Chatbot Episode 1: Making Boston Dynamics’ Robots Dance
Evan Ackerman: I’m Evan Ackerman, and welcome to ChatBot, a robotics podcast from IEEE Spectrum. On this episode of ChatBot, we’ll be speaking with Monica Thomas and Amy LaViers about robots and dance. Monica Thomas is a dancer and choreographer. Monica has labored with Boston Dynamics to choreograph a few of their robotic movies through which Atlas, Spot, and even Handle dance to songs like Do You Love Me? The, “Do You Love Me?” Video has been viewed 37 million times. And in the event you haven’t seen it but, it’s fairly wonderful to see how these robots can transfer. Amy LaViers is the director of the Robotics, Automation, and Dance Lab, or RAD lab, which she based in 2015 as a professor in Mechanical Science and Engineering on the College of Illinois, Urbana-Champaign. The RAD Lab is a collective for artwork making, commercialization, schooling, outreach, and analysis on the intersection of dance and robotics. And Amy’s work explores the inventive relationships between machines and people, as expressed by motion. So Monica, are you able to simply inform me– I believe folks within the robotics discipline might not know who you might be or why you’re on the podcast at this level, so are you able to simply describe the way you initially obtained concerned in Boston Dynamics?
Monica Thomas: Yeah. So I obtained concerned actually casually. I do know individuals who work at Boston Dynamics and Marc Raibert, their founder and head. They’d been engaged on Spot, and so they added the arm to Spot. And Marc was sort of like, “I sort of assume this might dance.” They usually have been like, “Do you assume this might dance?” And I used to be like, “It may positively dance. That positively may do plenty of dancing.” And so we simply began attempting to determine, can it transfer in a means that looks like dance to folks watching it? And the very first thing we made was Uptown Spot. And it was actually simply determining strikes that the robotic does sort of already naturally. And that’s once they began creating, I believe, Choreographer, their software. However when it comes to my pondering, it was simply I used to be watching what the robotic did as its regular patterns, like going up, happening, strolling this place, totally different steps, totally different gates, what’s fascinating to me, what appears lovely to me, what appears humorous to me, after which imagining what else we may very well be doing, contemplating the angles of the joints. After which it simply grew from there. And so as soon as that one was out, Marc was like, “What about the remainder of the robots? May they dance? Perhaps we may do a dance with the entire robots.” And I used to be like, “We may positively do a dance with the entire robots. Any form can dance.” In order that’s once we began engaged on what became Do You Love Me? I didn’t actually understand what an enormous deal it was till it got here out and it went viral. And I used to be like, “Oh—” are we allowed to swear, or—?
Ackerman: Oh, yeah. Yeah.
Thomas: Yeah. So I was like, “[bleep bleep, bleeeep] is this?” I didn’t know how to deal with it. I didn’t know how to think about it. As a performer, the largest audience I performed for in a day was like 700 people, which is a big audience as a live performer. So when you’re hitting millions, it’s just like it doesn’t even make sense anymore, and yeah. So that was pretty mind-boggling. And then also because of kind of how it was introduced and because there is a whole world of choreo-robotics, which I was not really aware of because I was just doing my thing. Then I realized there’s all of this work that’s been happening that I couldn’t reference, didn’t know about, and conversations that were really important in the field that I also was unaware of and then suddenly was a part of. So I think doing work that has more viewership is really—it was a trip and a half—is a trip and a half. I’m still learning about it. Does that answer your question?
Ackerman: Yeah. Definitely.
Thomas: It’s a long-winded answer, but.
Ackerman: And Amy, so you have been working in these two disciplines for a long time, in the disciplines of robotics and in dance. So what made you decide to combine these two things, and why is that important?
Amy LaViers: Yeah. Well, both things, I guess in some way, have always been present in my life. I’ve danced since I was three, probably, and my dad and all of his brothers and my grandfathers were engineers. So in some sense, they were always there. And it was really– I could tell you the date. I sometimes forget what it was, but it was a Thursday, and I was taking classes and dancing and controlling of mechanical systems, and I was realizing this over. I mean, I don’t think I’m combining them. I feel like they already kind of have this intersection that just exists. And I realized– or I stumbled into that intersection myself, and I found lots of people working in it. And I was– oh, my interests in both these fields kind of reinforce one another in a way that’s really exciting and interesting. I also happened to be an almost graduating– I was in last class of my junior year of college, so I was thinking, “What am I going to do with myself?” Right? So it was very happenstance in that way. And again, I mean, I just felt like— it was like I walked into a room where all of a sudden, a lot of things made sense to me, and a lot of interests of mine were both present.
Ackerman: And can you summarize, I guess, the importance here? Because I feel like— I’m sure this is something you’ve run into, is that it’s easy for engineers or roboticists just to be— I mean, honestly, a little bit dismissive of this idea that it’s important for robots to have this expressivity. So why is it important?
LaViers: That is a great question that if I could summarize what my life is like, it’s me on a computer going like this, trying to figure out the words to answer that succinctly. But one way I might ask it, earlier when we were talking, you mentioned this idea of functional behavior versus expressive behavior, which comes up a lot when we start thinking in this space. And I think one thing that happens– and my training and background in Laban Movement Analysis actually emphasizes this duality between perform and expression versus the both/or. It’s sort of just like the mind-body cut up, the concept that these items are one built-in unit. Operate and expression are an built-in unit. And one thing that’s useful is basically expressive. One thing that’s expressive is basically useful.
Ackerman: It positively solutions the query. And it appears like Monica is resonating with you somewhat bit, so I’m simply going to get out of the way in which right here. Amy, do you need to simply begin this dialog with Monica?
LaViers: Positive. Positive. Monica has already answered, actually, my first query, so I’m already having to shuffle somewhat bit. However I’m going to rephrase. My first query was, can robots dance? And I really like how emphatically and superbly you answered that with, “Any form can dance.” I believe that’s so lovely. That was an excellent reply, and I believe it brings up— you possibly can debate, is that this dance, or is that this not? However there’s additionally a means to take a look at any motion by the lens of dance, and that features manufacturing unit robots that no person ever sees.
Thomas: It’s thrilling. I imply, it’s a very nice technique to stroll by the world, so I really advocate it for everybody, identical to taking a time and seeing the motion round you as dance. I don’t know if it’s permitting it to be intentional or simply to be particular, significant, one thing.
LaViers: That’s a very huge problem, significantly for an autonomous system. And for any transferring system, I believe that’s laborious, synthetic or not. I imply it’s laborious for me. My household’s coming into city this weekend. I’m like, “How do I act in order that they know I really like them?” Proper? That’s dramaticized model of actual life, proper, is, how do I be welcoming to my friends? And that’ll be, how do I transfer?
Thomas: What you’re saying is a reminder of, one of many issues that I actually get pleasure from watching robots transfer is that I’m allowed to mission as a lot as I need to on them with out taking away one thing from them. If you mission an excessive amount of on folks, you lose the individual, and that’s not likely honest. However if you’re projecting on objects, issues which are objects however that we personify— or not even personify, that we anthropomorphize or no matter, it’s only a projection of us. But it surely’s acceptable. So good for it to be acceptable, a spot the place you get to do this.
LaViers: Effectively, okay. Then can I ask my fourth query though it’s not my flip? As a result of that’s simply too excellent to what it’s, which is simply, what did you study your self working with these robots?
Thomas: Effectively, I realized how a lot I really like visually watching motion. I’ve all the time watched, however I don’t assume it was as clear to me how a lot I like motion. The work that I made was actually about context. It was about what’s taking place in society, what’s taking place in me as an individual. However I by no means obtained into that college of dance that actually spends time simply actually being attentive to motion or letting motion develop or discover, exploring motion. That wasn’t what I used to be doing. And with robots, I used to be like, “Oh, however yeah, I get it higher now. I see it extra now.” A lot in life proper now, for me, isn’t contained, and it doesn’t have solutions. And translating motion throughout species from my physique to a robotic, that does have solutions. It has a number of solutions. It’s not like there’s a sure and a no, however you possibly can reply a query. And it’s so good to reply questions generally. I sat with this factor, and right here’s one thing I really feel like is a suitable answer. Wow. That’s a rarity in life. So I really like that about working with robots. I imply, additionally, they’re cool, I believe. And additionally it is— they’re simply cool. I imply, that’s true too. It’s additionally fascinating. I suppose the very last thing that I actually liked—and I didn’t have a lot alternative to do that or as a lot as you’d anticipate due to COVID—is being in house with robots. It’s actually fascinating, identical to being in house with something that’s totally different than your norm is notable. Being in house with an animal that you just’re not used to being with is notable. And there’s simply one thing actually cool about being with one thing very totally different. And for me, robots are very totally different and never acclimatized.
Ackerman: Okay. Monica, you need to ask a query or two?
Thomas: Yeah. I do. The order of my questions is ruined additionally. I used to be serious about the RAD Lab, and I used to be questioning if there are guiding ideas that you just really feel are actually vital in that interdisciplinary work that you just’re doing, and in addition any classes perhaps from the opposite aspect which are price sharing.
LaViers: The same old means I describe it and describe my work extra broadly is, I believe there are plenty of roboticists that rent dancers, and so they make robots and people dancers assist them. And there are plenty of dancers that they rent engineers, and people engineers construct one thing for them that they use within their work. And what I’m serious about, within the little litmus take a look at or problem I paint for myself and my collaborators is we need to be proper in between these two issues, proper, the place we’re making one thing. Initially, we’re treating one another as friends, as technical friends, as inventive friends, as— if the robotic strikes on stage, I imply, that’s choreography. If the choreographer asks for the robotic to maneuver in a sure means, that’s robotics. That’s the inflection level we need to be at. And so meaning, for instance, when it comes to crediting the work, we attempt to credit score the inventive contributions. And never identical to, “Oh, properly, you probably did 10 p.c of the inventive contributions.” We actually attempt to deal with one another as co-artistic collaborators and co-technical builders. And so artists are on our papers, and engineers are in our applications, to place it in that means. And likewise, that adjustments the questions we need to ask. We need to make one thing that pushes robotics only a inch additional, a millimeter additional. And we need to do one thing that pushes dance simply an inch additional, a millimeter additional. We’d find it irresistible if folks would ask us, “Is that this dance?” We get, “Is that this robotics?” Quite a bit. In order that makes me really feel like we have to be doing one thing fascinating in robotics.
And from time to time, I believe we do one thing fascinating for dance too, and definitely, lots of my collaborators do. And that inflection level, that’s simply the place I believe is fascinating. And I believe that’s the place— that’s the room I stumbled into, is the place we’re asking these questions versus simply creating a robotic and hiring somebody to assist us try this. I imply, it may be laborious in that setting that folks really feel like their experience is being given to the opposite aspect. After which, the place am I an skilled? And we’ve heard editors at publication venues say, “Effectively, this dancer can’t be a co-author,” and we’ve had venues the place we’re engaged on this system and other people say, “Effectively, no, this engineer isn’t a performer,” however I’m like, “However he’s queuing the robotic, and if he messes up, then all of us mess up.” I imply, that’s vulnerability too. So now we have these conversations which are actually sensitive and somewhat delicate and somewhat— and so how do you create that house the place folks do you’re feeling secure and comfy and valued and attributed for his or her work and that they will make a observe document and do that once more in one other mission, in one other context and— so, I don’t know, if I’ve realized something, I imply, I’ve realized that you just simply have to essentially speak about attribution on a regular basis. I deliver it up each time, after which I deliver it up earlier than we even take into consideration writing a paper. After which I deliver it up once we make the draft. And very first thing I put within the draft is everyone’s identify within the order it’s going to seem, with the affiliations and with the—subscripts on that don’t get added on the final minute. And when the editor of a really well-known robotics venue says, “This individual can’t be a co-author,” that individual doesn’t get taken off as a co-author; that individual is a co-author, and we work out one other technique to make it work. And so I believe that’s studying, or that’s only a wrestle anyway.
Ackerman: Monica, I’m curious if if you noticed the Boston Dynamics movies go viral, did you’re feeling like there was far more of a concentrate on the robots and the mechanical capabilities than there was on the choreography and the dance? And if that’s the case, how did that make you’re feeling?
Thomas: Yeah. So sure. Proper. When dances I’ve made have been reviewed, which I’ve all the time actually appreciated, it has been in regards to the dance. It’s been in regards to the choreography. And truly, sort of going means again to what we have been speaking a couple of couple issues in the past, plenty of the evaluations that you just get round this are about folks, their reactions, proper? As a result of, once more, we will mission a lot onto robots. So I realized rather a lot about folks, how folks take into consideration robots. There’s plenty of actually overt themes, after which there’s particular person nuance. However yeah, it wasn’t actually in regards to the dance, and it was in the course of the pandemic too. So there’s actually excessive isolation. I had no thought how individuals who cared about dance thought of it for a very long time. After which each on occasion, I get one individual right here or one individual there say one thing. So it’s a completely bizarre expertise. Sure.
The way in which that I took details about the dance was sort of being attentive to the affective expertise, the emotional expertise that folks had watching this. The dance was— nothing in that dance was— we use the buildings of the traditions of dance in it for intentional motive. I selected that as a result of I wasn’t attempting to alarm folks or present folks ways in which robots transfer that absolutely hit some previous a part of our mind that makes us completely panicked. That wasn’t my curiosity or the objective of that work. And actually, in some unspecified time in the future, it’d be actually fascinating to discover what the robots can simply do versus what I, as a human, really feel comfy seeing them do. However the emotional response that folks obtained instructed me a narrative about what the dance was doing in a backward– additionally, what the music’s doing as a result of—let’s be actual—that music does— proper? We stacked the deck.
LaViers: Yeah. And now that brings— I really feel like that serves up two of my questions, and I’d allow you to choose which one perhaps we go to. I imply, considered one of my questions, I wrote down a few of my favourite moments from the choreography that I assumed we may talk about. One other query—and perhaps we will do each of those in serie—is somewhat bit about— I’ll blush even simply saying it, and I’m so glad that the folks can’t see the blushing. But additionally, there’s been a lot nodding, and I’m noticing that that received’t be within the audio recording. We’re nodding alongside to one another a lot. However the different aspect—and you’ll simply nod in a means that offers me your—the opposite query that comes up for that’s, yeah, what’s the financial piece of this, and the place are the ability dynamics inside this? And the way do you’re feeling about how that sits now as that video continues to only make its rounds on the web and set up worth for Boston Dynamics?
Thomas: I might love to start out with the primary query. And the second is tremendous vital, and perhaps one other day for that one.
Ackerman: Okay. That’s honest. That’s honest.
LaViers: Yep. I like that. I like that. So the primary query, so my favourite moments of the piece that you choreographed to Do You Love Me? For the Boston Dynamics robots, the swinging arms at first, the place you don’t totally know the place that is going. It appears so informal and so, dare I say it, pure, though it’s fully synthetic, proper? And the proximal rotation of the legs, I really feel prefer it’s a genius means of getting round no backbone. However you actually make use of issues that appear to be hip joints or shoulder joints as a means of, to me, accessing a great wriggle or a great juicy second, after which the Spot house maintain, I name it, the place the pinnacle of the Spot is holding in place after which the robotic wiggles round that, dances round that. After which the second if you see all 4 full—these distinct our bodies, and it appears like they’re dancing collectively. And we touched on that earlier—any form can dance—however making all of them dance collectively I assumed was actually sensible and efficient within the work. So it’s a type of moments, tremendous fascinating, or you may have a shaggy dog story about, I assumed we may speak about it additional.
Thomas: I’ve a shaggy dog story in regards to the hip joints. So the preliminary— properly, not the preliminary, however once they do the mashed potato, that was the primary dance transfer that we began engaged on, on Atlas. And for folk who don’t know, the mashed potato is sort of the toes are going out and in; the knees are going out and in. So we bumped into a few issues, which—and the twist. I suppose it’s a combo. Each of them such as you to roll your toes on the bottom like rub, and that friction was not good for the robots. So when we first started really moving into the twist, which has this torso twisting— the legs are twisting. The foot ought to be twisting on the ground. The foot isn’t twisting on the ground, and the legs have been so turned out that the form of the pelvic area regarded like a over-full diaper. So, I imply, it was wiggling, but it surely made the robotic look younger. It made the robotic appear to be it was in a diaper that wanted to be modified. It didn’t appear to be a twist that anyone would need to do close to anyone else. And it was actually wonderful how— I imply, it was simply hilarious to see it. And the engineers are available in. They’re actually seeing the motion and attempting to determine what they want for the motion. And I used to be like, “Effectively, it appears prefer it has a really full diaper.” They usually have been like, “Oh.” They knew it didn’t fairly look proper, but it surely was like—as a result of I believe they actually don’t mission as a lot as I do, I’m very projective that’s one of many ways in which I’ve watched work, otherwise you’re pulling from the work that means, however that’s not what they have been taking a look at. And so yeah, you then change the angles of the legs, how turned in it’s and no matter, and it resolved to a level, I believe, pretty efficiently. It doesn’t actually appear to be a diaper anymore. However that wasn’t actually— and in addition to get that transfer proper took us over a month.
Ackerman: Wow.
LaViers: Wow.
Thomas: We obtained a lot sooner after that as a result of it was the primary, and we actually realized. But it surely took a month of programming, me coming in, naming particular methods of reshifting it earlier than we obtained a twist that felt pure if amended as a result of it’s not the identical means that–
LaViers: Yeah. Effectively, and it’s fascinating to consider find out how to get it to look the identical. You needed to change the way in which it did the motion, is what I heard you describing there, and I believe that’s so fascinating, proper? And simply how distinct the morphologies between our physique and any of those our bodies, even the very facile human-ish trying Atlas, that there’s nonetheless plenty of actually nuanced and fine-grained and human work-intensive labor to enter getting that to look the identical as what all of us consider because the twist or the mashed potato.
Thomas: Proper. Proper. And it does have to be one thing that we will mission these dances onto, or it doesn’t work, when it comes to this dance. It may work in one other one. Yeah.
LaViers: Proper. And also you introduced that up earlier, too, of attempting to work within some established types of dance versus making us all terrified by the unusual motion that may occur, which I believe is fascinating. And I hope sooner or later you get to do this dance too.
Thomas: Yeah. No, I completely need to try this dance too.
Ackerman: Monica, do you may have one final query you need to ask?
Thomas: I do. And that is— yeah. I need to ask you, sort of what does embodied or body-based intelligence provide in robotic engineering? So I really feel like, you, greater than anybody, can communicate to that as a result of I don’t try this aspect.
LaViers: Effectively, I imply, I believe it could deliver a few issues. One, it could deliver— I imply, the primary second in my profession or life that that calls up for me is, I used to be watching considered one of my lab mates, once I was a doctoral scholar, give a speak about a quadruped robotic that he was engaged on, and he was describing the crawling technique just like the gate. And somebody stated— and I believe it was roughly like, “Transfer the middle of gravity contained in the polygon of help, after which choose up— the polygon of help shaped by three of the legs. After which choose up the fourth leg and transfer it. Set up a brand new polygon of help. Transfer the middle of mass into that polygon of help.” And it’s described with these figures. Perhaps there’s a middle of gravity. It’s like a circle that’s like a checkerboard, and there’s a triangle, and there’s these legs. And somebody stands up and is like, “That is mindless like that. Why would you try this?” And I’m like, “Oh, oh, I do know, oh, as a result of that’s one of many methods you possibly can crawl.” I really didn’t get down on the ground and do it as a result of I used to be not so outlandish at that time.
However right this moment, within the RAD lab, that might be, “Everybody on all fours, do this technique out.” Does it really feel like a good suggestion? Are there different concepts that we might use to do that sample that could be price exploring right here as properly? And so really rolling round on the ground and transferring your physique and pretending to be a quadruped, which— in my dance courses, it’s a quite common factor to observe crawling as a result of all of us overlook find out how to crawl. We need to crawl with the cross-lateral sample and the homo-lateral sample, and we need to maintain our butts down– or maintain the butts up, however we need to have that optionality in order that we appear to be we’re facile, pure crawlers. We practice that, proper? And so for a quadruped robotic discuss and dialogue, I believe there’s a really literal means that an embodied exploration of the concept is a totally respectable technique to do analysis.
Ackerman: Yeah. I imply, Monica, that is what you have been saying, too, as you have been working with these engineers. Generally it appeared like they may inform that one thing wasn’t fairly proper, however they didn’t know find out how to describe it, and so they didn’t know find out how to repair it as a result of they didn’t have that language and expertise that each of you may have.
Thomas: Yeah. Yeah, precisely that.
Ackerman: Okay. Effectively, I simply need to ask you every yet another actually fast query earlier than we finish right here, which is that, what’s your favourite fictional robotic and why? I hope this isn’t too troublesome, particularly because you each work with actual robots, however. Amy, you need to go first?
LaViers: I imply, I’m going to really feel like a celebration pooper. I don’t like several robots, actual or fictional. The fictional ones annoy me because– the fictional ones annoy me due to the disambiguation challenge and WALL-E and Eva are so cute. And I do love cute issues, however are these machines, or are these characters? And are we shedding sight of that? I imply, my favourite robotic to look at transfer, this one– I imply, I really like the Keepon dancing to Spoon. That’s one thing that in the event you’re having an off day, you google Keepon dancing to Spoon— Keepon is one phrase, Ok-E-E-P-O-N, dancing to Spoon, and also you simply bop. It’s only a bop. I find it irresistible. It’s so easy and so pure and so proper.
Ackerman: It’s considered one of my favourite robots of all time, Monica. I don’t know in the event you’ve seen this, but it surely’s two little yellow balls like this, and it simply goes up and down and rocks backwards and forwards. But it surely does it so to music. It simply does it so properly. It’s wonderful.
Thomas: I’ll positively be watching that [crosstalk].
Ackerman: Yeah. And I ought to have expanded the query, and now I’ll increase it as a result of Monica hasn’t answered but. Favourite robotic, actual or fictional?
Thomas: So I don’t know if it’s my favourite. This one breaks my coronary heart, and I’m at the moment having an empathy overdrive challenge as a normal downside. However there’s a robotic set up – and I ought to know its identify, however I don’t— where the robot reaches out, and it grabs the oil that they’ve created it to leak and pulls it towards its body. And it’s been doing this for a number of years now, but it surely’s actually slowing down now. And I don’t assume it even wants the oil. I don’t assume it’s a robotic that makes use of oil. It simply thinks that it must maintain it shut. And it used to pleased dance, and the oil has gotten so darkish and the crimson rust coloration of, oh, that is so morbid of blood, but it surely simply breaks my coronary heart. So I believe I really like that robotic and in addition need to put it aside within the actually unhealthy means that we generally establish with issues that we shouldn’t be serious about that a lot.
Ackerman: And also you each gave wonderful solutions to that query.
LaViers: And the piece is Sun Yuan and Peng Yu’s Can’t Help Myself.
Ackerman: That’s proper. Yeah.
LaViers: And it’s so lovely. I couldn’t keep in mind the artist’s identify both, however—you’re proper—it’s so lovely.
Thomas: It’s lovely. The motion is gorgeous. It’s superbly thought-about as an artwork piece, and the robotic is attractive and heartbreaking.
Ackerman: Yeah. These solutions have been so surprising, and I really like that. So thanks each, and thanks for being on this podcast. This was an incredible dialog. We didn’t have practically sufficient time, so we’re going to have to return again to a lot.
LaViers: Thanks for having me.
Thomas: Thanks a lot for inviting me. [music]
Ackerman: We’ve been speaking with Monica Thomas and Amy LaViers about robots and dance. And thanks once more to our friends for becoming a member of us for ChatBot and IEEE Spectrum. I’m Evan Ackerman.