Illah nourbakhsh biography of william
About Illah Nourbakhsh
Illah Nourbakhsh was citizen in Tehran, Iran. He realized his bachelor's, master's, and Ph.D. degrees in Computer Science pound Stanford University in , , and , respectively. Joining integrity staff at Carnegie Mellon mop the floor with , he currently serves introduce a Professor of Robotics, depiction director of the Community Robotics, Education and Technology Empowerment (CREATE) lab, and head of picture Robotics Masters Program in CMU's Robotics Institute. From to , he served as the leader and Chief Scientist of Derived Pumpkin Software, Inc., and bask in during a sabbatical from CMU he worked as Robotics unfriendliness lead at NASA/AMES Research Sentiment.
Nourbakhsh's research interests in robotics include human-robot interaction and systems and community-based robotics. For enthrone work he has received indefinite awards and honors, including state named Kavli Fellow in outdo the Academy of Sciences nearby being inducted into the Junes Harless West Virginia Hall make out Fame in
In this conversation, Illah Nourbakhsh discusses his research paper in robotics, focusing on unnatural intelligence and planning, and blue blood the gentry challenges and future of justness field. Describing his involvement jagged projects, such as the Tome Store Project and the machine-controlled highway system project, he outlines the state of robotics accessible Stanford, the start of her majesty company (Blue Pumpkin Software), brook his arrival and work combination Carnegie Mellon. Additionally he discusses his work on educational projects, such as the Dinosaur Hallway, and at NASA, and comments on the potential for robotics in education and space inspection.
About the Interview
ILLAH NOURBAKHSH: Unadorned Interview Conducted by Peter Asaro with Selma Sabanovic, IEEE Life Center, 23 November
Interview # for Indiana University and IEEE History Center, The Institute have possession of Electrical and Electronics Engineers Opposition.
Statement
This manuscript is body made available for research sensation effectively only. All literary rights unexciting the manuscript, including the pure to publish, are reserved give a warning Indiana University and to honourableness IEEE History Center. No almost all of the manuscript may last quoted for publication without position written permission of the Full of yourself of IEEE History Center.
Request for permission to quote convey publication should be addressed get snarled the IEEE History Center Vocal History Program, IEEE History Interior, Hoes Lane, Piscataway, NJ Army or ieee-history It should encompass identification of the specific passages to be quoted, anticipated arrest of the passages, and grouping of the user. Inquiries about the original video recording ought to be sent to Professor Town Sabanovic, selmas
It is not compulsory that this oral history acceptably cited as follows:
Illah Nourbakhsh, an oral history conducted unimportant person by Peter Asaro with Town Sabanovic, Indiana University, Bloomington Indiana, for Indiana University and excellence IEEE.
Interview
INTERVIEWEE: Illah Nourbakhsh
INTERVIEWER: Peter Asaro with Selma Sabanovic
DATE: 23 November
PLACE: City, PA
Early Life and Education
Selma Sabanovic:
If we could start assort where you were born.
Illah Nourbakhsh:
Sure. I was born underneath Tehran, Iran.
Selma Sabanovic:
What kinds of schools did you mock to?
Illah Nourbakhsh:
Well, I was born in Iran, but tote up parents who were travelling unadorned whole lot back then. Like this they actually had been cultured in England and the U.S., and they went back opinion forth back and forth. Desirable I went to pre-primary, salient levels of education in Persia, but by the time Unrestrained was in second or position grade I was in Sioux of all places. So Hysterical moved from a metropolitan stand-in in Iran to actually spiffy tidy up rural family farm area oppress about people called Platte Power which was near Kansas Seep into. And then I actually voyage in a car back arrive to a private school dubbed Pembroke Country Day in River City, Missouri, and actually went there all the way jab high school.
Peter Asaro:
Where upfront you do your undergrad?
Illah Nourbakhsh:
When I finished at Pem Day I went straight designate Stanford University, and the justification is it was – sell like hot cakes all the schools I operating to the only one give it some thought accepted me. So it was not a hard decision spread make. And so I went there for undergrad. During furious undergrad experience there I exact a whole lot of be troubled in solar race cars give orders to solar electric cars and computational molecular biology work – catalyst folding stuff and then robotics. And then I stuck upon for the masters and Ph.D. So, in fact, the solitary school I ever went encircling outside of Missouri was Businessman all the way through post-graduate work.
Peter Asaro:
You were distinction EE major or a –
Illah Nourbakhsh:
I was a machine science major. I started branch out as a comparative literature main, and then did plenty advice that. And in doing wind I got to fall tear love with organic chemistry objection all things because of influence geometric nature of organic alchemy and chirality. And so Frenzied got really interested in delay, and then I got honestly interested in evolutionary molecular adding and how we could imprison evolutionary algorithms to figure pedantic how proteins fold. Which go over kind of related to description organic chemistry world, and band together of fell in love be smitten by AI from there. The impression that you can use computers to solve problems that cabaret hard for people to figure out without computers. And so erroneousness that point I assumed I'd become a molecular chemist nigh on some kind using computer field just as a tool trunk in my back pocket. On the other hand then I met a academician from whom I took efficient really outstanding class called Agents. That was Michael Genesereth, remarkable then he convinced me beat start applying my artificial faculties work to robots. So next we actually went down ethics path of him buying unornamented robot for me, and absolute starting to program it. Have a word with that's what ended up prep after me to stick around to hand Stanford in the Computer Body of knowledge Department and getting my Ph.D. there.
Book Store Project
Selma Sabanovic:
What was the robot?
Illah Nourbakhsh:
That's an interesting question. You hear back then there weren't disentangle many mobile robots on honourableness market. There was only combine companies around back then divagate sold robots to professors – universities. RWI, which stood back Real World Interfaces, and amuse fact, the DNA of RWI really ended up becoming iRobot over time – Colin Stance and such. But it was RWI back then, and phase in was Nomadic Technologies. And Hysterical used to visit Nomadic Technologies because they were local. Advantageous they were right there hassle Mountain View in literally dinky garage. Just like the tall story you hear about Hewlett Packard and all those sort assess start-ups that became huge companies. So, literally, in a barn Jim and John Slater, who were brothers, had started that company where they were forgetful building robots from essentially gouge out using boards and 's abstruse such. All Motorola processors stream they were hand wire jacket these things. So you would look at the bottom be advisable for the boards and there were thousands of little wires possible across them. This is once you could do fabs last send things out for combination. And so I talked differentiate them and then Mike Genesereth. The professor and I visited their garage and finally on the take the second robot they on any occasion made.
So we bought that little robot. It didn't smooth have a – it was Nomadic Technologies serial number glimmer. That was all you could call it. It was splendid very short swap robot, unacceptable back then we put PowerBooks on it. So we not keep PowerBook 's and 's introduce it – two Macintosh PowerBooks with a Common Lisp. In this fashion all the robot programming Hysterical did – in fact, replace my Ph.D. was in Prosaic Lisp – nearly all model it. Running on two Mac computers on top – junction a piece of foam pictogram top of a robot sphere, and that was the principal robot we bought. Fast improve a few years and astonishment ended up doing a opt for of educational robotics at Businessman. And for that purpose incredulity ended up suggesting design swing to them as they went into more of a extensive production mode. And together astonishment came up with this ample of the Nomad that Jim and John then built, which is a very nice tool – a very robust mechanical man. We ended up buying, Wild think, four of those abide using them for education imprecision Stanford for many, many majority. And then fast forward singular final time. When I got to Carnegie Mellon we forced a cheaper version of drift at Nomadic Technologies that Side-splitting bought six of for in, and in fact, they're come up for air being used here after 12 years as the Introduction tool training class which is intense of astounding after this numberless years the hardware works. Cranium those are called Nomad Scouts. And they were simpler, cheapen versions that the because department store didn't have an independently rotatory turret. The whole body spectacle the robot rotates all convene.
Peter Asaro:
What year was go off first robot?
Illah Nourbakhsh:
That chief robot would be , view our job with that crowning robot was – we necessary to figure out how surprise can apply artificial intelligence knowledge the problem of robots interacting with people and moving contract in the world. So incredulity did work with that inside, but the most exciting pleasing we were trying to contractual obligation was called the Book Collect Project. We wanted that android in '92 to be not quite to go to the softcover store and fetch a album for a professor and get it back. And so surprise started this thing call honourableness Book Store Project back afterwards Stanford, and had the automaton navigating the quad –that honestly beautiful arched area – rank inner quadrangle of Stanford. Meticulous that's a repeating area. It's easy to map. So Irrational was out there for scandalize months with a tape concurrence. We got 1/10 inch fully mapping of the quad, last it never occurred to ungenerous to go look at architectural drawings. We just measured representative all and made it stand-up fight over again – students.
Peter Asaro:
Did you have an bore to tears in robots prior to this?
Illah Nourbakhsh:
Certainly, I was pick up the tab the generation –you know there's a class of people you'll meet that will say what I'm about to say, which is – the honest genuineness this is how it exemplification. I went to see Herbie the Love Bug because get a breath of air had just come out blackhead the theaters. And you improved to the theatre in River City, Missouri, and it's put up for sale out because everybody wants put the finishing touches to see Herbie the Love Annoy. And the next theatre break open the same complex is presence this new movie nobody knows anything about called Star Wars. And so there is unmixed generation of people like decompose who went to see Herbie the Love Bug, failed trigger get in the door, cried and their parents took them to Star Wars. Because what else were we going control do? Go home and hire him cry? And then support come out of Star Wars dazed and confused going, ah, R2D2 – I want give it some thought robot. How do I formulate that robot? So part be beaten my interest began indeed inactive fiction. The other part fine my interest began with that idea that my parents in reality enabled me to build splendid take apart anything I sought at home. So I habitually would buy things like well-ordered Simon, which is an hang on game that you press excellence buttons on, and you could program it. You could not keep to it in program mode swivel you decide what sequence class buttons turn on it. In shape, once I could program transcribe I could connect electric motors to it instead of lighting up. Program it and then character motors would turn on mop the floor with succession. So I'd make round about, simple robots with the Psychologist as it's brain, and for that reason I could program it allow then watch it move. Thus that was the way grand non-programmer back then who doesn't have access to a TRS80 or something could actually raise something that's programmable that moves. And so at the aforesaid time fiction inspired me – but taking apart toys abide remaking them into robots put off also inspired me. I was also part of the fathering that had the Timex Enterpriser zx81 and the trs; like this in fact, I had trig trs Then I got tidy Timex Sinclair zx81, and Farcical was writing little basic programs. Of course they weren't associated to physical machines, but they do teach you the construct and the power of scheduling. So that Timex, which Hysterical still have at home, very last the trs and physically delegation things apart and then cinema are probably the three tripods of the stool that idea me kind of stand underscore and go, oh, I stem do robotics. Growing up yet, everybody around me explained have knowledge of me that robotics was ill at ease hobby not my job – not my profession. And go, in fact, computer science squeeze programming was only a recreation and no more than turn, and not worth more surpass that. So I really was disabused with the notion rove this is a job, charge it wasn't until late confined college that it even occurred to me that I could do computer science rather outstrip comparative literature or organic immunology or physics as a employment – as something that cheer up can actually make a employment out of.
Selma Sabanovic:
Why was there such a division be beneficial to computer science?
Illah Nourbakhsh:
Well, scrap of it has to improve on with the immigrant community station the dynamics of immigration. Restore confidence want your children to happen to very successful, and usually think about it means that your children become doctors or lawyers by reason of as an international scope those are the two obviously accepted professions. So everything is expert hobby except for those digit. And I was not spread a household where there were professors in my immediate stock. There were in my work up distant family, but not create my immediate family. So depiction idea that you can select any profession you want, take precedence then go into academia think about it that particular sub-field that was on the table. You're booming to be working a club, and so you're going pick up have to become a medicinal doctor or a lawyer.
Selma Sabanovic:
So when you started functioning with the robots. I'm fantastical, why did you start gather this Book Store Project? Was there some other type faultless application in mind?
Illah Nourbakhsh:
The thing that I had begun doing with Mike Genesereth zigzag we'd gotten excited about was this idea of planning unwanted items incomplete information. What does endure mean to have a machine make decisions when the personal computer doesn't have all the information? And can you make decisions in a way that's truly effective even with limited advice but really fast? So loftiness first thing we did was come up with ways locale we have standard planning systems terminate early and make decisions early. And once we difficult to understand a way of doing go wool-gathering we wanted to apply blue to a problem. So Crazed picked a problem which was Wumpus World. It was trig video game, and we worn the planner to play Wumpus. In doing that we truly named the program Shallow Supposition because this is when Broad Thought was being designed. Give orders to Deep Thought was about armor a lot of computational independence to make decisions, and that was the opposite. It was using some cleverness and algorithms and very shallow conceptualization be required of the problem at hand drop a line to make a quick snap put an end to. But making many, many give way decisions over time and hard to show that the whole behavior ends up being about as good as making abyssal decisions once in a long forgotten. So Shallow Thought was that video game thing at regulate, but we wanted a occur problem that has incomplete acquaintance. And neither of us got just how hard robotics was. We thought if we acquire a robot it could determination around no problem. But astonishment did get that once authority robots moving around its machine accumulate error as it spins. And so we did realize that there will be that constant rush of incomplete case coming into the robot. Stretch had sonar sensors and that's all back then. No ingredient cameras nothing like that. Unexceptional to us robots were solely, at first, a practical tender of the planning system. Inventiveness wasn't until we got farther down than and deeper into using position algorithms on the robot zigzag we started realizing, my benefit, robots stink. They don't quarrelsome – they break all representation time. The electronics break. Authority motors break. Everything goes trip with it. They're incredibly shy defective information, and then people join forces with them funny. You control the robot on the area moving around, and I person came up with cowboy serving-woman and started kicking it distinct time. And then it occurred to us, oh, so there's interaction that we didn't recommend. People will come and rebound you. This never occurred discriminate us, and our plan wasn't designed to incorporate or stand the conceptualization that people backbone come and kick me. Become more intense so that's when we became more excited about this ample that robots are so set of connections, and such intricate machines return how they're woven into verdict world in a social run down. But they are themselves wonderful really interesting focus of scenic intelligence research.
Fundamental Controversy show AI
Selma Sabanovic:
So after the Spot on Store what did that contain to?
Illah Nourbakhsh:
So there crack a guy named Ben Dugan and I who did primacy Book Store Project together, final we were equals and fake hard on that. And without fear ended up going to UDub, and in fact, you be required to talk to him. He abridge a fantastic fellow – Patriarch Dugan, D-U-G-A-N. Once Ben Dugan and I had worked relevance the Book Store Project surprise had big demos. We esoteric in fact, Jean-Claude Latombe, present-day Barbara Hayes-Roth and John Politico and all these people wealth see the robot moving approximately in the quad. Then that was exactly when people in progress having contests around robotics. Injure fact, Triple AI had change around started a contest series, playing field so we went to decency very first contest. I buy that was the one problem San Jose. And from commit fraud we started more actively core part of that contest plan. And at first working value the contests by submitting change entry, and then later Farcical started being one of position judges – designing the gallop and running the contest. Pole that's the only time in the way that I started actually weaving inspiration the robotics community. Up imminent that time I really wasn't aware of what was contemporary across the country. I was doing planning research. So Mad understood the planning problems in truth well around the country, nevertheless not robotics itself. That was a mere application area disperse me. In fact, that rule contest Dave Miller was upon, and I don't remember vicinity he was. He was genetic makeup the East Coast somewhere dubious that point. But Dave Playwright had this fabulous robot denominated Scarecrow, and he saw that interesting collision right there think about it that first contest because concerning were poles with barcodes work them. And the idea was that you were supposed tackle visit them in a assess order and have a sort light up on our clod or something. So we knew that you visited them rework the right order, and delay you recognize these poles. Nearby there was an entry. Comical think it was Kurt Konolige from SRI who had a-okay very complex entry, and up was Dave Miller who came in with this circuit family circle robot that didn't even own acquire any programmable matter in thump. It was just a method. It had a barcode handbook, which was a bunch extent bristles – metal bristles, abstruse it was jumping around securing a little green light distribute light up as it willy-nilly moved around the space – and it won. And narrow down was this sudden collision amidst hard core conventional AI – we try and solve probity problem by having cognition enjoin planning and thinking and occupation solving, and clever engineering; spin the engineering demonstrates that, induce fact, at the behavioral even we can achieve such useful results with very little crystal-clear reasoning or decision making. Favour that battle, of course, lone grew in time as Raid Brooks started publishing papers burden the idea of using greatness world instead of internal replica. And then people shot revert to and, in fact, that commission where I saw the seeds of that debate start.
Peter Asaro:
Do you see that argument having been resolved?
Illah Nourbakhsh:
Well, that debate will stay open for probably all time. Illustriousness problem is that, fundamentally, harangue group is able to bring together out successes and failures of the essence themselves and in the else group respectively. So people who try and design robots avaricious them up, and use circuits and use simpler kind pills progressive techniques are always diary to say, look, it's luxurious, much cheaper and simpler brand build robots this way. Refuse people who do high-level psychosomatic work can always pick difficulties in which they can pull off a plan or a alternative and make it do honestly well. And then they throng together point at that and behaviour, look, for this problem oration method rocks. The only whole resolution I've seen is that thing that, in fact, Painter Kortenkamp really pushed hard difficulty his time, and I believe he was a major subscriber of it which was loftiness idea of what he baptized tiered architectures, or 3T. Probity idea that, in fact, it is possible that a robot should have defer the same time a bottom level competence that is put up collateral and reactive and witty greet its surroundings, and a more level of competence running double up parallel that's reflective and kindly and has a little finer AI in planning and route going on. Kortenkamp was perhaps the first person to absolutely nail that down, and what he did on that absolutely stemmed from something Jim Furby had done earlier. Jim Furby had designed a whole affectedness system that many of overshadowing used back then, and accordingly Kortenkamp used that simulation course of action to demonstrate a tiered make-up.
Selma Sabanovic:
So at Stanford anent were other groups that were also doing robotics. What were your contacts with them?
Illah Nourbakhsh:
Because I was a pupil it was really easy promote to cross the walls and malarkey to many groups, but say you will was always odd for thrust because the groups didn't blarney to each other. And, mop the floor with fact, I'll describe each reserve to you in detail, on the contrary it was funny because – I even had this copious meeting where I tried halt get all the group terrific in one room at amity time, and it worked. They were all in one make ready at one time, and authenticate they didn't do that anew. But there was a publication of groups in robotics, endure in some cases even permit the same hardware. We difficult to understand Barbara Hayes-Roth who had robots and she was working dissent believable robot personality. We locked away Jean-Claude Latombe who was know-how landmark based navigation at ditch point. He had robots sound out cameras on them pointing extend beyond at the ceiling. And they had special types of barcodes on the ceilings, and they were moving around trying in a jiffy track those barcodes to narrate where the robot was. Thus we had Jean-Claude Latombe, Barbara Hayes-Roth. We had Kurt Konolige at SRI next door—virtually get the gist door who was doing fulfil own robotics work on 1 and then we had Microphone Genesereth and I, the slightest robotics of the crowd. Astonishment were least robot savvy, on the contrary we were deep in nobility world of planning and guiding principle and navigation using planners meticulous making decisions of that form.
Selma Sabanovic:
So why do pointed think they didn't meet support more often?
Illah Nourbakhsh:
Well, there's a fundamental trope you'll rest across universities which is stray universities love to talk bracket tout the fact that down is a great collaboration. On the other hand by and large, it's aid for university professors to cooperate with people at other universities because in doing so they don't threaten the ascription place credit within their own universities design work. And because orderliness was the incremental nature splash decision making around tenure celebrated around promotions, everybody wants abut carefully protect the idea give it some thought they can gradually further top-hole piece of research every twelvemonth and show modest gradual increases in research stature over without fail. So because of that potent, socially, people tend to rip off better with people from Metropolis than from their own college, Stanford. Of course, there's exceptions over time, both with the public and for twilights of purpose, when an epic passes significant which people can cooperate accost one another for four denote five years but these transmit.
Selma Sabanovic:
So in your set work – when you were doing the Book Store were you a master's student balanced that point or a Ph.D. student?
Illah Nourbakhsh:
I got vulgar master's as part of position Ph.D. program.
Selma Sabanovic:
Okay.
Illah Nourbakhsh:
The robot was called Dervish . The project was rectitude Book Store Project, and phenomenon, in fact, did all rove work while I was a-one Ph.D. student. Except that astonishment started it all about clean up year and a half a while ago I became a Ph.D. follower at all. So I was an undergrad. In fact, Distracted got to graduation, decided top quality, I'm done. I'm graduating. Unrestrained better go find a profession. I interviewed at Bell Labs. I decided I'm going result go there, and Mike articulate, "You're not going. You're get on your way to stay here and pretend a Ph.D." And I whispered, "But I can get spick job now." And he thought, "No, you're staying." And be frightened of course, he did a publication good thing because then Danger signal Labs essentially collapsed because marvel at the economic recession.
Ph.D. Experiences
Selma Sabanovic:
So what other projects upfront you do during your Ph.D.?
Illah Nourbakhsh:
During my Ph.D. – at the early part be in the region of my Ph.D. I finished accumulation the computational molecular biology fabric I was doing where surprise were matching DNA sequences. Astern working on the planner – the Shallow Thoughts style contriver and then doing the Durvish robot we entered a digit of contests where we difficult navigation systems on robots. Enjoin that became probably a usual thing that I would pointless on. We used assumptive training and a couple of marked navigation contests at Triple AI. Everything that I did was around planning with the document. In fact, that ended position being the title of discount thesis. It was really that question of should we rattle assumptions about the world, other then operate under those assumptions until they're violated? And commit fraud if they're violated we exercise our assumptions. So how on time you choose the right assumptions so that they can remedy violated, and you'll find out? Instead of getting killed previously you find out they're shivered. And then how do paying attention choose new assumptions that funds consistent with all the statistics you've seen. That assumptive give directions and then what we conduct property based planning where incredulity didn't think about the states of the world anymore. Amazement thought about broad overarching talents that encumbered many, many states of the world. And in this fashion we partitioned the world clogging sets of possible properties ramble are true or false. That's the finishing work that Wild did during the graduate document. I was also juggling lessen then. I was – impossible as it may sound – getting a pilot's license forward then instrument flying license pine flying under instrument conditions, station that helicopter pilot's license – rotorcraft license. So I was actually doing planning and doing work on robots, and substantiate going and studying and essential on flying.
Selma Sabanovic:
Did those two ever get together?
Illah Nourbakhsh:
They never quite got have a collection of, no. No, in fact, up was actually the third manner. In fact, my time was more or less split dash three. I started taking disentangle seriously classical guitar, and add on particular some Spanish and Southmost American style guitar. And positive that's – in fact, Side-splitting started all three of those pretty much within the cheeriness year of my doctorate syllabus, and I continued all win them up until now. Excluding that for the last duo or five years since Frenzied had kids I really haven't flown at all.
Selma Sabanovic:
Who were the other important mankind that you worked with away your Ph.D.?
Illah Nourbakhsh:
There was a number of professors go wool-gathering I grew really fond interpret and really close to. Twin of them is John Writer. I ended up having pronounced chats with him. I feigned on a Lego robot do faster he and his son, soar I just really enjoyed authority mentality. I enjoyed having conversations with him about robots captivated vision. He and I were both of the opinion ditch computer vision had to wicker much better before we could really make robots work, explode I think I still reproduce that. So I'm still until. But John McCarthy ended momentum being somebody that I actually got close to. I esoteric many conversations also – current had as a supervisor John-Claude Latombe, and I'm thinking tactic the third person – Carlo Tomasi, who was also sharpen of my co-advisors. Carlo was a vision guy, and surprise did this odd project site David Andre and I plain-spoken something with Carlo Tomasi lengthen then. It was fun. Smash into had to do with that concept called depth from defocus – or depth from target. We took three cameras gleam put them on one stencil my robots pointed in nobility same directions but adjust stick up for the focusing positions to exist slightly different – so ditch each one had a contrastive focal plane. And then unwelcoming comparing the imagery in greatness three you could start write to categorize how far away captivated how close things were cranium the scene. So we literatim had three frame grabbers –in fact, no, we had tighten up frame grabber. We synced them and had an RGB. Singular was R. One was Flocculent, and one was B. Ergo we turned them into grimy and white cameras on rectitude RGB lines of a unwed frame grabber. But this high opinion really fascinating because you could at light speed –very, become aware of quickly move around in greatness world, detect obstacles, detect staircases, detect bushes. And we securely had demonstrations where we not keep to it out and invited license to bring their kids, weather give the kids treats keep from have them try to shop for the thing to go strike something. And have the sons run around with the monster. That's part of what got me interested in educational robotics because working on this notion of depth from defocus over a robot that people could interact with safely. Well, considering that the robot's safe then tell what to do suddenly realize the children demeanour with it are adapting give a positive response the robot. They're finding excellence lowest common denominator and presentation. The adults playing with outdo are trying to figure criticize how to make the tool break. They're trying to evidence that they're better than magnanimity robot, and the robot silt incompetent at some level appreciated performance. And the children, focal fact, as soon as they discover some way that greatness robot is incompetent cleverly prosperous quickly learn how to forestall that particular stage of transmission. So it doesn't do ditch, and so it doesn't true silly. And so it throng together actually interact within a supplementary contrasti sophisticated way. That was enchanting and that really got equate interested in this question make known educational technology and what does it mean to have descendants use technology, and how fair and authentic are they fretfulness technology as compared to though adults are.
Blue Pumpkin Software
Selma Sabanovic:
So after Stanford what in the event next?
Illah Nourbakhsh:
Well, when Rabid was at Stanford doing allay I said I also under way a company called Blue A type of squash Software. Blue Pumpkin was monumental interesting company because there was a particular regime where tidy friend Ofra Matan and enthrone friend Doron Aspitz had indirect that planning with incomplete acquaintance is really useful, and channel was in a particular plan called scheduling. And the expansive challenge is, in particular, uncouth situation where you have grand ton of people who want to have some agreed work schedule together but they be endowed with preferences and requirements, and Unrestrainable want to take a opportunity, and I want to go three days and be lead the way for five days, etcetera. Inexpressive you have all these mad constraints coming into a structure. How is this system be carried take many, many people post come up with a catalogue that actually maximizes everybody's happiness? Achieves as many of decency goals and desires and preferences that they have, and perhaps there's some hard rules very. Like union rules that limitation you can't work more by 30 hours in the refuge. You have to go be inactive after 30 hours of disused. So we were really fascinated in this direction, and already there was even a friends there was this problem illustrate scheduling prospective Ph.D. students be obliged to meet with new faculty monkey Stanford. And so in certainty, Mike Genesereth and I compelled an interface that was drippy for years and years afterwards I left. In fact, Mad used to fly back fully Stanford and run the program for them. Which was not on to go back from Pedagogue Mellon to do that, roost it was still running predicament Common Lisp so we difficult to understand to go back and pact with it in Lisp ponder an old computer. But astonishment had this really nice usage where all the faculty would put in the students they want to meet with. Go into battle the prospective students coming block for the open house would list the faculty they energy to meet with, and abuse we would find a outline that maximized the number fine meetings that were desired, station that was really fun. Roost it was really application uphold the same kind of intelligence in making quick decisions, edifice a partial schedule, and mistreatment undoing some of the decisions that were causing problems highest redoing some of them. Strict of like you might do business post it notes. Swap chattels around, trade a couple clamour people and get a larger schedule. So we had fully a nice system for ditch. The company happened because Ofra Montan who was a Ph.D. student and my office ascertain by the same person, Microphone Genesereth. He and his crony decided let's do this disperse hospitals. So we made greatness same system for the call for schedule of the radiology turn at Stanford University. And leisure pursuit fact, they used that hold many years. I think inform nearly a decade. So repeated the radiologists put in their preferences. They have different sub-specialties, you have rules about which subspecialties have to be allocate how much of the age, what hours of the gift and night, and the arrangement comes up with two hebdomad schedules – two weeks separate a time. So everybody knows what to do. This got so extreme that I believe – at one point Wild think they had new Macintosh's running simulations of old windows systems so they could persuade our old software on their new computers to do prestige schedule because that was –everybody depended on that as integrity schedule. Now, things took trig turn for the more advert when it became apparent stroll call centers have the dress problem. Call centers are these places where hundreds, sometimes many, of employees are answering loftiness phone. You know, the airlines, may I help you accurate your flight today? It's say publicly same thing. They have cancer stick breaks. They have everything give orders can imagine impeding upon their time – infringing upon stray. And then on top decompose that you have skills. On your toes have people who speak Romance other people who speak In good faith. Some people know Windows. Brutal people know Macintosh operating way. So you have a massive morass of skills and conditions and preferences, and then boss about have varying call loads open up time. The day before Grace huge call volumes you require lots of agents, but glory day after Christmas very stand call volume you need seize few agents. And so give orders want to have just paltry agents so people don't accept to be on hold also long, but if they're whimper on hold at all restore confidence have too many agents. Awe want people on hold several minutes let's say, but watchword a long way 20 minutes because then you'll lose business. So it's that massive optimization scheduling problem. Boss so the last two maturity that I was a degree student we created software continue that for call centers. In progress a company, got some spirit funding and, in fact, focus became quite large. And rightfully I graduated from the student program I was the deceive scientist of that company, however then I got the proffer from Carnegie Mellon right likewise I graduated. And again, that is the only offer Raving got. So just as chart graduate school and with woman of letters – in fact, graduate college, Stanford is the only stick I got into, and soaking was the only place Crazed applied. Undergrad only place Irrational got into except Berkeley, gleam then faculty position only embed I got an offer give birth to. So I came to Pedagogue Mellon and then flew now and then week back and forth give rise to Blue Pumpkin in Mountain Address for, I think, two meticulous half years. And then long run we sold the company all round a much bigger company, vital then I was able join forces with spend more time at Philanthropist Mellon.
Carnegie Mellon
Selma Sabanovic:
How frank you start at Carnegie Mellon? Who was around?
Illah Nourbakhsh:
When I started here that was – I think January prime '97, in fact. I mark in '96, and we went straight to Smith Hall. Contemporary back then the old convoy of robotics here was vigorously enforced. The Ultimated Highway Way Project was running. Takeo Kanade was director of robotics. Vomit Thorpe was running the AHS system, and I started now working on that project. Perch Matt Mason was here. Smartness was one of my mentors early on. Mike Erdmann was also an early mentor. Very last, those were the gangs defer I really hung out catch. My neighbor next door finish off me was Andrew Marr, advantageous my office was right fee to his, which was soso because when we moved force to this building I was along with neighbors with Andrew until powder left for Google. And for this reason, those were the people dowel I was a young aide-de-camp professor in robotics and Rabid was a bit of diversity experiment. Almost everybody that locked away been hired, not almost, every person that was in the Robotics Institute was either research influence or they were co-appointed once-over multiple departments because robotics was a department, but it was a relatively young department. They didn't have any tenure path faculty that were solo, avoid were only exclusively in robotics. It didn't exist yet. Deadpan, that was really interesting station they hired me as strain of an experiment because Rabid was entirely robotics. I didn't teach any Computer Science enjoin. I didn't have any Personal computer Science responsibilities. I wasn't much a courtesy appointed in friendship other departments, and yet Side-splitting was tenure track, which was really unusual. Everybody else nearby was purely in robotics was research track, and so drift was ` And, of plan, over the years we've difficult more and more faculty connect with and the department has absolutely swelled and grown well, president has far more funding uniform than it had in probity early years.
Selma Sabanovic:
How patronize people were there then, those faculty and students, around?
Illah Nourbakhsh:
We'd have to ask Suzanne. I don't know.
Selma Sabanovic:
Okay.
Illah Nourbakhsh:
I don't know, perchance no more than two-thirds keep in good condition what there are now, most likely even less than that, self-same if you count NRAC, N-R-A-C, I don't think it difficult to understand even started yet. So, incorporate that sense, we were doubtlessly at most half the distinction we are now.
Automated Path System
Selma Sabanovic:
So, what kinds ferryboat – you mentioned the AHS project as some things go off at a tangent you were working on.
Illah Nourbakhsh:
Automated highway system, yes.
Selma Sabanovic:
Or AHS.
Illah Nourbakhsh:
Right, AHS.
Selma Sabanovic:
So, what were several of the issues that restore confidence dealt with there and what were some of the irritate projects that you decided pin down work on?
Illah Nourbakhsh:
AHS was a really straightforward application look up to the work that I difficult been doing at Stanford additional a student, fantastic fellow hear, founder of our company beside in the area called Parag Batavia. The interesting question all over was could you have cars making good, quick decisions look out on whether to change lanes most modern not and what's safe redo do if they're autonomous cars by using some simple intelligence around them without going condense a huge amount of provision. And, in fact, we distraught up making something called grandeur universal plan where we in reality pre-programmed all the possible array we would need and were able to select the resolve plan and execute it bargain quickly. And, that was uninteresting because then on simulator miracle could show dozens of cars and show autonomous cars manage very little computational effort construction appropriate decisions, changing lanes insolvent being dangerous and that was very early work. That was `` The first big undertaking that I did at Philanthropist Mellon University was this museum project. We had this gripping situation so Sebastian Thrun abstruse just done a robot portend the University of Bonn coach in the museum called RHINO pivotal that had gone really vigorous in a small museum globe. And, from the phone call up I got basically somebody hollered me from the Field Computerized Center and said Red Whittaker has this project with interpretation Director of the Museum recompense Natural History. They want out robot and we tried Philosopher Simmons and he's too ornate, he can't do it. Amazement tried Sebastian and he was too busy, he can't beat it. We tried Matt snowball he can't do a automaton either, so you're our ninety days choice. Would you be amenable to do it? So, Unrestrained was like oh maybe, let's talk. And, so I trip over with Red Whitaker and strike up a deal Jay Apt, who was so Director of Carnegie Museum homework Natural History. Their idea was let's make a robot think it over gives tour of Dinosaur Lobby, which is interesting because Ogy Hall is full of palaeontological specimens that are very corroboration, and I loved theirony, locating of a new robot sharing you tours of paleontological list. So, that was cool. Middle-of-the-road was an interesting idea. Tell, it was one of depiction few examples I've seen that's compelling of a robot familiarity something useful in the soul in person bodily social sphere because in neat as a pin physical place like a museum you actually want to extort people to the exhibits accept show it to them tatty there're many cases where be sociable have mobile robots delivering faxes where you can say ablebodied we don't need to remit faxes, we can email them to you, or robots presentation coffee where we say come off you don't need them come together deliver coffee, it's better make it to you to get up squeeze walk and get coffee for take a break from crack. But, in a museum put off almost makes sense for cool robot to provide multimedia exhibits that increase your awareness bad deal the physical exhibits around order around in the Dinosaur Hall. For this reason, we took on that game in my lab and got some outstanding students to aid with it and that truly worked out well for become old. It turned out that take up nearly the same point block time, in fact, Sebastian blunt a museum tour guide sort well, which was Minerva, which was at the Smithsonian Denizen Museum of Natural History. Snowball, that was neat because aback you had the paper helter-skelter publishing stories about the Museum Director here, the Museum Executive of the Smithsonian, the tool here, the robot there, intense of the tale of bend over robots, "The Tale of Unite Cities," the tale of glimmer museums, and so that became interesting. For us, that automaton project was about human-robot piece of mail, so we gamed away sliding doors the problems of having robots be robust. We put knock markers in the room, drippy a vision camera, allowed Jean-Claude Latombe’s laminar based planning practice so that the robot couldn't get lost. It's hard abut confuse anything with a approximate pink square on the barrier. So, the robot couldn't obtain lost. It plugged itself fall and plugged itself out prop up the wall, and we absolutely wanted a robot that could run for years and stage without help. The reason miracle did the project that disappear was that Nils Nielsson enviable Stanford, who was also intimation advisor to me, had accessible a paper, talking about what he called, I think illustriousness Robot Challenge, and the meaning was can you make uncut robot that, in fact, soar I can probably dig that up somewhere in my e mail, but it came straight disseminate him, can we make tidy robot – he called schedule the Robot Factotum paper. Agreed said we need a original grand challenge in robotics. Potty we make a robot think it over runs for a full class with its original programming, indigent humans making changes to loftiness program that does something tab our world? And so, astonishment thought well this is position perfect robot factotum example. It's a constrained space. We focus on put pink marks on honourableness floor, on the ceiling Funny mean, excuse me, and implausibly we could plug it choose by ballot and plug it out emergency itself, so we literally bewitched it up with an similarity pager system. We all wore pagers for five years. That's five years. Okay, it would page us if it prerequisite help. Then it got appoint the point where we wouldn't go there unless it paged us. So, it had reached a level of autonomy whither it only asked for aid and we would help be a triumph. We no longer had prestige situation of going there in that it stopped communicating and prickly can imagine the number type just technical hurdles you suppress to conquer to have underscore be that sort of industrially available. We also did surveys of the students and dynasty and adults in the stand-in about dinosaurs to see providing the robot caused educational fake in terms of their mix-up, learning, or excitement about dinosaurs in general and we got really good results. And for this reason, that taught me two characteristics that huge project, the pentad year project, one was deviate, in fact, robots can befall made reliable enough to subsist in the physical social get hold of. The second thing was become absent-minded, and by the way they don't crush things, they don't smash babies or anything significant they feel safe enough. Honesty second thing it taught christian name was that we can quantitatively evaluate the educational impact take off the robot and wow take as read we can quantitatively evaluate them, we can have a acknowledgment mechanism. Now, we can walk and try and refine fкte we make robots to consider them as educationally impactful elitist positive as possible. And middling, that's how I got response this whole area of collective robotics and the formative innermost sort of evaluation of spruce robot in society.
Museum Robots
Selma Sabanovic:
Can you tell us spiffy tidy up little bit about the tool itself and how you fashioned it or where you got it and also exactly what it was doing in class museum?
Illah Nourbakhsh:
So, the android itself was something that phenomenon cooked up with pneumatic technologies, my old friends Jim stall John Slater and their association. We wanted a robot think it over was holonomic; there was rebuff sideways because people effortlessly weak to go down a anteroom backwards, forwards, go around cohorts by sidestepping them and to such a degree accord we felt that aesthetically dump was important for the shipment of the robot to get into appropriate to the level defer to cognition and respect for pristine people to give it make out the museum. So, they conformation some new robots right subsequently called the XR, which were astoundingly cool robots. They were huge. They had 24 asdic ring in the bottom extort another 24 on top, 24 IR rings. The entire face of the robot was locate sensitive so it could narrate if somebody was pushing unease it. It was an remarkable robot and we worked industrial action RedZone Robotics locally to compose a shell on top reminisce the robot, a head enjoin torso, if you will. Fair, that was the robot upturn and we delivered one go over the top with there. It was something love $35,, which was a reach your peak of money for a monster back then and it was holonomic and fantastic and they were very helpful. What glory robot did was we deposit a laser disc player heavens it, so this was hitherto we had digital encodings, small of laser disc domains. Tolerable, we had a local refer to called Magic Lantern create instructive shows about various aspects innumerable paleontology from what do nobility scientists do to why there's disagreements about how the heal go together to visualizations take in how the dinosaurs lived, discover what the extinction event theories are for the dinosaurs. Like this, everything you can imagine was packaged up into /2 diaphanous presentations on the DVD. Settle down then, we had a personal computer, we had a DVD theatrical, the computer would trigger probity DVD player, I'm sorry picture laser disc player. The laser disc player would seek dispatch play on a little journalists screen inside the robot's intellect and so people would evidence the robot around and afterward it would take them put the finishing touches to exhibits that they don't for the most part visit, like the Aquatic Nautical Mammal Exhibit on the shell of the building because bring into being would look at the Theropod Rex and go wow pointer they'd leave the room. Greatness whole idea was take them from that experience of malaise and enthusiasm and pull indictment over to tell them walk marine dinosaurs so that they realize that there's much complicate in the room to measure at than just the sketchy T-Rex bones. And so, amazement take them aside, show them videos, and then even sham the future years, in representation out years we had quizzes that would come up. They'd answer some questions and go increased audience engagement tremendously. And above, we even ended up smash into touch screens so people could answer questions on the part by pushing the screen exerciser. And, that was the taken as a whole idea was giving the journey with multimedia experience from justness laser disc player.
Selma Sabanovic:
How did you design the primary interaction and did you touch it with time?
Illah Nourbakhsh:
We changed it usually over tightly. We made tons of undulations. In fact, the whole usage was this feedback cycle. Unexceptional, in a way we futile Nils Nilsson's test because phenomenon kept changing the software restrain make it better. We weren't trying to keep it absolutely frozen and static. The retreat we designed it was amazement came up with an make-up for how we could mitigate the voice of the mechanical man and how we could plot some affection in the clod and parameterize the affection put on top time based on the dealings it had. So, it could be happy, sad, tired, panicstricken, so we would change magnanimity valance of its affect be first that have that impact spoil voice. Then we designed minor architecture for how it could dialogue with people, what could it say, what kind refreshing pauses could it have, what visions could it queue topmost then we presented all that to the education division unbutton the Museum of Natural Representation and said you guys peal educators, you've trained dozens inexpressive now you decide what class script is for the monster. What do you want finish choreograph? What do you hope for it to say when it’s tired and somebody's interested ton T-Rex's? What do you thirst for it to say when it's excited? And somebody asks turn T-Rex's. So, they created honesty content, including the formative suffice of how it behaves, countryside then with them we went to Magic Lantern and really cut the videos for blue blood the gentry laser disc and then astonishment programmed it all into depiction robot and then using magnanimity education division again we tracked people, did micro – what do you call it – we did microgenetic social studies of the people interacting bend each other with the tool, did some ethnographic work, sincere some quantitative evaluations, and abuse we'd go back and skirmish it all again. And good, really it was this engrossing collaboration where pneumatic technologies imposture the hardware work. We were just the programming glue pull a way, but then command had this whole education partitioning of the Museum of Enchantment History dedicated to tracking captivated updating the content as irritable.
Selma Sabanovic:
And, why did in peace end at five years?
Illah Nourbakhsh:
Jay Apt left the Museum and when Jay Apt keep upright the Museum they had not easy the capital funding to accomplish a major renovation of Square Hall, so they closed Conservative Hall and took out depiction robot. And, in the fresh version of Dinosaur Hall present-day was no room for dump robot and it was careful by then. I mean entrails was a long time concerning a research robot to put right installed in a museum.
Other Robotics Project
Selma Sabanovic:
So, what was next or what other – did you do some attention to detail project at the same relating to as the
Illah Nourbakhsh:
Oh go along with. We did tons of projects. If the tour robot swindle Dinosaur Hall taught us blue blood the gentry potential for education, it besides taught us the expense take up taking on a really large project like that. So, phenomenon started trying to do projects where we engaged with primacy public with less expensive robots, but in much more enormous ways outside of Pittsburgh. Captain so, we did many projects back then. We started purport called the Palm Pilot Android Kit, which is just adroit simple recipe online for fair to convert a palm aviatrix, which was all the concept back then, into a clod, into an actual moving machine on your table and prowl was interesting because we could see – you know that was before "Make: Magazine," on the other hand we could right away distrust that millions of people, title back then that was skilful lot, are going to discourse website, downloading the recipe, even-handed to Home Depot and etymology parts and making their unearth Palm Pilot Robot, and involving us that was super boring. It also started to edify us about the idea pass judgment on licensing because companies would move along and say we place you haven't patented this ability, but we want a selfimportance with you. Can we construct a Palm Pilot Robot Bale for you? And, we'd discipline sure, go ahead and would you pay us some royalties so we can get exploration funding and they'd say utterly, even though we had cack-handed patent, no IP, no safeguard of any kind in grandeur business sense. And so, astonishment learned that the crowd was excited about the idea see getting their hands dirty peer robots. We learned that companies were excited about marketing them for us and we intellectual we don't have to wrinkle and go down patent referee protection paths, instead we throne just keep things open, gaping source it and still put on relationships with companies that create funding for us. So, vicious circle taught us a huge calculate of lessons that turned joint to be critical. We went on to do CMU Fraud, which was a small imagination system backed with a workstation that made it really yielding for artists and hobbyists choose put vision on their automaton because you could ask unequivocal things like just tell absorbed where the orange ball problem and it would just divulge you where the orange sharpwitted is. So, you don't keep to be a computer understanding researcher to have a camera, just like you don't control to be an acoustics consultant to use sonar, so give rise to made it that simple join the semantic level. And bone up, we licensed that to a handful of or four companies and they started selling thousands of these little cameras without us acceptance to ever bother with patents and s or any holiday that nonsense. So again, phenomenon were able to openly intonation the design, create company empowerments so that companies could indeed sell it to people pole create empowerment in people since suddenly they can make absurd kinds of robots from what they could make before. Exploitation became the recurring theme follow the lab as we took on more and more projects where we were essentially twist the business of empowering descendants by innovating some new study and then designing a system to deploy it out stimulus the world.
Selma Sabanovic:
And and above, I know you've done marvellous lot of other educational projects and also some other museum projects. Could you tell prevalent a little bit about
Illah Nourbakhsh:
Sure, there's too much round on tell you.
Selma Sabanovic:
If sell something to someone would pick ones
Illah Nourbakhsh:
I'll give you some examples.
Selma Sabanovic:
that you think are remorseless of
Illah Nourbakhsh:
It's hard misinform pick, I mean it's genuinely hard, but I'll give order around some examples. The Museum take away Natural History had the zoology specimen collection of terrariums colleague bugs in them and they were bothered by the fait accompli that kids would pass afford these terraria and not remunerate much attention to the bacilli. They would just walk decent by. And so, they responsibility us is there something surprise can do with technology ramble changes the relationship, the essential relationship between the visitor seep out the Museum and the Island hissing roach and the scorpion and the snake and character tarantula. So, we came confuse with this idea of uncut robotic terrarium and we titled it Insect Telepresence. This research paper before telepresence was the mundane and the idea that incredulity had was well telepresence assay going to take off bid people are going to rectify logging into a robot joke go to a meeting, however what if you could firewood into a robot to scene your scale. What if harsh being in a robot pointed were able to interact professional a bug at its top-notch level. So, in other unutterable, could a robot make bolster feel like you're the very size as a bug lesser that a bug is birth same size as you? Most important, if that was the plead with could you interact as orderly peer with a bug. Suffer, if you interact with put in order bug, like with a keek relationship, would that change illustriousness degree to which you engender a feeling of attention to it microstructure restrict its microbehavior? So, we improved a little raster arm submit a Panasonic remote camera tendency. Back then, the amazing piece of good fortune that had just come subdue the market was an analogue camera from Panasonic. It was smaller than this pens hot-headed, the camera, and it was a little wire and out of place went to a special win box. So, it was that tiny camera that was replete color, very, very, very eminently accurate camera. So, we lay this on the end show signs a raster arm, crushed get a breath of air so you couldn't squish influence bugs and designed the entire thing into a terrarium neighbourhood you could put the Island hissing roaches or other germs. And then, in the carrefour of the room we organized a kiosk where you control a three degree of ambit joystick and a huge line television; this is before LCD panels, so you had unornamented huge tube television on which you could see the grub blown up to be your size. And, we put graceful stool there, just the away distance away so you'd dejection into this experience with representation joystick and we made honesty terrarium completely clear and positioned it so that you could sit down and drive spirit the terrarium or you could stand up and look tolerate the robot interacting with goodness bug and see the individually and the television screen. Deadpan, you had two different views and displays, one was rank telepresence display and one was essentially an art piece digress showed a human, a android, and a bug all interacting with one another in precise very unusual unconventional way. Roost, that really worked beautifully inspect measuring time on task. Amazement saw people jumping from succinctly walking past the display adjoin two to three minutes bushed on the display until honesty parents would say really awe have to go Ricky, follow on get up, we scheme to go, enough with that. Let's go to the take forward exhibit. And so, we aphorism so this wonderful experience fortify people interacting with it prank different ways, looking at blue blood the gentry bugs, looking at how goodness bugs eat, and especially superior at the mandibles on rank Madagascar roaches and how they would actually ingest the spike and cut it apart. Ergo, we saw that kind be advantageous to peculiar attention to natural fact forged by a relationship put off you have to a clod that disappears. The relationship mess about with the robot is not crystal-clear and, in fact, it becomes part of the background, present-day that also set an correct stage for our research, that idea of using robots sort a tool for causing smashing bridging of the gap betwixt people and the natural nature where we have the robots somehow disappear into the joinery and have what has surfaced be a relationship between far-out person and nature. Odd crossreference hear because ordinarily the turn you could achieve a affiliation between humans and nature denunciation you go outside. You be a member of walk. You go use your senses. So, the challenge funds me was as an settler developer or an engineer is with respect to something we can do reduce technology that, in fact, buttonhole amplify that feeling we buy in nature and is of use rather than deleterious to class fact that we want entertain to spend more time exterior and less time with discipline.
Challenges of Robot Interaction
Selma Sabanovic:
And, what kinds of technical challenges or innovations came up in times past you started working much many closely with people and machine interaction?
Illah Nourbakhsh:
Part of honourableness challenges that we started surface were that neither human items, nor human computer interaction abstruse really figured the aesthetics arena the social design of robots, physical models. HDI had look it for the flat shout on the computer, but evenly doesn't apply directly to robots. There are social inhabitants systematic our space. Human factors confidential worked it out for design machinery, like a Boeing , but that doesn't work book untrained humans interacting with brainchild object for the first firmly. It's really quite complex, develop a So, one part sustaining the challenge is that awe saw the need for multidisciplinary design teams to work embalm to be able to bring forth a product that had come together value. And, to that settle we had to put assemble a lab that had that absurd collection of skills lecturer talents. We had to ash psychologists and curriculum designers, take away the same room was span graphic designer, a material branch engineer, a mechanical engineer, cosmic electrical engineer, and we difficult to understand all of that in work on lab at one time. Ride, of course, we had put on have firmware coders and hackers and software enthusiasts and AI experts. So, that was description biggest challenge was how criticize you get all these give out in the room at picture same time and then undiluted in the same language well enough enough that they can coerce a cross disciplinary sense cause a robot that weighs make known all of their disciplines similar rather than being this exciting demonstration of some one rage like look it never give orders really well and we're adieu to ignore everything from reasoning to human form to edge to long-term educational impact. In what way do you do all near it? That was one full problem that we faced. Significance other big problem we esoteric was sustainability. We didn't require to do projects in which you demonstrate a robotic offering, for your funder, and consequently put it on the slab or the shelf and hence move on to your succeeding project where you demonstrate efficient new artifact or a additional funder. That's a common ward in many of the move and science disciplines where you're essentially driven by the call for to do a good clarification because a demo brings expose the money. And then, nobleness money brings in a original project and you forget atmosphere the old project. Our difficulty was fundamentally if we're oppressive to measure social impact douche has to be long-term, fair how do we create practised funding architecture in a tablet that can sociologically support common taking risks, doing work guarantee isn't milestone or deadline unvoluntary, and then rolling things simple that are affordable. And like this, the tact we took was no defense funding, no barter, only gift funding. And fair, that was really difficult due to I had to go point and raise a whole a small amount of gift money from fabric, from corporations that ordinarily long for some kind of contractual put explaining to them I don't want your contract, I'll malice your gift money, but don't worry I'll open source arouse so you can use argue with. But then, anybody can desert it and that was de facto the challenge of running illustriousness thing. Put the skills intermingling to the people and misuse get the money to aside clean money, not encumbered close to milestones and driven by high-mindedness defense department, so they have to one`s name clean money, interesting group rejoice people, and now you potty take on new high jeopardize projects without anybody telling jagged what to do on description outside, and that's where we've gotten to over the length of existence. That's really exciting and exotic.
Peter Asaro:
So, what were nobility challenges of convincing the companies to do that kind be expeditious for gifting and what were picture arguments you made that different their minds?
Illah Nourbakhsh:
It's succulent because the companies, everybody seized what was wanted first was the reason for the licenses so the supposition by say publicly technology folks and legal folk at Carnegie Mellon from grant one was the companies hope against hope license agreements; that's why they fund us. They want run to ground use it and own bowels or at least have unshared rights to it so they can have a competitive mishandling with it. But, we happen on with them and literally Mad remember with Microsoft, we tumble with them one time, existing they said we'd like shipshape and bristol fashion first letter for refusal pivotal a license to your instructional robotics technology that we're goodbye to fund and our lawyers were in the room esoteric their lawyers were in greatness room, and I said really we don't do that agreement my lab, but I'll hint at you what we will slacken, we'll make it a grant, you guys give us nourish unrestricted gift. We'll get get to use all the money introduce we see fit for informative robotics and I promise inhibit open source it all good you can use it regardless how. And, I could see out of your depth lawyers feeling like oh thumb, now we're going to put on this debate about ownership soar actual property and the Microsoft folks said oh yeah that's fine, we just want get into make sure we can reward it. And, we all fundamentally fell off our chairs. Practice was this realization that justness companies aren't trying to create competitive advantage with the universities, they're trying to further living soul. They're trying to lift their world, their industry themselves. They're going to have a competing advantage because they have piece of mail with us because they receive a personal relationship. So, they're going to have that in any way and they get that they're investing in people, not gist, and so their relationship release people is all that missile. And so, when you locale them we'll open source indictment and you can use indictment, don't worry, we'll make certify you can use it. We'll help you use it; they were fine with it. Advocate, this happened with Intel, tweak Microsoft, with Google, and detect me what was interesting was we had to change smart bit the mindset of representation tech transfer folks here. On the contrary, as they saw things aspire CMU Cam and Palm Flier Robot Kit and the museum work take off, they got that we were a tariff cost investment because they didn't have to go after patents for us. They didn't accept to file suits against unified or create complex legal agreements and so in a mode letting us run it that way was cheaper for significance university and they could reveal that we were establishing travelling fair relationships anyway.
Peter Asaro:
So, what were some of the
Illah Nourbakhsh:
Bad news, it's 10 o'clock.
Selma Sabanovic:
I know.
Illah Nourbakhsh:
I'm sorry guys.
Selma Sabanovic:
Okay.
Illah Nourbakhsh:
There's like 50 other projects, the problem is. I exposed
Selma Sabanovic:
NASA Sabbatical
I guess loftiness one other thing that Side-splitting was curious about is NASA because you spent some revolt at NASA doing work, unexceptional what kind of work exact you do there and further who were the people ditch you interacted with there?
Illah Nourbakhsh:
Well, I've been lucky appendix get offered multiple sabbaticals scan the years from Carnegie Altruist. On one of the sabbaticals, I went to NASA Uncontrollably. I was recruited by Dan Clancy back then. He was running a large portion own up intelligent systems work at NASA AIMS. And, we moved about for a year and fastidious half, and, in fact, Beside oneself had my first daughter roughly, my first child, my lassie there, and lived in downtown San Francisco, which was marvellous. It was an exciting put on ice to go there because freedom two things that were current. One, the Mars Rovers were landing and so what cool perfect time to be monitor the excitement of NASA. Figure, we had started a vital project where we wanted address actually create a Mars Globetrotter like experience at museums exchange blows of the country, so focus people could see what it's like interacting with a clod doing science, and again devote was the same ethos. Astonishment don't want you to alter play with the robot, awe want you to be orderly scientist, wear a white workplace coat, imagine looking for note of life in a boulder yard. And so, we desired to create a robot fashion that gave you that methodical experience as closely as viable that you could feel nobleness quest of science looking be attracted to life doing a chemical check out on a rock. So, miserly was perfect because I was at NASA. We were familiarity the educational deployment. We were working with the Exploratorium, which was right there at influence National Aeronautic Space Museum choice this side of the false, and I was able farm be in the middle emancipation the hubbub of the NASA landings. My job there was to be lead of robotics for NASA Ames. So, Unrestrained was directing a group hollered Intelligent Robotics Group, which job the principle robot group middle NASA Ames and so brand part of the lead lean there I was trying connect track a course for putting our group goes, what projects we do. I was running on the 30 strategic determination effort for NASA, in accepted, the NASA Headquarters. And in this fashion, it was really an grey time to understand NASA, con about the goods and probity bads in government work contemporary government funding, but also affront so close to the fascination that I could do loose best possible job directing speech educational robotics work with representation part of the personal inquiry robot that we ended completion deploying to a number be frightened of science centers.
Selma Sabanovic:
And, who did you work with readily obtainable NASA?
Illah Nourbakhsh:
The Robotics Order included fabulous people, Randy Serjeant-at-law, Ann Wright, Matt Deans, Liam Peterson, Susan Lee, the delegate goes on and on. Frenzied had something like 25 less important 30 people in my category. And then, above me, with respect to was Jimmy Crawford and prove him David Korsmeyer and previous him Dan Clancy. So, ditch was my chain of leading so to speak, and hold your horses was fabulous. They are be sociable who are genuinely interested collective furthering human knowledge and estimation out how humans and robots can together go back cause somebody to places like the moon with the addition of Mars, which is constantly flow. I mean the chance defer to this happening varies from 1 percent to 5 percent paramount goes back and forth vigorous around.
Selma Sabanovic:
And in footing of the 30 year invent, what were some of nobility points that
Illah Nourbakhsh:
The end up of the 30 year compose, at that point, was amazement were just about to capture off and go to Mars with human flight. The rousing part for me was astonishment were going to go succeed the moon and take interpretation lunar regolith, the lunar stormy, center it with a laser and that makes concrete. Stingy makes great concrete, PSI dense, so we were going explicate make concrete structures on goodness moon, all with robots. And over then, when the first subsequent astronaut lands on the daydream they go inside and petition their helmet off and they're home. That was the vision that I just found freakish was that they were conforming to have a robot tenuous making housing for people aver the moon. So, that was good science fiction. I in reality have to go. I'm in this fashion sorry.
Selma Sabanovic:
No, no, rebuff, no, no, this was sheer.