insideQuantum, the podcast telling the human stories behind the latest developments in quantum technologies. I’m Dr. Steven Thomson and as usual, I’ll be your host for this episode today. We’re continuing our launch week series of talking to quantum technology researchers here at Freie Universität Berlin. And in particular, we’ll be digging into the concept of quantum error correction. It’s a pleasure to be joined today by Dr. Philippe Faist, a post-doctoral researcher in quantum information and a co-founder of the Error Correction Zoo. Hi Philippe, and thanks so much for joining us today!
🟣 Philippe Faist (00:36): Hi Steven. Thanks for having me here.
🟢 Steven Thomson (00:39): So before we get too far into error correction and just exactly why it needs a zoo, let’s first talk a bit about your journey to this point. Can you tell us, first of all, what first got you interested in quantum physics?
🟣 Philippe Faist (00:51): I guess physics has always been running somehow in my family. My grandfather was already a physicist, my dad’s a physicist, so it felt natural to go study physics. During my master’s studies — so in the later years of my studies — I was torn between two different things that I was really passionate about. One thing I was really interested in was the fundamental laws of physics. So I took a lot of theory courses like particle physics or the standard model. And on the other hand, I was really interested in tinkering in the lab and working out things in the lab and playing out with all of these great devices and machinery. And I really loved that. So I figured, oh, I’ll probably go into experimental research. At some point during my master’s, I took a lecture by Renato Renner, actually with whom I would then do my PhD.
🟣 Philippe Faist (01:41): And he was giving a lecture on quantum information theory and I felt, wow, this is incredible. You can really touch on deep questions in physics. Like what is the second law of thermodynamics from an information perspective and so on. And so it, it really drew on my appeal to the fundamental laws of physics on the theoretical side. And I really, you know, I was really interested by that. That really kind of opened my eyes to that field of quantum information. And I asked him to stay for a Masters project. And from then on, it worked out, you know, for me to stay for a PhD there. That’s how it started.
🟢 Steven Thomson (02:15): So you were almost an experimental physicist then?
🟣 Philippe Faist (02:18): Oh, I wouldn’t, I wouldn’t go that far. I did a semester project, I guess we called it back then. I did three weeks in the lab twice during my studies. I loved it, but it was still baby stuff. You know, I was probably soldiering little wires on a printed circuit board or reading out stuff from an oscilloscope or things like that. So, you know, I, I wouldn’t claim to be an experimental physicist in any way.
🟢 Steven Thomson (02:41): Do you think it’s good to have a little bit of a grounding in some experimental techniques before moving into theory though, to have actually, you know, touched some experiments and have some idea about what it means to do experiments, even if it is a little bit simplified?
🟣 Philippe Faist (02:56): Yeah. You know, I, I think it’s, it is actually very illuminating to have gone through that experience simply because as a theorist, you can then listen to an experimental talk and instead of seeing all of those plots and thinking “Oh, okay, that’s cool”, you can actually start appreciating more all of the work that went behind actually calibrating your devices, you know, getting all of that equipment set up in order to be able to actually take those measurements and you get a slightly better perspective about everything that’s behind everything you’re being told.
🟢 Steven Thomson (03:28): It’s the unglamorous stuff, isn’t it? The stuff that never gets published in Science or Nature but probably takes 90% of the time. So at what point did you decide that you wanted a career in quantum physics, then? You found it interesting to study and obviously wanted to study it further, when did that turn into thinking “Okay. I really want to work in this area”.
🟣 Philippe Faist (03:48): You know, I think it happened gradually and in a certain, in some kind of natural fashion, I was studying physics. I got interested in quantum information theory, as I told you. And I asked, you know, I, I got interested in doing a Masters project on quantum information theory. And then that worked out pretty well. So I asked my advisor if I could stay for a PhD. And he offered me a PhD position. So in a certain sense it went very naturally from there on.
🟢 Steven Thomson (04:17): So it’s a case of always one more question, always one more thing that you want to do. And then before, you know, it suddenly you’re, you’re in…
🟣 Philippe Faist (04:24): You’re in, exactly — knee deep.
🟢 Steven Thomson (04:28): So if you hadn’t gone into physics, then if you hadn’t made all these decisions and taken the path that you did, what do you think you might have done instead?
🟣 Philippe Faist (04:37): Good question. I’m not sure I thought that deeply about other options. I mean, I was tempted at some point…so I played the violin quite a bit especially during my high school years. And I was very tempted to continue as a professional musician to some extent, but I also…I was very strongly drawn to physics, so I’m not sure how seriously I had actually considered that option, but it’s an option that I entertained for a bit as an alternative to physics.
🟢 Steven Thomson (05:08): I think a lot of physicists seem to also be musicians in their spare time.
🟣 Philippe Faist (05:12): Yeah. I don’t know how that happens.
🟢 Steven Thomson (05:14): Musicians and hikers.
🟣 Philippe Faist (05:16): Hikers and climbers. Yeah. Boulder, you know, bouldering, a lot of people are into that.
🟢 Steven Thomson (05:19): Okay. So we’ve talked a little bit about your journey and how you got to this point. Can you tell us a bit about what it is that you actually work on? So what’s the field that you work in. What’s the big picture goal of this field? What’s the challenge that your field and your work is trying to address?
🟣 Philippe Faist (05:37): I’ve done quite a bit of different things along, you know, the past few years. My PhD thesis was mostly concentrated on quantum thermodynamics and how to understand, for instance, the second law of thermodynamics from using modern tools of quantum information theory. Then during my postdocs, I actually went into different areas. For instance, quantum error correction. Maybe the latest thing I’m really interested in is a bit of a natural continuation of these ideas around quantum error correction, it’s quantum complexity. And there, the idea is really to try to understand the physics of many body systems kind of beyond the point where they look like they have thermalized. If you take some complicated system with many parts, it will, if you start it in some very simple state after a while, it will just look like it’s thermalized. You know, like your cup of coffee, it’ll just cool down to room temperature after a while. But we know in quantum mechanics, if the system is closed, then the system continues to evolve and it retains some memory of the initial state. We’re trying to understand and develop all of the important tools to get a handle on how this system behaves after this point where it looks like it has completely thermalized and reached room temperature, but actually there might be some memory of the initial state that you could try to recover if you have sufficient computational power.
🟢 Steven Thomson (06:51): So it’s like — to come back to this cup of coffee — you leave a cup of coffee in a room, you go away, you come back, it’s at room temperature, and you have essentially no way of knowing what was the original temperature of that cup of coffee. But you’re saying if the cup of coffee was in some kind of isolated quantum system, it will cool down. But if you somehow understand this process of…this evolution, that’s cooling down of it, you could reverse engineer this evolution process and figure out how did this cup of coffee start? What was the original temperature? What was the original states? And it retains some kind of memory of that?
🟣 Philippe Faist (07:24): That’s right. That’s right. And in certain sense, these are all questions somehow related to “what is the second law of thermodynamics”? How can you have some kind of second law of thermodynamics in a system that’s isolated, right? In which you’d expect that entropy has to stay constant from the laws of quantum information, but where we do observe some form of entropy increase. And somehow, really the idea is to relate this apparent entropy increase to…what is accessible to someone who’s looking at the cup of coffee. You know, the, the person looking at the cup of coffee will not be able to take a microscope and track every individual trajectory of every atom and molecule in this cup of coffee, but rather will have some very coarse measurements about that. And somehow we’re trying to push these ideas that have, you know, already been out there in statistical physics for a long time. We’re trying to push these ideas to regimes of higher and higher complexity to say, well, what if I have some near term quantum computer that I could use to kind of reverse engineer the evolution for longer times than I would need just to establish some values of simple things that I could measure, like energy or temperature.
🟢 Steven Thomson (08:33): So it sounds like there’s two parts to this. One is kind of fundamental, how do you describe physical processes in terms of information theory? And then the other here — you mentioned quantum computers — is a little bit more…I hesitate to say ‘practical’ when I’m talking about quantum computers, but let’s, let’s say practical for lack of a better word. How can you understand what quantum computers are doing through these information theory processes? And then I guess, how can we, how can we get something sensible, get something useful out of these noisy devices that we have the best approximation, I guess, to a quantum computer that we currently have. Does that seem like a fair statement?
🟣 Philippe Faist (09:11): You’re spot on. One of the ways that we’re thinking about this type of problem is really we’re thinking, well, what could a quantum algorithm do if I’m only allowed to place a certain number of elements in this algorithm, we call these unit elements, gates, you know, there are just units of processing. There are some simple calculation on just two units of your computer and you combine them normally to build a big algorithm. And if you’re restricted in how many gates you can apply, maybe for instance, because you don’t want your algorithm to run for eternally long times, then you’re limited in computational power. There are some things you will not be able to do. So really we’d like to understand what you can do, basically the set of all possible operations you can imagine doing if you’re restricted in these number of units, computation elements.
🟢 Steven Thomson (09:55): So, okay. We mentioned there “gates”. Maybe if we just explain what they are in a kind of classical way. So in a classical computer, you have these bits, that are zeros and ones. A gate, then, is some sort of operation that you can apply to these zeros and ones to, I dunno, to add them together or something like this, some kind of simple elementary operation, is that correct? Okay. And then in quantum computers, you do essentially the same thing. You have these same building blocks, these same gates, but you’re acting on quantum bits rather than classical bits, which have this much richer, more complicated behavior. And that’s what leads to the interesting features, but also causes the problems, I guess?
🟣 Philippe Faist (10:32): Yeah, that’s right. I mean, they have, they’re…what’s funny is that they’re extremely different to manipulate. You would use very different techniques to understand what you can do with a classical algorithm versus what you can do with a quantum algorithm. And that has to do with, you know, the structure of quantum mechanics and what you’re allowed to do in quantum mechanics. That’s definitely one of very, very interesting aspects of this type of research.
🟢 Steven Thomson (10:53): Okay. I see. What would you say then is the biggest challenge in your field at the moment — let’s say kind of near term, five to 10 years?
🟣 Philippe Faist (11:03): There are a lot of challenges. I think the way I see it, there’s, there’s kind of a big vacuum. We understand pretty well the physics that can happen, if I can say at low complexity scales — you know, where you look at states that you could prepare potentially with fairly few of these gates — but somehow it seems to me like we have very little understanding of what we can actually do with more than just that level of gates. For instance, if you have a chaotic system, a system that’s very kind of mixes up information in everywhere, those are systems that will be able to explore regimes of complexity that are higher. Then you could have a system — and you’re a very prominent expert in this — for instance, where if your information stays localized in different areas, those are states you’d expect to still be a fairly low complexity. And I mean, I think really understanding the physics of what’s going on past those timescales is something that excites me right now.
🟢 Steven Thomson (11:54): And is this — the physics of these highly complex systems — is this something that we have other ways to probe at all? Or is this a complex problem for any method? Be it information theory or any other theoretical or experimental technique that we have?
🟣 Philippe Faist (12:11): I guess this is still very much new ideas that are around it. I should mention these ideas came from quantum gravity and, you know, Lenny Susskind for instance is very interested in understanding black holes and worm holes and quantum gravity. And there, he really started introducing this idea of complexity as being something physical that we should look at. So I guess you could say there is another approach to understanding this question of say complexity growth in systems that comes from quantum gravity. We can use our intuition, and the idea is that you could get about maybe what theories of quantum gravity would say about worm holes in black holes to say something about complexity. You can start playing around with near term quantum devices. Maybe one thing I didn’t mention is quantum machine learning, or, you know, near term quantum devices, they’re naturally setups where you would want to do these simple algorithms, you know, these algorithms with a limited number of gates. So these are all settings where you could try to explore and better understand what you can do and how to better get a handle on these states of kind of that have some level of complexity that go beyond the simple states that we usually encounter in physics.
🟢 Steven Thomson (13:19): So this is a field that has connections to almost all areas of physics. I guess this is not some kind of very academic technical question. This is something that has applications to technology, to machine learning, quantum computers from your past research on this topic. Then, is there anything that particularly stands out to you as something that you are really proud of? Not necessarily your highest impact or most cited work, but like for you personally, is there a piece of work you’ve done that means a lot to you?
🟣 Philippe Faist (13:48): Well, you know, yeah, I did a few things. I’m not sure to which extent they’re, you know, impactful or I can be proud of them in the sense you’re suggesting. I think maybe one piece of work where I felt, you know, this, I felt proud at least on the moment. So while I was at Caltech, we had this project on studying quantum error correction and studying how quantum error correction is affected by symmetries that you can have at the level of the encoding. And in that context, there were a few things already that were known. But we were really trying to get some form of general trade off in how basically…I should probably explain some background here. We were interested in error correcting codes. In an error correcting code, you want to encode some logical information onto some physical hardware, and then you want to do computation with your encoded states because at that level you want, you want to carry out some computation and there is a famous theorem in quantum computing that says that basically you can’t do that.
🟣 Philippe Faist (14:54): There are…you know, if you wanted to do that by applying just gates on each physical qubits on each physical unit of hardware then you will not be able to do any interesting computation at the, at the logical level, at the level of the information you’re protecting. So really the question becomes, how can you find ways around that? And we studied a particular way of going around that, which might be well, how, what if we loosen a little bit, our requirements on the quantum error correcting code and required the codes not to be perfect, like not to protect the information perfectly, but only to recover it with a good enough probability. And if that’s the case, well, the theorem that, that says that you’re not allowed to do computation by applying local gates — you’re not allowed to do interesting computation, I should say.
🟣 Philippe Faist (15:43): Then that theorem breaks down and we didn’t really know how to adapt that theorem to this more general setting where you allow the quantum error correcting code to have some imperfection. And I guess that my contribution there was to come up with a certain point of view or a certain idea that that is actually very common in information theory. But that, you know, turned out to be kind of — in this context of quantum error correction — an interesting way to look at the problem. And that’s always to look at, you know, in quantum error correction, we’re interested usually in thinking of encodings, we’re interested in various schemes of what types of errors you can correct, what types of encoding and decoding you can have.
🟣 Philippe Faist (16:30): But another interesting thing you can look at is what goes to the environment. That means like what information leaks out and is lost. And something very specific about quantum information theory is that you can say a lot about a system by just studying what leaks out of that system, what information leaks out of that system. So that’s what we did. That’s kind of the approach that I took to study that problem, and that helped us get this result about covariant codes which I mentioned, this kind of trade off between error correction and continuous symmetries. And I should mention, of course it was a collaboration with a lot of people. So that was kind of my 50 cents in the collaboration.
🟢 Steven Thomson (17:06): Nice. Okay. So there are two, two things in there that I’d like to dig down on a little bit more deeply. So one is error correction. We’ll come to, I think, more on error correction in a moment. The other one is there, you mentioned this principle of how exactly you do computing and the difference between physical and logical qubits. So then this is like saying that if you have if you want to store some information or do some computation, you need more physical qubits to store the information, then you might think it’s not a case of just having, I dunno, one qubit for every piece of information. You need extra qubits to allow the system to actually do something useful. Is that correct?
🟣 Philippe Faist (17:48): Yeah. I mean, think of it even classically, if we’re separated by, you know, a hundred meters and I’m screaming or you’re screaming something to me, right. I might say “What? I didn’t hear you!”, and you might have to scream that thing again, two or three times before I’m able to actually get what you’re saying. And that’s a form of error correction. You know, we’re trying to communicate over some noisy environment. So a lot of the information that you are trying to convey to me actually gets lost instead of arriving, you know, into my ears and that you can remedy by using something we call an error correcting code. So typically if you have a communication line, an optical fiber, you will use some type of error correcting code to redundantly, encode your information in a way that you can actually still recover the interesting information at the other end, even if some of the information carriers that you sent through get lost. That allows you to reliably convey information for one party to another.
🟢 Steven Thomson (18:42): So perhaps the simplest, you know, very, very toy model of error correction then is, is the one you just use about repeating a message several times if pieces of the message are lost in each communication, but you repeat the message enough times, eventually you can reconstruct the entire message.
🟣 Philippe Faist (18:56): That’s right. That’s a very simple form of classical error correction. Now, you, you mentioned of course, quantum error correction before, too, there’s again a very interesting theorem in quantum information theory that tells you that you can’t do that. You can’t just repeat a message many times that’s called a “no cloning theorem” and it tells you basically, you can’t just repeat things. So you have to come up with more clever techniques for encoding your information on these, on your physical hardware, such that if you lose some of the components of your hardware, you can still recover that logical information. And it was not at all obvious historically that it would be possible. So it was really a big breakthrough back in the nineties when people showed that it was actually possible to do that.
🟢 Steven Thomson (19:35): And the no cloning theorem — that tells you, if I paraphrase, that you can’t copy a quantum state without changing it? And that’s quite different to classical information where you can have as many copies of a bit of data as you like, and they’re all essentially identical to each other, but in a quantum system, the act of measuring, the act of looking at that data changes it. So if you want to make a copy or lots of copies, then they’ll all, I guess, be different to each other. And then that’s a disaster. That’s not what you want.
🟣 Philippe Faist (20:03): That’s right. That’s right.
🟢 Steven Thomson (20:05): I see. Okay. And we’re talking about error correction in very sort of general terms here. We’re talking about error, correcting codes and different ways of doing error correction. Why are there different ways of doing error correction? Why…you know, you see an error, ideally, we just want a way to fix that error straight away. What is it about this problem that gives rise to this whole family of different potential ways to fix the errors? And why are there so many, why are they so different from each other?
🟣 Philippe Faist (20:31): Great question. Let’s go back to your example of the repetition code, where I just repeat my information for a classical error correction. Well, in that case, if I repeat each letter or each word three times, for instance, then that will mean that I will have to triple the size of the information that I’m sending to you. If I want to convey a sentence of 10 words, I will suddenly have to send you 30 words to be able to convey that information. Well, it turns out that there are codes that are much more efficient so that maybe I would only have to send, and I don’t know, 15 words instead of 10 to redundantly encode the information in a way that protects it to a similar level of protection to what I would get with the repetition code. So that’s one example of a better code than the repetition code in terms of the encoding size.
🟣 Philippe Faist (21:18): There are many parameters on an error correcting code that we might be interested about — many features, if you like. So first is how many physical carriers do I need to encode a given amount of logical information. Then there might be how robustly protected the encoded state is, how many carriers can I lose while still being able to recover the information? Then there’s another aspect which is, can I encode to the state and can I decode it easily? You might have a very good error correcting code, but if it’s really hard to encode or decode, that might not be a very interesting thing to do. So you have all of these parameters and they trade off. You know, they’re sometimes you have codes that are very good at encoding lots of information, but they might not be very robust to losing a lot of physical carriers.
🟣 Philippe Faist (22:06): So one of the reasons you have this variety of error correcting codes is that they kind of all specialize in different features and some codes are good at one thing, some other codes are good at something else. Maybe one code is simply very easy to describe, and that might be even useful just for pedagogical purposes, if you’re teaching error correction. So, you know, there is a huge variety of error correcting codes that trade these different parameters off. And then there’s another reason why you have so many codes, which is it’s still a very active field of research. People still realize that there are new things to be discovered, new, better codes to discover. And to give you an example, even classical error correction has new results coming out. I mean, 10 years ago or so there was a discovery of the polar codes, and that’s now the technology that’s being used, if I understand correctly, in 5G networks. So you do have a lot of progress going on also at the classical level, even though the field of classical error correction has been there for very long. And of course, quantum error correction is a much younger field and it also has a lot of new research coming out every day. It’s a very active field of research with a lot of new results, all of these new results, they generate new codes and you kind of make your field of error correction bigger.
🟢 Steven Thomson (23:17): So all these different codes, then — you mentioned that they have slightly different advantages or disadvantages. Would it be correct to say then that some of these might be useful for, I dunno, let’s say you’re building a quantum computer in some particular hardware setup. Well, maybe the type of hardware you’re using — be it trapped ions or superconducting qubits — the type of hardware you’re using says, “Okay, right. This is an easy — relatively easy — manipulation to make. Therefore we want an error correction code that involves this sort of process.” And then if you’re, I dunno, sending information, you mentioned optical fibers, then perhaps optical fibers. You would use a code that involves different types of processes, maybe more suited to optical fibers. So then is this the reason why there are so many different types of error correcting codes? Because there are so many different physical platforms and different situations where errors can creep into quantum systems, because I guess quantum systems are super vulnerable to errors in a way that classical systems are not so trying to protect from all these different possible errors and all these different physical situations is the reason that there are so many error correcting codes.
🟣 Philippe Faist (24:17): Absolutely. And you mentioned the hardware, that’s something I forgot to mention. You know, these codes sometimes are really suited for one type of hardware and not another type of hardware. And that’s definitely something that adds to this richness of the landscape of error correcting codes.
🟢 Steven Thomson (24:34): Talking then of the rich landscape of error correcting codes. This might be a good opportunity to talk about the Error Correction Zoo. Can you tell us then a bit about what is the Error Correction Zoo?
🟣 Philippe Faist (24:45): Of course. It’s a project to collect categorize and to organize all of the different error correcting codes that we’re aware of into one website currently. And that way, if you’re interested about a particular error correcting code, you can look up that error correcting code on the Error Correction Zoo, and get all of the basic information about that code and pointers to the official references, where to look up where these codes were introduced, how they’re defined, what they’ve been used for and so on. So it’s really one central repository for collecting these error correcting codes and trying to shed some light on how they are organized and how they’re related together.
🟢 Steven Thomson (25:24): Why did you feel like there was a need for a project like this to organize and categorize all the different codes that are available?
🟣 Philippe Faist (25:31): There are a lot of error correcting codes. You know, you mentioned that there is the…the landscape is really huge and I think it’s, it’s always good when you have a field of science that has generated a lot of results to at some point, start taking a step back and saying, okay, can we kind of get a bigger picture about this? Can we start organizing these results in a way that is useful either for someone who comes from the outside and says, okay, how should I get an idea of what’s going on in quantum error correction, for instance, or for even someone within the field of quantum error, correction, to kind of easily look up, how is this error correcting code again, related to this other one and so on. So it’s just a general valuable resource to kind of take a step back from the details of each individual paper and get some big picture of quantum error correction.
🟢 Steven Thomson (26:18): That definitely sounds like something worthwhile and something I think a lot of other fields would benefit from as well, taking this big picture approach. I think there are certain aspects of many body physics — the field that I work in — where yes, people can’t see the woods for the trees, I guess.
🟣 Philippe Faist (26:33): You should start a many-body zoo!
🟢 Steven Thomson (26:37): Maybe! But I can definitely see why it would be useful to have this overview of all the different options that are available. If someone were interested in error correcting codes and they were to go onto the website, then what would they find? They log into the website and and they they’re faced with hundreds — thousands? — of error correcting codes. How many codes are on the website? Do you know?
🟣 Philippe Faist (26:57): We’re at around a hundred to a hundred and fifty. I think it’s increasing, as we’re adding more and more codes.
🟢 Steven Thomson (27:04): Okay. So if someone goes onto the website and they’re faced with these hundreds of codes, what can they do with it then? How would they navigate this website? What kind of information can they get about them?
🟣 Philippe Faist (27:14): We’re trying to design the website in a way that’s not just a bunch of pages, but that kind of tries to show you the relations between these codes and how to get there. So, first thing you can search for something. If you’re interested in a code that has a particular feature, you can search for it. You can click on different links that, you know…there’s a very basic organization that got inspired from, you know, organization of life forms and biology with domains and kingdoms. So you can have the classical domain or the quantum domain, and each domain has its own kingdoms — the qubits, the qubit kingdom, the continuous variable kingdom and so on. So they’re kind of related to the different types of hardware you could build your code on. And that is something that you would see for instance, on the landing page. On the landing page, you would also have a random code presented to you. Just picked at random. That was inspired by Wikipedia’s random article feature.
🟢 Steven Thomson (28:05): Is there any way to, to sort of visualize this, I’m thinking of you might be familiar with these visualizer tools that show you all the papers on the arXiv and which ones cite which. Is there anything like this for error correction?
🟣 Philippe Faist (28:15): We do have something like that. You can, there’s a page. I guess we called that the code graph, it’s an automatically generated diagram with all of the error correction codes that are in the zoo and they have kind of lines that relate them or arrows that relate them depending on how they’re related. Some codes, we felt are special cases of other codes. So we decided they would be children of the other codes. So you’d have an arrow from the child to the parent in this graph. And you’d also have this structure of domain and kingdoms kind of outlined. That was part of this effort in giving you an overview of the different codes and how they’re related.
🟢 Steven Thomson (28:52): If anyone listening wanted to get involved in this, or if they go on the website and they see that there’s a particular code missing that they would like to add. Is it possible for them to get in touch with you and maybe contribute some of these features?
🟣 Philippe Faist (29:03): Absolutely. Actually all of our code base and all of the information about the error correcting codes, they’re hosted on GitHub and there’s a public repository for that. So if someone is interested, they can, of course send pull requests or create issues and, you know, contribute to the Error Correction Zoo in that way.
🟢 Steven Thomson (29:21): Have you had much feedback from the community in this way, much engagement with people contributing codes?
🟣 Philippe Faist (29:25): We have, yeah, absolutely. Even from this group here in Berlin, and that was really wonderful to see. It seems like people are getting interested in the Error Correction Zoo and that’s, you know, to a big credit…I do wanna mention here — I should have mentioned this earlier — I do wanna mention Victor Albert, who is really the main leading force in this Error Correction Zoo, it was really his initiative. I do wanna credit him for that. It was really his initiative to start this Error Correction Zoo. And he did an absolutely fantastic job in talking about it to everybody and promoting it. In January there was the big quantum information conference that was held in California. And there were a lot of people there and he had printed out stickers, you know, and handed out stickers to people about the Error Correction Zoo. So this is very much, you know, his, his initiative. And he did a great job at promoting it. And it seems like people are actually really responding to that.
🟢 Steven Thomson (30:17): So maybe touching on something you said there Victor works in the US at the moment. You’re here in Germany, but you’ve also worked in the US and also in Switzerland, if I remember your record right. So your career has taken you to quite a few different places, a few different countries. Were there any surprises when you moved between countries, any culture shocks, be it academic or cultural, anything that you weren’t expecting until you got there and took you by surprise.
🟣 Philippe Faist (30:44): You know, yeah, absolutely. You know, different places have different cultures obviously. And that really permeates in the academic environment and in research. You know, in Europe, we’re very focused on tradition and we’re very conservative in terms of ideas. If you try something new and it fails, people might have the impression, “Oh, but you know, it’s, that was kind of risky. You should have known that it could have failed”. I felt that in the US, there’s a very different approach to that. I felt there’s a much bigger optimism for new things that haven’t been proven yet and it’s it’s more of the culture of “well, let’s try it out and see if it works and if it doesn’t well, okay, let’s move on and try something else”. And I, I think I’ve learned really much more to try to be optimistic about new things without necessarily those new things having already been established as something promising. I try to really make that exercise all the time and say, well, that’s something new. I see this, I dunno, this new idea. Well, okay. There are millions of reasons why it might fail and, you know, I guess the European approach would be, oh, but it has to prove itself still because your, all these reasons it might fail. The approach that I felt was much more prominent in, in the US, or at least in California. Again, this is only my impression. I don’t know if this is, if you ask anyone else that they’d have the same impression. But the impression that I had was the approach would be — and what I really tried to exercise myself — is to say, okay, what are the reasons it might succeed? You know, kind of take you the opposite way and say, okay, and if it does succeed, what is it promising for? What will it be able to do? Where is this thing likely to go? If it manages to overcome all of these challenges. And I think it drives your research in a different way. And I think it kind of makes you look forward much more than trying to cover your tracks all the time and looking backwards to, you know…it helps you kind of really adopt new ideas, I think.
🟢 Steven Thomson (32:33): Would you say then that having had the opportunity to travel lot and work in different countries has, has been beneficial to you and to your career?
🟣 Philippe Faist (32:39): Absolutely. It’s been essential for me. It’s been essential. I mean, I don’t want to suggest that everyone has to do it right, but I’ve learned so much going to, into my first postdoc there at Caltech that I feel really transformed in a certain way, if I can say that? It really opened my eyes to a lot of things that I feel I would not have been able to learn if I had stayed in Europe.
🟢 Steven Thomson (33:00): Touching then on, I guess the difficulty, I suppose, of moving between countries, which is something that a lot of us do in academia…I think you’re one of the few group members here who actually has a family. How do you juggle this? How do you juggle having a family, having kids, and also having a, I guess, quite a high pressure career where you have to move around countries a lot and uproot your life every few years? That sounds like it must be a really difficult balance to strike.
🟣 Philippe Faist (33:27): It. It can be a bit of a challenge sometimes. You know, in academia, you’re building yourself as a researcher, as much as you’re producing something. So it’s, you’re not just…your work is not just to do something, it’s really to build yourself, your identity as a researcher. And so that kind of naturally blurs the line between your personal and your professional life. And of course, that’s true for, for everybody in academia, not people, not only people who have families. Now having a family, of course, definitely forces you to reset your priorities to some extent. And I think it did a lot of good for me because it, it forces…it really reminds me that there’s something bigger that I’m building here than just some research output or career. And I have something I can really look forward to coming back to when I get back at home after a long workday.
🟣 Philippe Faist (34:15): And, you know, I think there’s something really nice to that. So of course now academia might force you to move around a bit between different geographical locations, places. That can be a bit of a challenge sometimes, but it’s also a bit of an adventure, you know, it’s like when, when with my wife it was questioned that we’d move to California. It was a bit of a question, well, “Hey, do you wanna spend three…do you wanna live three years in California?” And you know, when, when you view it that way, it can be, it can be pretty nice. So yeah, there are definitely definitely challenges, especially if you have kids. Now you have to start worrying about how’s the daycare system, or how is the, how is the school system or healthcare system, depending on where we go, because that might be much more relevant.
🟣 Philippe Faist (35:01): So yeah, there are definitely challenges. It takes a lot of mental energy. You have to reset your social life and everything at all of the different locations that you go to. And with kids, that’s all kind of…there’s kind of a force multiplier that makes everything a little more difficult, but I think it’s part of progression in life and you’re moving to the next stage. And I think there’s something nice about that. My wife has also reoriented her career recently. And so, yeah, that definitely requires us to find a balance sometimes between “Hey, can you watch, can you watch Zoe this evening? I need to go do this event.” And sometimes the other way around if I have to do something. So yeah, there’s definitely a balance to strike there.
🟢 Steven Thomson (35:42): One of the other sort of challenges, I guess, talking about family life and moving and physics is that physics in particular is a field that has been very male dominated for a very, very long time. And obviously I think there’s still a long way to go before we reach any kind of equality, but more and more discussions are starting to happen about this. You’ve worked in a couple of different countries. Have you seen the attitude to equality change either over the years or in the different countries that you’ve worked in?
🟣 Philippe Faist (36:12): I think overall our field, you know…of course there’s a big issue here. We want to make sure that our field is diverse, that really all of the population groups are well represented in our field. My impression is that our field is not the worst in this category, but that’s definitely not an excuse for, you know, just sitting down and not doing anything. One thought that I did have on that subject was that I often hear this discourse as being “Well, how do we make our environment more favorable to women or what can we do to be able to better include women in our environment?” What strikes me is that there could be a nuance. Maybe this is just my impression, but there could be a nuance in that discourse that there is something about women or about other population groups like minorities, that, kind of we have to accommodate for, like we should accommodate for.
🟣 Philippe Faist (36:58): And I think really what’s going on is that there is some internal dynamics within the system that is already established that really privileges white men for staying within. And kind of, there’s some kind of internal dynamics that protects this. I shouldn’t say protects, that kind of…that’s a barrier for anyone else to penetrate. And I think one of the ways I would really approach the problem, you know, is really how do we break these dynamics? How do we…you know, it’s more of an introspection question, I think about ourselves. What are we doing wrong in setting up our system in a way that it’s so hard for either women or other minorities to get into the system and stay within? And that could be, you know, when you’re hiring someone and you’re trying to judge how good of a scientist they’re going to be. Well, maybe because if you’re interviewing a man, then it might be easier to connect just because of some common background or something, or some common cultural or, or I dunno, some personality or whatever…You might have some kind of implicit bias at that level. I, I would imagine that trying to crack down on those types of self preserving dynamics is really one of the most important things to address nowadays.
🟢 Steven Thomson (38:07): A few last questions to wrap up, then. We’ve talked a lot about the different areas that you’ve worked in, the different topics that you’ve worked on and the future promise for these fields. What advice would you give someone wanting to get into these fields? Someone who listens to you talking and thinks, “Wow, this stuff sounds really exciting. Where do I start? How do I, how do I learn more about these? How do I get into working in these fields myself?”
🟣 Philippe Faist (38:30): You know, what makes it tricky is that every career path is extremely different, you know, in academia or in research, especially in our field, you can do so many different things. Nowadays, you can go do an internship at a quantum company…that wasn’t really an option when I did my PhD. So, you know, if you’re a PhD student and you want to get into research and quantum information, it’s a lot about, you know, getting to know people, getting to know…networking, going to conferences. I think that’s a, I think we’re a field that has still a lot of new ideas and concepts that are to be explored. And even the, the technical tools, the foundation of, of our field are still evolving to a good extent. It’s still a very young field. So I think there is definitely room for new ideas and for new approaches. And don’t be shy of trying to think big and try to put your own mark and network with people, go to conferences.
🟢 Steven Thomson (39:19): I think probably the networking and speaking to people aspect is…that’s certainly a big one. And I would say personally, that’s something I did not do very well early in my career. I don’t know if I do it well now, but certainly I didn’t really click, I think, to how much research is community driven and done by people, done by people who talk to each other across time zones, countries… And this, I think creates a lot more opportunities for you than if you were just sitting in an office toiling away by yourself, posting papers on the archive. I think, yeah, absolutely talking to people is, the way forward.
🟣 Philippe Faist (39:53): You know, at the beginning of my PhD, I was in a lucky position where I could go to conferences because my advisor had lots of funding. And at some point my — well, then girlfriend, who then became wife — she was like, do you really need to go to all of these conferences? And I was like, you know, that conference I went to, that’s when I spoke to that person that unblocked me on that, that problem, this other conference I went to that’s when I learned about this other problem that I’m now working on. So, you know, I was really, I could really, at the beginning of my PhD, I could really kind of associate each conference with one significant breakthrough in my own development.
🟢 Steven Thomson (40:29): As a final question, then if you could go back in time and give yourself one piece of advice, what would it be?
🟣 Philippe Faist (40:37): I think I would go back to my PhD years and, well, I don’t know how…if I would’ve been ready to accept that advice at that time, but really I had, I guess I had some of that attitude of, I was interested in my things and I, of course I saw that there were other interesting things going on and maybe there were and I, of course I didn’t judge people for being interested in those other things, but somehow I felt, “ah, that’s not interesting to me. I, that’s not something I want to look into deeply”. And I, I think I really failed to appreciate a lot of developments and a lot of interesting things going on at that time, which maybe if I had, you know, been a little more interested, I would’ve learned more about quantum error correction, quantum complexity already back then. And it’s not something that I would have to basically learn from scratch later on.
🟢 Steven Thomson (41:20): So being open to new perspectives, I guess.
🟣 Philippe Faist (41:22): Absolutely. Yeah.
🟢 Steven Thomson (41:24): So if our audience wants to learn more either about you or about the Error Correction Zoo, where can they find you on the internet and on social media?
🟣 Philippe Faist (41:33): I am on Twitter. I guess that’s a way you can reach out to me. I have a email address that’s I guess, on my website. Yeah, I mean, there are standard channels. You can reach out to me.
🟢 Steven Thomson (41:43): Okay, perfect. Well we will leave links to your Twitter profile, your email and your websites on our own website, which is insidequantum.org, and anywhere else that we leave a transcript of this episode. Thank you so much then Dr. Philippe Faist for joining us today to discuss error correction and quantum technologies. Thank you.
🟣 Philippe Faist (42:02): Thank you, Steven. That was wonderful.
🟢 Steven Thomson (42:04): Thank you also to the Unitary Fund for supporting this podcast. If you’ve enjoyed today’s episode, please consider liking, sharing and subscribing wherever you like to listen to your podcast. It really helps us to get our guest’s stories out to as wide an audience as possible. I hope you’ll join us again for our next episode. And until then, this has been insideQuantum, I’ve been Dr. Steven Thomson and thank you very much for listening. Goodbye.