Erik Demaine, the genius that plays for a living (21-04-2016)

Mathematician Erik Demaine talks about understanding complex problems and design solutions through play and gaming.

More videos with Erik Demaine

automatically generated captions
00:00:04 Right. Yeah well origami has a surprisingly rich mathematics and geometry to it.
00:00:16 It's I originally got interested in origami because it just posed a lot of interesting mathematical questions you have
00:00:23 this sheet of material and you have very simple rules you can't stretch it and you can't tear it
00:00:27 and what can you do just by a reconfiguration just by folding and so it's very kind of a simple set up
00:00:35 but the answers turn out to be surprisingly complicated and you need to use a lot of powerful geometry
00:00:41 and algorithms to figure out what you can fault in many senses of you can really fold anything out of a sheet of paper
00:00:48 and you can prove that mathematically. And that's sort of where we got started.
00:00:52 It's very exciting
00:00:54 and finding more interesting ways to make structure the fold between different shapes also has a lot of practical
00:01:00 applications in science and medicine
00:01:03 and engineering where you want to build some kind of structure that can change transform it shape from one thing to
00:01:10 another so maybe you want to fold it down to some small size for storage or transportation.
00:01:17 Like if you want to put something inside the body.
00:01:20 Maybe you need to transport through small blood vessels and so you need to make it very compact
00:01:25 or you want to deploy something into space.
00:01:27 You want to fold it small so it fits inside your space shuttle and then going to unfold it when it gets there.
00:01:34 I think it's even more exciting is you imagine like buildings or gadgets
00:01:40 or things that can transform from one shape to another and serve different functions depending on what you need.
00:01:44 Maybe your house a room in your house can transform from a kitchen to a bedroom and this kind of thing.
00:01:53 One of a couple of the areas that we were exploring are things like printing. Now to robots.
00:02:00 So there are a lot of rapid prototyping machines that are designed to make flat sheets of material.
00:02:09 And how can you use them to make three D.
00:02:11 Robots and other structures
00:02:13 and folding is a good way to do that you can transform your two dimensional sheets into some cool three D.
00:02:18 Structures so we one of our goals in this printable robot project is to make robots that can for like ten
00:02:26 or twenty dollars of materials you can cut them and make them within a couple of hours.
00:02:32 So everyone can make their own robot and customize their robot to do whatever they want.
00:02:37 Another fun application is in making nano scale structures so we have out of the whole computer chip fabrication
00:02:46 technology we have really good ways to pattern two dimensional surfaces at with nano scale features like nanometer
00:02:52 resolution. But we're not so good at making three D. Structures at that scale and so folding offers another way.
00:02:59 That's more in process and experimental but. An exciting possibility for folding is for you something that is.
00:03:11 What makes your position as a professor sort of roses. Yeah.
00:03:19 Yes I know lots of different players interested in different aspects of folding maybe more practical side I'm more on
00:03:25 the theoretical side and developing new mathematics
00:03:28 and tools to show to help those sort of kind of underlying technology for people to build on to make useful things.
00:03:39 So we especially like to prove what we call universe ality results where we say in this kind of regime of origami
00:03:45 design or folding design.
00:03:47 You can make anything you want and we give you a computer algorithm to do that
00:03:51 and so you can come in with your specifications like oh I'd like something that looks like this and it
00:03:56 and the algorithm will give you how to fold exam. Actually that thing.
00:04:02 And you know we get very general results they're not always the most practical because often we don't take into
00:04:09 consideration things like the thickness of the material or other kind of structural issues
00:04:14 and something we're trying to get to but by kind of simplifying looking at the core geometry we can get very general
00:04:21 and powerful results and then that those can be adapted to more practical scenarios where you are going for something.
00:04:34 You would think that there's a limitation but I don't know every year it.
00:04:38 I'm amazed at what origami artists come up with there's new people with new ideas
00:04:44 and it seems like almost limitless possibilities
00:04:48 and especially if you start with a large enough sheet of paper you can really fold really really complicated things
00:04:55 and there's still aspects we don't understand.
00:04:58 For example area we look at a lot is curved crease folding so most are Grammies made with straight creases curved
00:05:05 creases are a lot harder to understand and analyze and we're starting to make progress on the mathematics
00:05:12 but there's still a lot we don't know we don't have any good design algorithms to say Oh I'd like to fold something
00:05:16 that looks like this. Here's the curve creases you need to do that.
00:05:20 So instead we've been experimenting a lot with just playing around trying different curve crease patterns
00:05:26 and see what they produce and trying to be able to model that mathematically.
00:05:30 And that led my father
00:05:32 and I into the sculptural side of paper folding so most of the sculpture we made is make is around curved crease
00:05:40 folding initially we're just experimenting trying to figure out what's possible and what what can be done
00:05:46 but we kept making all these beautiful forms and so I started to embrace that as a purely sculptural.
00:05:53 Endeavor as well but there's a lot of back and forth.
00:05:56 I think that will do something sculpture really that will inspire new mathematics.
00:06:00 We discuss we understand something better about curved creases mathematically that inspires new sculpture
00:06:04 and so it's a lot of fun to go back and forth between the two cards right between art and science I think in general.
00:06:12 That's a big appeal to why people like to explore origami
00:06:16 and mathematics together because you have this sort of scientific purpose maybe an engineering application
00:06:24 or the beauty of the mathematics but then one of the applications is also to make sculpture.
00:06:32 So it's really exciting to see these kinds of collaboration is a lot of engineering teams are bringing on origami
00:06:37 artists at to help design new folding structures so artists have a lot of practical experience of how to make
00:06:45 interesting folding structures and then they know the the literature which is a lot of people folding stuff
00:06:52 and then that can inspire and inform new scientific discoveries. So your work your age. These are.
00:07:03 Yeah that's where we like to live is right on the edge of knowledge where we we have a lot of tools
00:07:08 but there's still something we don't understand.
00:07:10 And so we try to push push that frontier of what's what's known on the scientific side
00:07:16 and we use sculpture to kind of help explore that area more tentatively we can we can often make things that we don't
00:07:24 yet fully understand.
00:07:25 And so that lets us go a little beyond the frontier and sort of explore what's out there and see what's possible
00:07:31 and then.
00:07:33 Hopefully eventually understand that part mathematically where you where your science part formed you well what was it
00:07:40 like for you just what was there. Yeah well in general we're looking at unsolved problems.
00:07:50 I mean part of some sense one of the hardest parts is to figure out what the right question is. So you might want.
00:07:59 There.
00:08:00 Two types of questions about folding structures one is I give you a structure
00:08:04 and I want to understand its properties and sort of analyze what it does how good it is what it folds into
00:08:10 and the other side is the design side so you have some more high level specification of what you'd like to fold
00:08:17 and then you want to automate the design of a structure that folds with those parameters designs maybe the more
00:08:24 exciting side
00:08:25 and there's many different ways you might formulate what you want to fold sort of the classic origami design problem is
00:08:34 is shaped design. I say I give you a three dimensional shape. I want to fold that thing.
00:08:40 What's a good way to fold that thing.
00:08:41 And we're still we're still finding good algorithms for that we have some general procedures that work
00:08:47 but they may not be so efficient as one of the standard measures of efficiency is if I have a square of a particular
00:08:54 size material. How large of a version of that shape can I fold.
00:09:00 I don't want to fold a really tiny thing because that means I'm kind of wasting a lot of my material five a big square
00:09:05 filled with little microscopic things not very efficient material usage.
00:09:10 So how can we optimize that scale factor we still don't know the best way to do that.
00:09:15 There's a sense in which we can't know exactly how good we can do that
00:09:18 but we can hope to approximate the best solution.
00:09:21 So that's something we're still actively working on for example our current.
00:09:25 Favorite technique is called organizer and it's it's also free software
00:09:31 and it's an algorithm we've been analyzing over the last several years to give an arbitrary three D.
00:09:37 Shape that gives you a way to fold exactly that shape.
00:09:43 It seems to be a good method but we don't know it's the best method
00:09:45 and then there are many other questions based on other types of goals you might want like maybe you want to have a
00:09:52 folding structure that can make two different shapes.
00:09:54 We don't know much at all about that question or you want to make a folding that actually.
00:10:00 Works with really thick material because you're making out of sheet metal
00:10:02 or you want to make a practical mechanical structure.
00:10:07 We're still understanding that we've made some progress on
00:10:10 but there's still a lot of questions we don't know the best way to deal with with these kinds of practical issues
00:10:18 and so that as becoming really relevant these days because a lot of people are trying to build these structures
00:10:23 and sometimes it works sometimes it doesn't like to understand that threshold then
00:10:28 and ideally automatically design structures that always work really well in practice.
00:10:42 Yeah I mean I think we like to build objects and we.
00:10:46 And it's even cooler when those objects can change shape so almost anywhere you imagine a gadget of some sort.
00:10:51 I think folding could offer some interesting perspectives on on reconfigure ability.
00:10:59 C one one area we haven't talked about is protein folding which is a kind of origami it's a little bit different
00:11:09 but it's kind of essential to how just understanding how life works and also potentially drug design.
00:11:16 So every living thing that we know of in this world is built up out of lots of little proteins kind of making life
00:11:22 happen and proteins are centrally one dimensional pieces of paper that quite a lot into complicated three D.
00:11:28 Structures in that three D.
00:11:29 Structure kind of determines how it interacts with other proteins and what what its function is.
00:11:34 And we don't really understand that process of folding kind of a one dimensional strip of paper into these three D.
00:11:41 Structures how nature does it how we could do it how we could design proteins that fold into geometries that we want to
00:11:47 like combat. You can imagine some disease comes along new disease.
00:11:51 You could design a protein to fight specifically that disease but we don't know how to design.
00:11:58 Proteins that folded the way we want to.
00:12:00 And so we're trying to understand how proteins fold in order to sort of just understand how biology is functioning
00:12:06 but also so that we can kind of control it in useful ways to kill viruses and things like that.
00:12:12 So that's that's an exciting but difficult interaction.
00:12:17 I would really like a sort of universal programmable gadget you know like we have lots of gadgets where you can
00:12:24 download software updates like your smartphone you can download software updates and it does new things
00:12:30 but we don't yet have a gadget where we can download new shapes
00:12:34 or new geometries you can imagine a kind of universal gadget that can take on any shape I mean it has to preserve mass
00:12:43 that you could imagine it unfolding and becoming a large thing at folding into a more compact.
00:12:50 Structure changing shape maybe it's a chair or one one moment and it becomes a bicycle the next moment
00:12:56 or I mean anything in principle is possible.
00:12:59 It's we need to figure out what the practical regimes are
00:13:01 but instead of having a separate gadget that does different functions
00:13:06 or separate separate furniture that does different things you could imagine having fewer objects that are more
00:13:12 reconfigurable so that that excites me like I really like gadgets.
00:13:17 But I can have a gadget that can do more different things or be more customizable I think that's really exciting.
00:13:26 I owe a lot of people in the field.
00:13:28 Got into folding because they've been folding since they were kids and doing origami
00:13:32 and then they learn about mathematics
00:13:34 and think oh oh maybe we should combine these two I came from the other side so I was a beginning graduate student at
00:13:42 University of Waterloo and I was just curious. I was looking for interesting problems to solve.
00:13:49 I knew that I really like geometry and algorithms and. My father remembered an old.
00:14:00 Unsolved problems that he had read about
00:14:01 when he years ago from a column by Martin Gardner who used to write for Scientific American about mathematical games
00:14:10 and it's a problem that comes from the magic community and the concept is you take a piece of paper you fold it flat
00:14:18 and make one complete straight cut and then unfold the pieces
00:14:22 and magicians like Houdini could produce a five pointed star lots of different simple shapes
00:14:29 and Martin Gardner I was wondering you know what are the limits can you make anything by this process
00:14:35 or what can you do
00:14:36 and so that's the problem we started working on like OK I've got geometry in algorithms now seems like a cool unsolved
00:14:44 problem to work on and it turned out to be fairly challenging took us a year or two to solve
00:14:50 but it also was very exciting that we got our first universality result
00:14:55 and we showed that you can make any poly gun any sheet metal straight sides
00:14:59 or actually you can make several shapes all at once. Just by a one straight cut after folding. So those very exciting.
00:15:07 It's fun problem motivated by magic.
00:15:09 It turns out to have some practical applications also there are some designs for airbag folding collapsing airbags flat
00:15:16 that are based on the same kind of algorithm that we didn't intend that at the time
00:15:23 and it really got us excited about this world of folding where it seems to have very rich
00:15:27 and complicated mathematics but also those kind of fun and visual
00:15:31 and you can you can demonstrate these things you know you can fold a piece of paper make a kite and make a swan
00:15:37 or whatever shape you want to so it has an attendee ability that everyone can kind of appreciate even if they're not a
00:15:45 mathematician you can say hey look we solved this magic problem that's cool.
00:15:50 You know something for you where it was well before right. Well. Yeah it's going to start.
00:16:08 I guess my father became a single parent when I was two years old
00:16:13 and so we've been close for a long time especially from when I was ages seven to eleven
00:16:20 and we started traveling together and visited many different places.
00:16:25 Mostly east coast of United States and just traveling for fun.
00:16:29 There was no particular reason other than seeing different cultures within the United States
00:16:34 and exploring which was really fun and throughout that time my dad treated me as a peer. So we would.
00:16:42 Jointly decide where we're going to go next to how long to stay in a place some places we'd stay just for a few days
00:16:48 other places we'd stay for years and that was a really fun and bonding experience for us.
00:16:55 Growing up and also because we're traveling a lot.
00:16:58 We try to at home school and home school turned out to work really well for us.
00:17:04 I would spend only like an hour a day doing sort of the breadth of regular school
00:17:09 and so that I have many other hours during the day to explore things
00:17:13 and very quickly for me exploring was computer programming.
00:17:17 I got really excited about that essentially how to video games I played a lot of video games I was curious how they
00:17:24 were made and my dad knew a little bit about computer programming to get us started
00:17:27 and then we'd go to the library to learn more.
00:17:30 This was all before the Internet and so I was sort of racially learning about computer programming
00:17:37 and having a lot of fun there and then when school got out I would go and play with kids and things like that.
00:17:42 So that was a really great time for me growing up and I went very fast in the computer science
00:17:48 and eventually mathematics side of things right. So right. Yeah I asked over that when I was.
00:18:00 Five or so and six years old my dad and I had our first collaboration we like to say with the Eric
00:18:07 and dad puzzle company.
00:18:08 So I helped design wire take apart puzzles and my dad would make them bending wire
00:18:17 and then we sold to twenty stores across Canada and we split the income fifty fifty and it was a lot of fun.
00:18:25 That was definitely the beginning of my interest in puzzles which is still to this day something I had like a lot
00:18:33 and probably also the beginning of my interest in mathematics and geometry
00:18:37 and things like that although that came much later as we were twelve. Yes yes. So after we ended this travel.
00:18:48 I wanted to learn more about computing in computer science I learned was a thing
00:18:54 and you have to go to university to learn about it.
00:18:56 So there was some complication but I started undergraduate at twelve
00:19:02 and took lots of classes because I at that age you can really soak in a lot of material and so I ended up finishing
00:19:08 when I was fourteen and then went to graduate school and got a master's and Ph D.
00:19:15 By the time I was twenty and then went on the job market and became a professor here at MIT. You're right. Yes.
00:19:32 Yeah it's really. We really value. Having fun and enjoying the work that we do.
00:19:39 There's a very there's essentially no line between the work that we do and the things we do for pleasure.
00:19:46 So it's all mixed together just with different kinds of outcomes maybe becomes a math paper maybe it becomes a
00:19:54 sculpture maybe it's there's no outcome we're just doing it for fun but.
00:20:00 It's all for fun and the philosophy is that if we do work that we enjoy
00:20:06 and find pleasurable then we'll do it very well excel at it and that has been a useful guiding principle
00:20:15 and I would encourage everyone to do the same it's definitely it may seem risky at times
00:20:19 and certainly there was a worry that the work that we do is to recreational like you know we're studying the
00:20:25 mathematics of a magic trick how could that be useful for anything but it turned out to be unexpectedly.
00:20:33 But I think a lot of specially in mathematics there are just a lot of basic questions that are very curious
00:20:40 and you want to know the answer to
00:20:42 and if they're basic enough the sort of very simple set up like paper folding is a very simple set up a very few rules
00:20:47 about what's what you're allowed to do and yet it's very complicated to understand it's a nice context for.
00:20:55 I think basic research tends to become useful eventually even though you may not see the applications ahead of time
00:21:01 and so mathematicians tend to be attracted to like very simple questions that have complicated answers.
00:21:07 Those tend to be also useful questions to answer always but if you solve enough of them.
00:21:12 Many of them will become practical and so even though you do it for fun.
00:21:17 It tends to have useful applications as well so that you might be worried by a lack of applications
00:21:24 but turns out to be OK. It's for real life. It's really sweet.
00:21:35 I mean we have pretty much ideal set ups where we can work on what we enjoy and get paid for it
00:21:41 and have fun doing it and have all the resources to do it. We're very lucky.
00:21:52 Yes So the glass blowing interest comes from.
00:21:55 My dad's background which is more on the visual arts side so before I was born.
00:22:00 In the late sixty's early seventy's he had the first glass studio in Canada.
00:22:04 It's called the father of Canadian glass
00:22:07 and so he had to have a studio made lots of glass work it was it was the early days in the studio movement of glass
00:22:15 blowing in North America
00:22:16 and so he was experimenting exploring what's possible trying different recipes to make glasses and glass colors
00:22:23 and things and then he didn't blow glass for many years until
00:22:30 and I never really saw him blood last until we came to MIT fifteen years ago
00:22:35 and we discovered he'd MIT has a glass blowing studio called The Glass lab and so my dad got curious to try
00:22:44 and glass flying again.
00:22:45 And so he started teaching there became one of the instructors and started blowing glass again
00:22:52 and I got to see him blow glass and watched him make things and it's so beautiful and amazing to watch
00:22:58 and then eventually like maybe maybe I should try clasp like that said Yeah you know you should at least see what it's
00:23:05 like but be careful. It's addictive.
00:23:07 So I quickly got into a glass blowing and now we blow glass together and make things together and it's a lot of fun.
00:23:22 It's a little more difficult because there's a lot of physics going on with glass blowing which is not exactly my forte
00:23:28 but we're always looking for interesting math and connections between mathematics and glassblowing and we've found.
00:23:36 We've found some interesting books there I think there's still a lot more to be explored.
00:23:40 I would love to have algorithms to automatically design interesting because this sort of operations you can do are
00:23:46 glassed in glass blowing a very simple. You know you can.
00:23:50 You're turning your piece you can swing it around you can play with sort of gravity in this way you can heat different
00:23:56 parts and cool other parts and that totally changes the shape that you produce.
00:24:00 But it's a very complicated relationship and so it's hard to model all of that mathematically.
00:24:06 But we've found some interesting regimes where it's simple enough that it's mostly geometric what's going on
00:24:13 and so we can use computers to help design new patterns in glass.
00:24:18 So we have some free software called virtual glass that we've been developing where you can design what are called
00:24:24 Glass came patterns very simple.
00:24:29 Conceptually simple but hard to visualize where you set up some essentially straight lines of color and glass
00:24:35 and then twist them.
00:24:36 And so you get some really cool twisty patterns they've been used in glass flying for for centuries. But.
00:24:45 Pretty much everyone who makes glass cane follows one of standard set of patterns
00:24:50 and so we were curious whether there were more patterns for Glass can that were possible in the software lets you
00:24:56 explore those patterns and lets you try new things and sometimes you try a new thing
00:25:00 and it looks kind of like an old thing.
00:25:02 So it's not really interesting but other times you try a new pattern and it looks amazing in the software
00:25:06 and that tells you here.
00:25:08 This is something we should spend the time to actually learn how to make in real life software doesn't tell you exactly
00:25:13 how to make it but it gives you a kind of schematic and then you have to do the glass blowing hard work
00:25:19 but at least you know that the thing you're trying to make is really beautiful and so it's worth working towards.
00:25:26 So you can rapidly try lots of different designs and software to find the one you want and then go physically make it.
00:25:33 So it's really hard.
00:25:38 Yeah I think it's I mean I think in general working on the boundary between two different fields you find interesting
00:25:44 areas that.
00:25:46 People tend to specialize in just one area and so they miss the things that the boundaries
00:25:52 and so we've had a lot of fun exploring these boundaries
00:25:54 and I think it comes partly from our different backgrounds my dad with the art background me with the more math
00:25:59 and science. Background and we're always talking to each other and so we see we see the connections.
00:26:06 When I started graduate school I was doing this sort of more theoretical mathematical work my dad's side
00:26:12 and saying looks. That's interesting.
00:26:14 This kind of creativity you're going through
00:26:16 and solving unsolved mathematical problems is very much like the kind of thing that I go through in designing new
00:26:23 sculptures or thinking about new art to build
00:26:27 and so we started working together then he got he I taught him to become a mathematician.
00:26:33 And he taught me to become an artist and so now we work on both together
00:26:37 and it's really it's a lot of fun for us to collaborate in that way but also leads to really interesting questions
00:26:44 and inspirations where instead of just thinking OK the math is the serious stuff
00:26:50 and everything else is just you know side project we think of everything is like main projects
00:26:55 and they inspire each other in ways that we couldn't predict. So I'm just working for years. We're definitely.
00:27:15 Yeah I mean you could say frontiers of science and art maybe.
00:27:20 Or that interplay but you know we're always as scientists we're always excited about the unknown
00:27:26 and I mean that's as soon as we understand something fully it becomes almost boring
00:27:33 and we want to move on to the next thing I mean we write down what we know and publish it
00:27:37 and share it with the world so they can build on top of it
00:27:40 but then we're always excited about the next question which we don't understand that's that's really what drives us is
00:27:48 the price that we don't quite understand or like that seems a little strange. And we're curious about and.
00:27:57 Yeah that's that's where we explore next. You know years at. Wells you might. Yeah it's a good question.
00:28:18 I think in the in the folding regime. I work in many different areas but in the folding world.
00:28:26 I think the biggest challenges right now are taking the nice mathematical geometric design algorithms that we have
00:28:34 and adapting them to to real world materials.
00:28:38 So we're starting to look at how does the thickness of the material affects what we can fold.
00:28:44 How does the rigidity of material affect what we can fold off and you're making things out of plates and hinges.
00:28:50 So you can really only fold at the creases whereas on paper it's more flexible.
00:28:55 Between the creases So this is a world called rigid origami still trying to understand how to design within that space
00:29:02 but it's very practical and exciting and for us it's nice
00:29:05 and challenging because we don't know that's what that's what we don't know how to do
00:29:09 and so that's where we're attracted So I think in the next couple of years we'll make a lot of progress in that kind of
00:29:14 trying to take the rich and very general mathematics and adapting it
00:29:19 or how to deal with the parameters of real world materials where you know other areas where you
00:29:29 or where you're working well yeah I see there are a lot of so I mean like the traditional origami set up is you have a
00:29:45 square paper and all you can do is fold and it's really interesting to see what you can do just by folding
00:29:52 but there are a lot of practical setups like in our printable robots project where it's.
00:30:00 Also find to cut the material I mean why not. It's folding is very powerful it's a good way to go from two to three D.
00:30:06 but We don't have to start from a square of material probably we're starting from some kind of rectangle of sheet
00:30:12 material and why not also cut it in two dimensions before you fold
00:30:17 and that's exciting because it can lead to much more efficient foldings potentially use all of the material now
00:30:24 and you can make structures you couldn't make just by folding or you can you can make them much more efficiently
00:30:30 and in different ways it's also a little it's tricky from a mathematical perspective because now we have so much more
00:30:36 freedom we can cut and fold in some sense it's more freedom than we know what to do with
00:30:42 and so that's that's kind of a new direction of folding where we also are cutting because why not.
00:30:48 It's a practical thing you can do
00:30:51 and maybe there are some settings where you want to add lots of cut some settings where you want to add fewer cuts we
00:30:56 don't know the right balance between us and I think that's a new frontier we're still exploring
00:31:01 and trying to understand but potentially leads to much better ways of folding structures.
00:31:07 And going back to your you're very free for. Case don't have this space. If you go.
00:31:35 I think it played a big role I mean it's hard to know exactly
00:31:38 but I think growing up with so much free time unstructured time where I could just explore what interested me really
00:31:47 gave me a big edge. Instead of sort of wasting time which a lot of schools do just filling that time.
00:31:56 So as a kind of child care. It's set up.
00:32:01 There's social aspects which are good too but a lot of time I feel like is wasted in school
00:32:07 and so having the home school opened up this window where I could explore what interested me and
00:32:13 and really dive in deeply and that let me go far ahead in the computer science world
00:32:18 and I think in general could let students go really far ahead in the thing that excites them the most you still have to
00:32:25 add in the breadth and socialize with other kids and so on but really
00:32:31 and then going to university at a young age I think really gave me another edge whereas you can learn so much at a
00:32:39 young age and so when you get to university.
00:32:41 Suddenly there's really interesting things you're learning and it's really exciting
00:32:45 and I still remember the things that I learned back then.
00:32:50 So that's really powerful as a way to to get started and I think a lot of people could do it.
00:32:58 There's also a more general sense of.
00:33:02 Because we were improvising if we want a long travelling around we would talk to our neighbors learn about what they
00:33:10 knew about and if they knew some interesting topic they would teach me and teach my dad.
00:33:15 So I learned different aspects about the magic that way.
00:33:17 I learned different kinds of cooking that way
00:33:21 and that was a fun way to say it to appreciate different people of different backgrounds
00:33:28 and different knowledge sets and I think in directly that influenced me
00:33:33 and my dad to think a lot about collaboration
00:33:38 and in current day we we collaborate with a lot of different mathematicians
00:33:43 and different papers I've written papers I think over four hundred people at this point
00:33:48 and on the art side we're also looking for collaborators interesting ways to combine different ideas from different
00:33:54 minds we collaborate a lot with each other. Of course but also looking for outside inspiration. I thing.
00:34:00 When you combine multiple people together you can really you can solve problems that could not be solved individually
00:34:07 on the mathematical side this is because there's just so many areas of mathematics. You can't really know all of them.
00:34:13 But some problems require lots of different tools to solve and so you can either go
00:34:19 and learn about that tool it takes a long time
00:34:21 or you could just collaborate with the person who ARE THE knows the tool
00:34:24 and they can solve that piece of the problem really. Well you can solve your piece you combine the right people.
00:34:29 You can solve big problems relatively easily
00:34:32 and on the art side you get inspiration things that no one person could make because they have the creative voice from
00:34:39 multiple people you have to be willing to let go of your own ego to do this and I think that probably for my dad
00:34:47 and I came from this period where we're just kind of exploring together and being open to the people that we meet
00:34:52 and learning from them. Not that I know of. And it's an interesting challenge to try to model.
00:35:06 Fun or humor or surprise. Mathematically I've heard.
00:35:10 I know I have some friends who are trying to answer that question but I don't know of one sort of I usually go by.
00:35:21 You know it when you see it kind of definition. So it was like. Each Other. Large I see here.
00:35:37 Yeah it's certainly a fascinating topic to think sort of at a high level of like mathematics for example has a kind of
00:35:48 a branch mathematical logic where tries to understand where we try to understand mathematically what mathematics is
00:35:55 and why it works or when it works when it doesn't work. But.
00:36:01 Of course the mathematics we practice in real life is a kind of a social dynamic you know do you believe
00:36:08 but someone claims they have a proof written down there
00:36:10 but to really check the proof you have to check it very carefully and it's humans are perfect
00:36:16 and so it's there's a social dynamic to the.
00:36:20 The body of research we we create and in some ways it makes it more fascinating
00:36:27 and colorful that that kind of mind share of what we know
00:36:31 or what we think we know is always kind of changing usually we're adding things we think are true
00:36:36 or that we claim are true.
00:36:37 Sometimes we take them back away we look at an old theorem and people have been building on
00:36:42 and realize oh actually that proof is wrong
00:36:45 and then there's this flurry of activity trying to fix the proof make a new proof so that the results that are built on
00:36:51 it are the result may still be true but sometimes we need to find a new way to prove it.
00:36:56 Sometimes the results and being false. That's that's more. It's occasionally scary but it's exciting.
00:37:04 Always trying to discover new things but also make sure they're really correct.
00:37:09 And definitely to me one of the appeals of mathematics is that you there is at least a sense of real truth of ultimate
00:37:17 truth that in principle if you're doing it correctly
00:37:22 and you prove something you really know that it is without a doubt true.
00:37:25 There's no other area where you can be as certain but still even then
00:37:30 or stuff like quite certain because humans make mistakes all the time. Yeah.
00:37:37 So I mean certainly we can see a lot that we can kind of build up lots of evidence that something is true by
00:37:45 constructing lots of examples and sculpture and and more practical engineering structures and so on
00:37:50 but to know that it's always true is a little bit different to know that it's usually true. This start.
00:38:01 Well you're right.
00:38:09 You have puzzles of remained an active interest
00:38:11 and in some sense all the mathematics we do is a kind of puzzle we have some set up of like what you're allowed to do
00:38:18 say with paper folding are some other simple mathematical structure
00:38:22 and the puzzle is you know what's possible what can you make these are kind of met a puzzle set sense
00:38:28 but even puzzles themselves like the kind of board game puzzles you get or the like sliding blocks
00:38:35 or these kinds of things are actually really interesting to study mathematically as well.
00:38:39 And so my dad and I and many collaborators like to explore the mathematics of games and puzzles
00:38:47 and we do it for video games like we studied Tetris and Super Mario Brothers
00:38:52 and other Nintendo games that I grew up playing now I can study them mathematically
00:38:57 and the sorts of things that we prove are that it's really hard to play these games perfectly.
00:39:03 So if you if I give you a level of Super Mario Brothers and say can you get from start to finish.
00:39:09 That's actually computationally difficult problem
00:39:13 and you can prove that solving that problem is really hard for a computer to do
00:39:18 and my philosophy is that humans are essentially a kind of computer
00:39:22 and so that tells you that it's also really hard for humans to play these games perfectly
00:39:27 or to solve these puzzles to play a Tetris game optimally
00:39:31 or to slide blocks around to get one block out of the box all these problems are really really hard
00:39:37 and I think it helps explain for humans.
00:39:40 Why we enjoy them because humans like a challenge things should be challenging but not too difficult
00:39:47 and proving these problems are computationally difficult.
00:39:50 They're still solvable given enough time
00:39:51 but in general you need an amount of time that grows exponentially with the size of the puzzle and so.
00:40:00 It means it's beyond a certain size it really becomes intractable and even in a small size it's a challenge
00:40:05 but still feasible.
00:40:07 I think that's why it's fun to have this kind of mathematical justification for why we like playing games
00:40:13 and puzzles and it's also a fun way to explore puzzles and games that I grew up with or know or love
00:40:20 and be able to study studying them from a mathematical perspective lets me essentially play the game
00:40:27 but in a more interesting way in some ways.
00:40:30 Usually we have to design new levels new puzzles within a design space in order to show that.
00:40:38 Oh we can build like logic gates and we can essentially build a computer within this game or puzzle.
00:40:44 And that's how you show that it's hard for a computer to play because computers are not it's really hard for a computer
00:40:50 to simulate a computer sexually.
00:40:53 That's sort of the hardest thing that they can do
00:40:55 and so we get to have fun by playing became by designing new levels and so on.
00:41:03 In order to prove these kind of interesting mathematical results that actually this game is really challenging.
00:41:09 And difficult.
00:41:15 So it's a bit of both Certainly I also just like playing games playing board games playing video games
00:41:22 and that's a lot of fun. Just as. Mediums to explore human experience I guess.
00:41:32 And I like the the role playing aspects I like the having fun with friends aspect
00:41:39 or exploring a world that's I mean these days.
00:41:42 Video games tell really powerful stories and so it becomes a new medium for storytelling.
00:41:47 So lots of more personal and just sort of fun aspects like that as I would be as a as a kid playing again.
00:41:54 And there's definitely a lot of nostalgia playing even playing old games as.
00:42:00 New Again they still hold up as being very exciting. But there's always even when I'm playing just for fun.
00:42:07 There's always in the back of my mind thinking I wonder if we can set this up as a clean mathematical problem
00:42:13 and analyze the complexity of this game
00:42:17 and some games are more amenable to this kind of mathematical analysis some of them require some adaptation to be a lot
00:42:24 of games have a lot of different elements it's really complicated mathematics is really good at getting at the core of
00:42:29 a problem. So it's a lot better when you have set up a simplified version.
00:42:33 Maybe you say oh let's just focus in on this one particular aspect of the game which if that's harder than the whole
00:42:40 thing is of course also even harder and so you can kind of isolate out the different parts
00:42:45 and tease out an interesting mathematical problem out of a real game or puzzle and then analyze that. So I mean.
00:42:52 It's it all fits together so as I'm playing a game. I'm always thinking about.
00:42:56 I wonder what I can tease out of this game and as I'm playing and having fun.
00:43:01 I'm also trying to think about that that mathematical formulation so.
00:43:05 It's good because then you get inspiration for new problems to solve just by having fun all day so.
00:43:17 Yeah I should get one.
00:43:19 But we made these wire take apart puzzles so each It's multiple pieces each piece is just made out of a piece of metal
00:43:28 wire that my dad would bend with pliers into shape and so one shade might be trouble clef
00:43:36 or some some recognizable shape or house I remember designing that one and then there be other pieces attached to it.
00:43:44 Everything's made out of wire.
00:43:45 Maybe some metal rings also and so these pieces appear to be interlocked and the challenge is to separate them.
00:43:52 And so while they look interlock there's actually some complicated procedure for pulling one piece at a.
00:44:00 Yeah there and you've solved the puzzle of that you have to put it back in and give it to someone else to solve.
00:44:04 So these are challenging the kind of a mix of geometry and apology in their design and.
00:44:14 Yeah there are a lot of fun you can be quite hard.
00:44:17 Some of them require hundreds of moves to solve some of them are easy if you know how but yeah.
00:44:27 So I think that the challenge of a human playing video games comes from different elements
00:44:34 and some parts are easy for computers to solve and other parts we sure are difficult for a computer to solve.
00:44:39 So if you imagine like solving a level of Super Mario Brothers there's there's kind of the physics of like the physical
00:44:48 aspect of pushing the buttons at the right time either pressing them really quickly
00:44:51 or exactly the right moment just before you fall off the ledge you jump
00:44:55 and land in the right place that sort of thing that computers are actually really good at doing
00:44:59 and there are people who exploit that in the tool assisted plays plays of games where they use computers to like slow
00:45:08 everything down and.
00:45:10 Time exactly the right moment to make a jump and things like that the computers can do some aspects really well
00:45:16 but there's kind of a there's a broader medal level in solving a puzzle
00:45:19 or solving a level in a video game where you need to plan out I should do this thing first
00:45:26 and then I'll go do this thing there's a time limit
00:45:28 and so it's really sensitive to how I should plan out the overall execution of the level executing it may be hard for
00:45:34 human easy for a computer but the planning part for a sufficiently complicated game is usually really difficult
00:45:41 and you can prove that that's computationally challenging now real world levels
00:45:46 and puzzles are usually designed to be right at the edge where you have to try several different options
00:45:51 but it's not impossible but in some sense
00:45:55 and that challenge comes out of this broader setting if you have a really large.
00:46:00 Well you can encode really hard problem inside that that puzzle and solving it.
00:46:05 You can show us how hard even for a computer to play.
00:46:09 So another example is like Tetris you have Tetris The usual the practical challenge is that you have this time limit.
00:46:16 You know the pieces falling you have to decide where to put it really fast computers are good at doing things really
00:46:21 fast but deciding where to put is actually really hard
00:46:26 and you can you can show that sort of long term planning of where you should put your pieces so that you won't run out
00:46:31 of space in your Tetris board.
00:46:33 That's actually competition intractable also so it's interesting I think real world games you have this interesting
00:46:40 mixture of making it hard for a human by giving time limits or physical execution can be challenging
00:46:45 and then you have this mastery aspect which very appealing to gamers.
00:46:50 But then usually there's this underlying computational difficulty that.
00:46:53 So you're solving this hard problem under time constraints that it's exciting for people. It's research for your site.
00:47:16 There definitely are some consequences I have a lot of games
00:47:21 or in some sense a like video games are often an abstraction of a real world problem.
00:47:26 Typical example is motion planning.
00:47:29 So you're either you have a bunch of robots
00:47:32 or you're a bunch of people trying to execute some goal you have a lot of objects you want to rearrange them into a
00:47:38 particular pattern
00:47:39 or maybe you're in a warehouse moving products around the every every product has a place it needs to go.
00:47:44 What's the optimal way for moving all these parts around.
00:47:48 That's those kinds of problems end up in a lot of video games also usually in a somewhat abstracted simplified form.
00:47:55 So proving those problems are hard shows also that.
00:48:00 These kinds of more real world problems are difficult as well maybe you can
00:48:04 or it helps you maybe try to isolate what are the what is special about the real world instances maybe your warehouse
00:48:11 is mostly two dimensional because you don't stack lots of things
00:48:14 or what are the it's speciality is of the real world instance that make them easier than the videogame So there's
00:48:22 definitely that kind of interplay
00:48:25 but I think a lot of people study the Myself included study the complexities of these games
00:48:31 and puzzles because it's fun and it's kind of a recreational pursuit.
00:48:35 So it's a little bit less serious than some areas of mathematics computer science but still we enjoy it
00:48:42 and it's kind of a fun.
00:48:43 I use it a lot as a way to get students excited about research because most people come in with their own background of
00:48:50 like what are fun games and puzzles that they grew up playing
00:48:54 and those inspired new mathematical problems either directly about those games
00:48:59 or about sort of the underlying principles
00:49:02 and these kinds of hardness personally call them to show these games are competition intractable are a nice way to get
00:49:09 started in research because you get to play with the game you get to use the expertise you have from having grown up
00:49:15 playing this game.
00:49:15 You probably spent way too many hours playing them
00:49:17 and that expertise is actually really helpful for solving the underlying math problem
00:49:21 and it can get people excited about. Oh this is this is computer science research I want to do more of this.
00:49:28 There's also the I think the broad appeal of you know there's some mathematical results that are hard for.
00:49:36 The general public to appreciate.
00:49:38 But you analyze a game or a puzzle that everyone has played or big segment the population is played
00:49:43 and they can appreciate like oh yeah I remember that being really hard.
00:49:46 Oh you can prove that mathematically Oh that's interesting.
00:49:49 I wonder how they do that and that can inspire people to enter the field or at least get a curiosity or
00:49:55 and learn about fields that they're not necessarily working in and appreciate that.
00:50:00 There's interesting things you can do about problems I happen to care about because most people like games
00:50:06 and so this isn't a nice kind of broad appeal connection where our years ten years.
00:50:21 It's hard to know exactly where my research will take me I definitely like MIT as a base because it's I mean there are
00:50:30 mazing students here amazing people doing all sorts of great and crazy things
00:50:34 and just a lot of flexibility to essentially do what we want and explore whatever we find interesting.
00:50:40 So what will be most interesting to us in ten years is hard to guess but this definitely is a nice.
00:50:49 Powerful base to do it from so different enjoying my time here.
TV Get inspired and watch tv episodes of The Mind of the Universe, made by Dutch public broadcaster VPRO
  • Browse through over 30 hours of interviews
  • Download the interviews, including subtitles
  • Remix, re-use and edit under CC-BY-SA license
  • Start exploring