Charles S. Sydnor Distinguished Professor of Computer Science
She’s an analyst. He’s an algebrist. She loves displays of unbridled creativity; he has journaled every day for thirty years. Between the two of them, Duke power couple Ingrid Daubechies and Robert Calderbank have changed the way society shares and processes information.
James B. Duke Distinguished Professor of Mathematics and Electrical and Computer Engineering
Art–Conservation and restoration, Inverse problems, Wavelets (Mathematics)
Miranda Volborth: This is Rate of Change, a podcast from Duke University dedicated to the ingenious ways that engineers are solving society’s toughest problems. I’m Miranda Volborth and I’m here with Ken Kingery, in conversation with Duke’s preeminent power couple.
Ingrid Daubechies: I’m Ingrid Daubechies.
Robert Calderbank: I’m Robert Calderbank.
Miranda: If you work in math or physics or engineering, chances are you’re already familiar with these names. if you’re not, suffice it to say that their work has influenced the way you communicate at a fundamental level.
More on that in a bit.
For the time being we’ll just say that these two are pretty famous, each in their own right. Ingrid, in fact, had a multipage profile in the New York Times magazine last year, titled “The Godmother of the Digital Image,” which featured a photo of the mathematician riding a bike across the parched Black Rock desert, at Burning Man.
We wanted to know more.
Ingrid: It’s most infamous for, uh, drugs, and wild parties and all that, and sex orgies, and I’m interested in none of those <laugh>. What I was really interested in was the completely unbridled creativity, and that’s really why I think it’s a blast. I really love going there. I love seeing all the creative stuff, the art, all the weird ways in which people put things together and do things and enjoy what other people do. And I really, really, really like it.
Ken Kingery: How many times would you say you’ve gone?
Ingrid: I have gone four times so far and I intend to go again when I can fit it into my schedule.
Robert: I think you have to have a certain comfort level with discomfort.
Miranda: I think that’s exactly right. And I think that I do not have that comfort level.
Robert: <laugh> I don’t have that comfort level either. I love to see the pictures. I’m enchanted by the crazy vehicles. But all that grit <laugh>, I’m just… I don’t really like crowds. So being in the middle of 50,000 people in the desert is not my thing.
Ingrid: The first time I said, “I want to go, do you want to go?” And Robert says, “Absolutely.” He says, “In the middle of the desert.” He says, “In August, in the hottest time of year, in a bunch of a wild crowd of crazies. Absolutely.” He says, “My dream vacation.” I mean…so, that decided that I’m going by myself.
I mean, I am completely energized and exhilarated when I’m there by everything I see. I mean, completely crazy things that you wouldn’t expect! So, one day I got up very early. I, who am not a morning person, because I wanted to see the artwork at dawn. And so, I’m bicycling through the mostly sleeping camp— because most people have these parties go to these late parties, and so are sleeping in the morning— and I go past the camp that was going to have a concert later, and they were doing a rehearsal and they were singing music by Offenbach. But they were yodeling it! I mean, which lends itself very well to Offenbach, but I’d never heard that before. And I had never thought of that, that I would be bicycling there and hearing people yodel to Offenbach at six in the morning. I mean, it was absolutely crazy. These things happen all the time. So, I really, really, really like it.
Miranda: And you feel like that kind of recharges you, carries you through the year?
Ingrid: Through things? Oh, absolutely. Absolutely. I mean, uh, just actually witnessing human creativity always recharges me. I mean, so, so seeing unexpected and, and creative things is what I get a lot of my energy from. Yes.
Ken: I’m just curious, do you have any sort of, uh, artistic, uh, inclination or is that all Ingrid?
Robert: I think it’s mostly all Ingrid… I mean, I think the, the most artistic thing about me is handwriting…
Ingrid: You… Oh, you have your journal with you?
Robert: I’ll show you.
Ingrid: Robert has this collection. I mean, for as long as I can remember he has a journal that he keeps, and he has one for every year. And it has a summary of his technical ideas, but also of things that happened in our personal lives. And so, I met a historian who said, “This is going to be a treasure trove for some historian of science in the future.” So yes, he has 30 years’ worth of these, one for each year.
And, and it’s almost calligraphed. I mean, this is just Robert’s normal handwriting when he’s thoughtfully writing. His letters to me when we were “courting,” as they say, were in that handwriting. It’s his standard handwriting when he’s not scribbling. And I’ve wondered whether we shouldn’t make it a font, the “Calderbank” font.
Miranda: I love that.
Miranda: Did you keep the letters?
Ingrid: Oh yes, of course I have all of those.
Robert: I mean, one of the reasons that I do that is, that sometimes when you do research or you do administration, you spend an entire day, and at the end of the day, it’s hard to figure out what you actually did, and taking 20 minutes at the end of the day to try and write down what it was that you accomplished is useful.
Miranda: Robert has been at the front of just about every breaking wave in communications technology since the nineties. He worked on voiceband modem tech at Bell labs, as the internet was coming into being; that technology was incorporated in more than a billion devices. If you ever heard a modem chirp and squeak, you can thank Robert for that memory. More recently, he was part of the team at AT&T labs that developed the group theoretic framework for quantum error correction. providing the foundation for fault tolerant quantum computation. So, the fact that he faces the same existential question as everyone else: “what the heck did I accomplish today, actually?” is kind of striking.
Ken: Do you find that that exercises a different muscle than thinking about math or physics? Or is it all kind of the same idea?
Robert: No, very much so.
Ingrid: It’s very much the same.
Robert: No, no, no. It’s different, but it’s, it’s also… I mean, often the idea comes… I mean, you’re thinking about this and the idea comes from over there. Actually, when you’re thinking about technical things, you are— you’re often trying to make that happen. There’s not a recipe for that happening, but when it does happen, it’s magical.
Ingrid: We don’t know where ideas come from, but it’s my experience also that they come from, um, freewheeling a little bit. I mean, having the particulars of the problem in mind, but not trying to focus too hard on them. I mean, trying to mentally step back a little and, and leave a little bit of freedom so that things can jiggle and then suddenly something… I mean, most of the jigglings are not productive. I mean, if we knew a way of making them very productive, we would bottle it and sell it. But it’s a mental habit, and so that’s why sometimes I’ve had some of my best ideas come— I mean, well, a standard thing is in the shower, but actually I had some of them come when I was nursing a baby. Because you have to be kind of relaxed to make that happen, to make it go naturally, to not be all crisped up and so on. And that’s exactly the kind of, of situation when, when, when things can go, when thoughts can, can flit from one place, from one flower to another.
Miranda: What was that idea?
Ingrid: Oh, it was about—actually, it was about constructing, combining several structures that I had worked with, in a very elegant way that then actually became something that became part of the JPEG 2000 standard. So it was, it was to do with special wavelet filters and how to structure them very nicely.
Ken: Well, thanks for all the JPEGs. <laugh>
Ingrid: Well, it’s JPEG 2000. Standard, JPEG consumer product JPEG is an older standard which is still based on a discrete cosine transform, which is Fourier transforms. The JPEG 2000, which is the one that’s used for more advanced internet applications— for transmission of high-quality sports broadcasts or digital cinema, that’s the one that’s based on waves. It’s really JPEG 2000.
Miranda: Ingrid is best known for her work in image compression, specifically with wavelets, which are brief oscillations useful to signal processing. A wavelet bearing her name is used in the JPEG 2000 format that she just mentioned. She’s a MacArthur Fellow, a recipient of the Wolfe Prize in Mathematics and a Baroness—a title granted by King Albert II of Belgium in 2012 for her achievements in physics.
Ken: Is there anything in that that applies to how an engineer should think, or how a mathematician should think and are, are those two different things?
Ingrid: Well, I’ll let Robert give his answer, but for me, it is all very related in that you try to put things together that maybe were not necessarily put together. And I think my background being in physics, and then working on mathematical applications and engineering, gave me a different perspective—a different point of view, which I have found very useful. And so, I encourage my students also to take as many classes, different classes, as they have an appetite and time for, because I tell them that knowing things in different areas is what’s really helpful to then find new solutions.
But, what do you think?
Robert: Yeah, I… look, I make my living at the border between electrical engineering and mathematics. What does that mean? It means that I hang out with engineers. Like, at Duke I hang out with people who have quantum computers, and I sort of make the calculation that if I listen to them long enough and carefully enough, then I’ll discover something that I can do that they can’t because I have this mathematical background. That’s what I tell my students, too: that I want them to become bilingual. So, I want them to be able to talk to the experimentalists and I want them to be able to talk to their own community.
Something that people don’t often realize is that science research, engineering research, it’s really community activity. And one of the things I think we both try and teach our students to do is to create communities, be productive parts of communities. It’s one of the advantages of working in different areas and talking to different people—because you have a technical area that you work in, but when you think about problems a little bit outside of that area, then they illuminate your area in a different way. And that’s worth something. I mean, this program that I started in the summer, the Data+ program…
Miranda: Data+ is a program of the Rhodes Information Initiative at Duke, which Robert directs. It’s a summer program for students who choose small project teams to work with for ten weeks over the summer, solving problems with data. Gathering data, analyzing it, visualizing it. The projects are so interesting and diverse. Here are a just a few projects for summer 2023: Measuring urban heat islands and their causes in Durham; Mitigating the Risks of Emerging Technologies; Quantifying the carbon footprint of Duke dining; landscape analysis of whole genome sequencing and newborn screening. There is something for everyone. And the experience is so meaningful that the program has become hugely popular at Duke and has been replicated in programs like Code+ and Climate +.
Robert: …and one of the things that we discover is that questions from the humanities tend to break the algorithms we have. And so that’s a lot of fun. We had climate projects, we have projects with the city of Durham, uh, which we’ve done for many years. And the magical thing about it is that it’s a community of students that are discovering data science at the same time. First and second year students sign up for the experience and they come out of the experience with this network of people who know stuff, who they can ask questions of. And I think that it gives them a whole support network for their time at Duke. And I think they might actually be more ambitious about the classes they sign up to take, because they have this support network in the background.
Ingrid: But so, but it also— and I think very importantly— teaches them that on a project it’s useful to have input from people who have a different background than your own, and that people who work on different projects but use similar tools, may be useful resources. I mean, and that is something. Both these things are very important and useful to Duke to have as an attitude for the rest of their lives.
Robert: Yeah. I mean, I think that, you know, we… Of course, students come to Duke, we teach ’em stuff. So, they’ll learn an algorithm and then we’ll test how well they know the algorithm. We’ll give them a data set and it’s a clean data set. And the answers are perfectly unambiguous. When they work in the real world with a real data set, it is messy. It is incomplete. It takes you four weeks to get it in the form where you could even ask it a question, and you have to do this. So, in a sense, it’s an experience where you integrate all of the things you learn to do something of use for a client.
I mean, I think that there, so there’s two things that I think are educationally transformative. One is this idea of a project with a client, uh, and the other is being part of a team to work on the project. And I think that the combination of the two is really transformative.
Ken: I’m curious if you guys bounce problems off each other. If you kind of work collaboratively at all, or if you just try to keep your professional lives separate.
Robert/Ingrid: So, this is a really… yes, yes. And no. <laugh> So, it’s yes and no.
Ingrid: Actually, sometimes it’s really interesting because most of the time we need very few words to communicate. Like, you know, if we both like a rug or a painting, then we resonate.
Robert: We just resonate. But with, uh, technical projects, I think we found it very frustrating because we needed a lot more words…
Ingrid: Because our skill sets are very different. Robert is basically an algebraist and I’m an analyst. And the way I describe that, which is caricature in both senses, is that analysts are really… the problems that we look at are, the metaphor is, you have an overstuffed sock. And you put in something else, and then you want to see where it starts bulging elsewhere. I mean, it’s very mixed and fluid. In algebra you have these beautiful structures. I mean, it’s like building something with…
Robert: Like a crystal.
Ingrid: It’s like a crystal. Something with a zoom tool or one of these toys that really fit together beautifully, and they fit in only one way. You build these beautiful things, and then you might take a hammer to it, and it might— something might break off, but the whole thing might be essentially the same. Or, it completely breaks. But there’s not a question of, “I hammer here and then something else starts bulging there.” It’s all or nothing. I mean, it’s a very different mindset. It’s much more on these beautiful crystalline structures, which is analysis is not really… Now, both of these are complete caricatures, but the result is that because our skill sets in mathematics are so different. What Robert does, I find really hard. And he claims to find that what I do is hard, when it isn’t.
So, <laugh> so the result is that… one of us says something that’s kind of obvious, and the other one looks puzzled and the first one becomes irritated and says, “What do you not understand about that?” So, we do very little of that because we get too— we get too frustrated. I mean, it works in both directions, actually—the irritation.
Robert: <laugh> Yes. Although we did write— I mean, we wrote one paper together. And it’s actually an interesting story, because we were at Princeton and there was a graduate student who had taken Ingrid’s wavelet class and had taken my course on data communications. And he came to us both and he said, “Okay, does now I want a research project that involves you both.” And the research project turned out to be lossless compression and it had an influence on the JPEG standard. And we did get quite frustrated in the writing of the paper <laugh>, but at the end it was actually quite a successful paper. It’s got like 2000 citations or something like that.
Ingrid: But we haven’t done it since.
Robert: <laugh> We haven’t done it since.
Ingrid: And, I mean, but I’ve been thinking that I might try to really get into some of what you’re doing now with quantum computing, again.
Robert: Mm-hmm <affirmative>. Yeah. So, for example, we have completely different skill sets in quantum stuff. So, I mean, Ingrid understands what hydrogen atoms really are. And for me, quantum computing starts when things become operators—
Ingrid: And well, and Clifford algebras…
Robert: And Clifford algebras. Yes.
Ingrid: Yes. And I say, “oh my God, Clifford algebras…” <laugh>
Miranda: I’m going to be honest here, I had no idea what these two are talking at this point, but they understood each other very clearly, and they didn’t need very many words to achieve it.
They just resonate.
It goes to show how valuable spending time with people outside of your own field of expertise can be.
Thanks for listening. Subscribe for more updates from Duke Engineering, and if you learned something from this podcast, please share it with others.