fbpx

Danica Kragic: Robotics and AI

by Music Tech Fest | MTF Podcast

In this MTF Podcast, Reuters journalist Jamillah Knowles interviews Danica Kragic, Professor of Robotics at the School of Computer Science and Communication at KTH Royal Institute of Technology. Danica brings a unique perspective to innovation process, and the relationship between artificial intelligence, creativity and the human experience.

She’s an internationally renowned researcher in robotics who also happens to be a seamstress and fashion designer. It’s hard to imagine someone who absolutely embodies the synthesis of cutting edge scientific research and human creative expression that we champion at Music Tech Fest more than Danica…

AI Transcription

SUMMARY KEYWORDS

robots, people, humans, robotics, understand, mtf, build, bit, seamstress, tech fest, ability, danica, sensors, code, environment, challenge, generate, design, kth, systems

SPEAKERS

Danica Kragic, Andrew Dubber, Jamillah Knowles

 

Andrew Dubber 

Hi, welcome along. I’m Dubber. I’m director of Music Tech Fest, and this is the MTF podcast. Now, MTF Stockholm in September took place at kth Royal Institute of Technology where they had some very brilliant professors working at the cutting edge of science and art, often both at the same time. Danica Kragic is a professor at the School of computer science and communication. She’s currently the lead investigator on a groundbreaking 10 year, 1 billion kronor AI research project funded by the minute and Alice Vandenberg Foundation, and bringing together academia and industry across a wide range of verticals. Her research covers robotics, computer vision, and machine learning. She also happens to be a fashion designer and seamstress in her spare time. Danica joined us as MTF’s woman in the lead of robotics and AI in the MTF Labs that took place in the decommissioned underground nuclear reactor chamber at kth. She also grabbed the moment to sit down with Reuters journalist and MTF host, Jamillah Knowles on the MTF interview stage, from MTF, Stockholm, Professor Danica Kragic.

 

Jamillah Knowles 

Alright, ladies and gentlemen, and everybody generally here at the Music Tech Fest, we’re going to start off with another interview. And it is my pleasure, again to be talking about design crossovers and technology with with a woman who I I have not met before, but I mean, extraordinarily impressed by I’m sure you will be too. Can we please have a round of applause, please, for Danica Kragic. So obviously, it’s part of what I do to stalk everybody thoroughly and about them. But we need to know what you do. So can you explain a little bit more about work. And also, I don’t like to call it hobbies, because it’s important work too. But tell us what you do.

 

Danica Kragic 

So to start with, my my office is like hundred metres from here. So I’m local. I’m from kth, some professor in computer science. And now for like 20 to 23 years, I’ve been doing research in the area of robotics. So I work with physical systems. And when I say physical, it’s a robot with arms. Usually on wheels, we don’t have Well, we have some robots on legs. But so what we do, really is we develop software that uses information from different types of sensors that robots are equipped with. And we generate what we call intelligent behaviour. So we want through bots to understand what they are. We want robots to interact with humans, we want robots to interact with the environment in terms of picking up objects, putting them down. And the grand idea how of me and many that are doing this type of research is that once we will have robots that can do all dangerous jobs that we humans do today and the to do with like robots to do in the future.

 

Jamillah Knowles 

So there’s a lot going on there and the research that you have and say going back not that long in robotics research. Robots quite often didn’t know where they were, which then leads to problems as to where they want to go. You don’t know where you are, you’re not sure where you’re going. So what are the advances that are making your work a bit easier is the other technologies that basically mean, you’re able to take more of a leap in this build.

 

Danica Kragic 

So I would start with something, because this is something that the humans take for granted a lot, our senses, and the ability to see the ability to feel the ability to feel, smell, hear, and so on. And we are so good and is to be integrate all these sensors all the time, in real time, and use this to interact with each other and interact with the environment. So building artificial sensing has for quite some time been a big challenge and still is. And also covering group with body with something that would look like or have the functionality of our skin. And cameras, yeah, they can generate images, but then you need lots of software in order to understand what’s in the image. So I would say it’s this interplay between building, artificial sensing and then also building algorithms that can understand what sensors actually measure and generate some form of a knowledge representation. But not not let’s say like this. Nothing of that. would be important if you did not have the ability to move about an environment. So you need to use sensors for the purpose in most cases for while moving around, you need to detect obstacles, you need to avoid them. So that’s what I mean by intelligent behaviour. And if you want to physically interact with the environment, not just push things around, but if you want to pick them up, having the ability to weld have some form of articulation, like in terms of arms hands. And I would say that that’s still what a big challenges in the area of robotics and I would say, industrial robotics, and she was here, she’s gonna say a little bit more about that, you can build a hand, but if you can’t cover it with skin, again, that there are still things trouble, otherwise, I wouldn’t have job. But challenges.

 

Jamillah Knowles 

And so, you know, that’s certainly a good way of looking at what’s happening inside a robot. But outside of robots, is us. And we are fleshy and soft. We are easily damaged, you can’t drop us from a great height,

 

Danica Kragic 

you can but

 

Jamillah Knowles 

we don’t we don’t bounce quite as well as a robot. But it also means that our interaction can be led by fear sometimes. So if you’ve got something large and metal coming towards you, what can we do to mitigate this? Because I mean, robots are very interesting, obviously, a huge part of our culture as well. But it can be a bit frightening. So what do you what do you advise when it comes to making robots a bit more acceptable?

 

Danica Kragic 

So it’s a very multidimensional question, because it has to do partially with our perception of technical systems. And to start with, if you have a technical system that solves effectively something for you, the brain just kind of done that has a tendency of accepting it. So even if you may be in the beginning, are scared of it, your interaction, spending time with using it understanding it makes also the perception of it easier. And I’m very much against building robots that look like humans, or our direct replicas of humans, we have that we have that in Japan. We also have now that UK, your Hanson robotics, and Sophia and so on. And why am I against that? Well, at least for some time, until we have robots having all the abilities of humans or even more or better abilities than humans, because if you are an untrained user, and by untrained, I mean somebody that is maybe not in the technical area, somebody that is potentially scared from the beginning, anything that looks like human you have human expectations on and they know how it is for us humans, when we meet the human, another human, that maybe doesn’t speak our language that comes from a different culture that doesn’t have the same values and things like that we become reserved. So it’s also between us, it’s not only between us and technical systems. So I would like to avoid fuzzing people, if I can say like that, especially now in the beginning, but the technology is still very, very young. So that’s one thing. The second thing is we can’t completely disregard human body, or be inspired by human body. And that has a little bit to do with environment around us. So everything that we see around us has been adapted to our bodies, the ability that we can sit makes the chairs of certain size of certain shape, the ability that we drink, or use mic or a pen. If we didn’t have any fingers, or we just had just one, the world around us would probably look completely different. So now you want to deploy a robot in the environment and wanted to do cleaning or picking up things, potentially ironing with an iron this whole, you know very much for me very much fun. And yeah, I can understand. But it’s the same. So you can’t really since you know, evolution has given us this body. We have throughout our history built things and objects to fit our body. Now he would like to be built something else that is as effective. And that’s not going to happen. Do you understand what I mean? So somewhere, we maybe need to learn how to address or develop the mechanical design, but potentially the face or it doesn’t need to look the same. And this is for those interested in mathematics. It’s a really interesting mathematical problem. And now you can see it as optimization problems. So how do you build another kinematic design or something Like that, that is as effective as you can get guarantee is as effective, but doesn’t necessarily look as a human.

 

Jamillah Knowles 

Interesting. Alright, so there’s I mean, there’s there’s been civics issues in different countries as well. So we’ve got like delivery robots now that are bringing your groceries to your home. That’s great. And questions like, where do we put a lane for a delivery robot in a city after we’ve got the bicycles, which I quite often look the wrong way and get run over in Sweden, sorry, cyclists, but also then the cars and then the pedestrians, and then you know, whomever else. So I certainly see what you mean trying to fit for robots into our world.

 

Danica Kragic 

And the problem is that you can’t just deploy technology and say, it’s gonna work from the beginning. And this is also what we expect, because we have seen lots of Hollywood movies, and in movies, everything is perfect. And everything worked from from the beginning, from the beginning, you can’t do the same now because we are just not used to this type of systems in the physical world. Of course, new technologies, virtual reality, augmented reality can be used to train our brains to get used to this things. But it’s still it’s a challenge for for our brain, because our brain is not used to maybe go around to the HoloLens whole day and, you know, generate kind of an overlay on the regular environment so that you actually see virtual robots coming toward you, when do you kind of like finding a new pattern for walking or something like that? So there are ways how you could potentially train everybody? Because you’re just not yet there? Okay, well,

 

Jamillah Knowles 

I want to turn a bit of a corner here in a chat with you and talk about discipline, and multi discipline. So obviously, I hugely admire your work in robotics. But as a seamstress, I also hugely admire your work in the field of creating clothing and fashion. Why do you do this, I mean, it’s not the easiest of skills. Hmm. But

 

Danica Kragic 

now my grandmother was a seamstress. And I spent lots of time with her when I was young. And from the age of four, five, she would give me pieces, you know, that that pieces of fabrics that she would have left from making something to somebody and I she allowed me to be creative from the beginning. And I was allowed to use a sewing machine and so on. But it felt that I invested time in something that I had, like a physical proof of, I read a lot and so on, I did some sports, I’m really bad at sports. But anyway, it just didn’t give me kind of like the same feeling of of me pointing and saying, Oh, I did that. And I suppose it, it was also a very good balance to programming and mathematics and everything that I educated myself in, because it’s, it’s a different type of creativity. And it’s also not serious, you know, I’m not kind of like a designer, professionally. So I can do things. And I, well, people can have opinions about how it looks specially my husband, he always has lots of opinions. But it’s very much about how you can do right, you can do wrong, but it’s only for you. So and I need that I need my free time activity to be something completely different from work.

 

Jamillah Knowles 

And so I mean, obviously, you’re in education, which is great. I mean, dedicating time to pass on your skills to other people of the generations is pretty huge sacrifice, because obviously, we could, I don’t know, glue you into big tech companies, and we’d never get to see your work again. But do you encourage your students then to consider different skills? I mean, as you say, working with with coders quite a long, you know, you’ve got long screen time, they’ve got fabulous minds. But is there a way to influence that in education to maybe give him some relief or ways to see the world a bit differently?

 

Danica Kragic 

No, absolutely. There are a couple of things. So so you know, as an academic, I’m very often asked, so your research kind of like how many startups a teacher did you do and so on. And then I say that, well, I feel that my responsibility is not to make startups myself but to make people or to build people that build startups. And I really strongly believe in that because you can’t, I mean, making a startup is really huge work. So you really need to keep your focus on make sure that that if you think that you’re recording kind of supervising people and so on in terms of having PhD students and so on mentoring, that that’s what you should do, and I really loved that. And I love that because I mean I have family, I have two kids, I love to tell them that there is also life, outside academia, which for some of them is especially in the beginning difficult to understand. But as Michela has said, you cannot build new things. If you do Don’t come in contact with people that challenge you. And yes, as an academic, you are challenged, but the rules used to challenge us like in terms of our citations, publications and so on. You know them. But then you change the area when you go to another area. And when you get the question that you understand, like, okay, you don’t know anything about my research, let me tell you, I mean, it challenges the way you describe, and you talk about your research. And that potentially also gives you a reflection that is a little bit different than it would have if you only talk to people that that do the same thing as you. So now, when we talk about future robots, when we talk about systems that are there, for humans and with humans, you need a good understanding of what not a human wants, but what many different types of humans won’t need or willing to accept. Because I think that that can directly affect your algorithms and how you think of designing the system for the future.

 

Jamillah Knowles 

And so just finally, to wrap things up, looking at different disciplines and putting them together, I mean, according to me, so it’s just my opinion, people can dismiss it, they want, but I think the clothing that you make is quite beautiful, I find it very appealing. When I talk to people in terms of code, it’s about being functional, and maybe elegant. Maybe we can do elegance in fashion and the way we’re not the same way that we do in code. Do you find ways to combine these? Or do you see more beauty in the code or robotics that you’re doing today? Or do you think that the separation is a good idea?

 

Danica Kragic 

I think that this is a this is also an important aspect, because today, more and more people like Michela said, and she learned or taught herself to programme out of other reasons that maybe I did. And I think that in order then to develop something that somebody else can use, you actually cannot think only about function, you also need to think about making beautiful code. And they also communicate that in a good way. And you can throw a pedal to literature, somebody needs to be able to understand what goes beyond just the technical because I think that that’s going to be very, very important when we start to work on human robot, real human robot interaction, although the field exists, and it has existed for 20 years now. But still really trying to understand this small differences and variations that will really make the difference. And this is exactly what you see in fashion and includes design. I mean, it can be very simple. But there is a detail that actually makes it so that everybody wants it. And it’s very recognisable to a certain designer. And I think that this is what we don’t have today programmers Don’t think about what is that makes my coat. Beautiful. So, but I think that it’s going to be needed in the future.

 

Jamillah Knowles 

I look forward to the future of the Gucci code. Thank you very much that it’s been wonderful talking to you.

 

Andrew Dubber 

Amazing stuff. Hey, thanks so much for listening to the MTF podcast, hope you’re enjoying and if so why not go back and find another one to listen to. There are some incredible people who’ve joined us at the festival. And we’re going to keep on adding more. Don’t forget to spread the word. Have a great week and talk soon. Cheers.