The HLF Portraits: Sanjeev Arora

Video thumbnail (Frame 0) Video thumbnail (Frame 9438) Video thumbnail (Frame 18876) Video thumbnail (Frame 22056) Video thumbnail (Frame 32445) Video thumbnail (Frame 39482) Video thumbnail (Frame 45740) Video thumbnail (Frame 57056) Video thumbnail (Frame 68372) Video thumbnail (Frame 79687)
Video in TIB AV-Portal: The HLF Portraits: Sanjeev Arora

Formal Metadata

The HLF Portraits: Sanjeev Arora
Title of Series
No Open Access License:
German copyright law applies. This film may be used for your own use but it may not be distributed via the internet or passed on to external parties.
Release Date

Content Metadata

Subject Area
The Heidelberg Laureate Forum Foundation presents the HLF Portraits: Sanjeev Arora; ACM Prize in Computing, 2011 Recipients of the ACM Prize in Computing in discussion with Marc Pachter, Director Emeritus National Portrait Gallery, Smithsonian Institute, about their lives, their research, their careers and the circumstances that led to the awards. Video interviews produced for the Heidelberg Laureate Forum Foundation by the Berlin photographer Peter Badge. The opinions expressed in this video do not necessarily reflect the views of the Heidelberg Laureate Forum Foundation or any other person or associated institution involved in the making and distribution of the video. Background: The Heidelberg Laureate Forum Foundation (HLFF) annually organizes the Heidelberg Laureate Forum (HLF), which is a networking event for mathematicians and computer scientists from all over the world. The HLFF was established and is funded by the German foundation the Klaus Tschira Stiftung (KTS), which promotes natural sciences, mathematics and computer science. The HLF is strongly supported by the award-granting institutions, the Association for Computing Machinery (ACM: ACM A.M. Turing Award, ACM Prize in Computing), the International Mathematical Union (IMU: Fields Medal, Nevanlinna Prize), and the Norwegian Academy of Science and Letters (DNVA: Abel Prize). The Scientific Partners of the HLFF are the Heidelberg Institute for Theoretical Studies (HITS) and Heidelberg University.

Related Material

Point (geometry) Trail Service (economics) Civil engineering Decision theory Multiplication sign Direction (geometry) Disk read-and-write head Expected value Frequency Mathematics Computer configuration Term (mathematics) Software testing Series (mathematics) Physical system Social class Inheritance (object-oriented programming) Gradient Bit Cartesian coordinate system Connected space 10 (number) Personal digital assistant Universe (mathematics) Video game Right angle Musical ensemble Family
Point (geometry) Principal ideal Satellite Observational study Multiplication sign Student's t-test Heat transfer Web 2.0 Selectivity (electronic) Position operator Physical system Email Information Gradient Physicalism Hecke operator Cartesian coordinate system Data mining Process (computing) Universe (mathematics) Computer science Video game Right angle Procedural programming Musical ensemble Pressure Family
Point (geometry) Group action State of matter Decision theory Multiplication sign Direction (geometry) 1 (number) Online help Parallel computing Student's t-test Mereology Coprocessor Field (computer science) Neuroinformatik Hypothesis Goodness of fit Mathematics Latent heat Bridging (networking) Term (mathematics) Algebra Form (programming) Task (computing) Email Algorithm Inheritance (object-oriented programming) Key (cryptography) Gradient Physicalism Bit Linear algebra Library catalog Cryptography Faculty (division) Process (computing) Software Telecommunication Computer science Right angle Video game console Quicksort
Axiom of choice Point (geometry) Existence Wage labour Divisor Cryptosystem Direction (geometry) Decision theory Real number Range (statistics) Set (mathematics) Parallel computing Solid geometry Field (computer science) Neuroinformatik Mathematics Different (Kate Ryan album) Computer configuration Core dump Theorem Mathematical optimization Dissipation Optimization problem Gradient Keyboard shortcut Bit Cartesian coordinate system Limit (category theory) Cryptography Theoretical computer science Equivalence relation Faculty (division) Phase transition Formal grammar Video game Procedural programming Quicksort Arithmetic progression
Complex (psychology) Group action Thread (computing) Multiplication sign Direction (geometry) Modal logic Range (statistics) Boom (sailing) Approximationsalgorithmus 1 (number) Bell and Howell Water vapor Travelling salesman problem Mereology Dimensional analysis Neuroinformatik Web 2.0 Mathematics Machine learning Different (Kate Ryan album) Core dump Series (mathematics) Data conversion Social class Area Collaborationism Algorithm Optimization problem Software developer Moment (mathematics) Bit Process (computing) Exterior algebra Order (biology) Computer science Normal (geometry) Right angle Quicksort Point (geometry) Turing test Trail Server (computing) Service (economics) Civil engineering Virtual machine Web browser Field (computer science) Hypothesis Element (mathematics) Power (physics) Goodness of fit Internetworking Natural number Term (mathematics) Authorization Boundary value problem Focus (optics) Shift operator Graph (mathematics) Information Projective plane Line (geometry) Approximation Computational complexity theory Faculty (division) Software Personal digital assistant Information retrieval Universe (mathematics) Musical ensemble
Point (geometry) Building Multiplication sign Orientation (vector space) Direction (geometry) 1 (number) Virtual machine Student's t-test Field (computer science) Power (physics) Mathematics Machine learning Term (mathematics) Different (Kate Ryan album) Natural number Energy level Extension (kinesiology) Algorithm Electric generator Artificial neural network Software developer Gradient Bit Surface of revolution Evolute Theoretical computer science Process (computing) Internet forum Computer science Quicksort Metric system Resultant Asynchronous Transfer Mode
Bit rate
[Music] professor how surprised are you to find yourself sitting here in Princeton and middle of your professional life you were born a long way away from here you tell me a little bit about your childhood yeah I was born in India in a city in Rajasthan called Jodhpur and my family moved around a little bit during my childhood we were in Delhi for a few years and Ohta because of your father's yeah what did he do he was a professor also in civil engineering in the civil engineering in the cities engineering college and our mother was at home children three three children yeah one brother one sister I'm always interested in the expectations that parents have for their children as you were growing up let's imagine you 810 years old what kind of expectations tightness were you getting from your parents so in India of course parents are very concerned about their kids education and the expectations are set according to that but apart from that I don't think there were any special expectations yeah we did well in our classes from fairly early age and so with an early passion evident in one direction or another or just enjoy school of taking all the courses yeah I don't think I had the special passion or anything she even right through high school I was just good at yeah and yeah I mean India you certainly back then there were no special camps or special classes right I didn't even know I was gifted it anyway I mean I was just yes good in my one class then there are hundreds of classes around the city I know that yeah so the teachers of some significance in that early pre-university period no not I mean yeah nothing well maybe maybe by high school I mean there was some yeah without the more connection with the teacher and with the subject but yeah 70 up to grade 10 or nine maybe not much yeah I spoke to another person of this series of interviews of Indian origin and in his household the whole expectation and push was to the Civil Service the idea that you would do what it was to get you to that point there was none of that I don't think civil service per se I mean clearly that was considered the epitome of success or something but but I think there was some general expectation among kids those days that so in the Indian competitive exams you you're tested on your general knowledge so you know that you know a lot about lots of things a generalist that's what you're tested for right and the special passion that's actually it was a disadvantage you need to know the broad well I suddenly know I helped yeah yeah so yeah there was no at least when I was growing up there wasn't a sense that you have to excel in pretty good at discipline well when you graduate from high school and was in a public school private school private school which is called public school oh it's your holdup customer in a English way but it was a private school in our terms and comes a time where you have to decide what to do after graduation how are you making that decision from high school or from high school so there again in India there are from an early age you you it gets from into your head that the two professions are engineer or doctor and in those days I'm not sure if the system is still the same now probably still that after ninth grade or 10th grade you you decide on the math track or the life sciences track okay to get into one of those two I'm just guessing you decided that's right yeah okay so you're now on your way to possibly being an engineer yeah or something about this subject okay so what do you do after high school how does one does what apply again is the application system similar to American universities is there a general test one less thing right so the system is completely different and still is the case there's a national exam and the premier institutions certainly for a technical education were the IITs the Indian Institutes of Technology and there was a general exam of national example for that you show two days in the year where you know there's one exam everybody takes it determines your life you know it determines you admission into these premier Institute's and you're gonna really get ranked nationally yeah I'm just guessing you did well I did well yes you had the option of going where to any of the Institute's which I went to IIT Kanpur
there was no good reason I mean this is way before or somewhat before World Wide Web this is 1986 so certainly in India we didn't even have emailed that then certainly in the small cities so there was no other information I mean you some you know somebody who when their who liked it you know those kinds of yeah yeah it wasn't based on anything yeah not great but it was an institution of prestige that's right one of the fight yeah right and okay you arrived but do you have to again instructing me in the university process there is this is very helpful you have to declare a major you basically have a prescribed course that you must take how does it go right so that's actually interesting so sometime in high school I started thinking maybe I'd like to do physics just because physics has such I think in retrospect looking back physics has a tremendous outreach you know popular books and everybody knows who I enstein is and know that so here's some what I Bish's would would be fair to call you at this age a business yes so I think by at that point because I'd done so well in the National exam I I mean that was the first time I realized that I was a potential yeah I mean certainly yeah to be among the top people in the country in that year so so yeah I had a college principal who here I used to chat a lot and he was urging me to do something with my life not as he would say you know and just do the same material go for the same material success that everybody wants to do and I was this thing yeah and and certainly just by reading and this you know the physics outreach and so on I mean I I had a sense for you know beautiful sign and that seemed like an attractive choice but since I scored very well in this national exam I could pick any major so you're the major you got dependent on your understanding in this exam Wow so so I could pick any major and the top major in those days and probably still today is computer science in India and so I had some familial pressure to at least try computer science and and so so that's why not completely so then around that time a cousin of mine had gotten into a mighty and this side no idea how this happens because some some university over in the u.s. admits you and gives you full scholarship which translated into Indian rupees is some ginormous son so how the heck that happens I had no idea but because this cousin of mine who is who who lived in Delhi and was you know what they were much more well-off so they were much more globally connected yes so this cousin had figured out that system SATs and and recommendations and the whole American system so he had got me into a mighty with full scholarship so that's so that's when I loved that this is possible and what this procedure is that my cousin and so then I decided to apply for an American University and and maybe do physics there because I could see that at at IIT doing physics or chemistry wasn't as glamorous exciting right because in those days the goal of going to premier university was so that you could go to grad school yours so I thought I would just short-circuit that even go as an undergrad and maybe study physics or something so this was not actually formally because of dissatisfaction with where you were but just seeing this new possibility and maybe translating a longer study into a shorter amount of time no I was interested in research I thought I would like to do research yes NIT are certainly you'd been holding position for four years until you get to go to grad school and do research whereas I had a sense that in the u.s. at a top university has you so what age were you when you make the a tale for a party eighteen eighteen yeah so as a freshman I applied for a transfer right and then I got accepted but then I didn't come in freshman year because my application will financially it didn't reach them somehow got lost in the mail oh yeah so so then I wrote to them and they deferred my admission by a year and so I came mastership sophomore year had they promised that money would come in a year or they were just willing to wait with you and see whether you would get financial support the next year yeah I I think yeah they gave me financial support the next year not they didn't promise me but I think I had the sense from my cousin I think that the the bigger hurdle is to get in at that time MIT was very good that way they would admit a small select band of international students and everybody who needed money with get whatever money they did so would you go I'm assuming your family is excited about this for you rather than saying don't go so far
but maybe then maybe there was some discussion about you're not going no okay yeah Indian parents said oh Fiat take down were thrilled I wouldn't say thrill I mean they're never thrilled to send their kid away so far away but yeah it was a good thing yeah I think that was so much of the first experience cultural shock when you come to MIT or do you immediately find yourself among into interesting people and you're happy and you're the homesick or is it a rough first year well intellectually it was 20 fantastic I mean you open this thick book of courses and you're free to take anything whoever heard of such a thing I mean India they tell you okay you come in these are six courses you take this year so that was thrilling couple of courses and you can pick anything socially you know it takes some time to adjust one thing I did was I got I got admitted to this Independent Living group kind of like a co-ed fraternity but it wasn't affiliated with the National Organization about 30 people in a house so that was nice in that house all right I might be sure they were in it yeah including one of the faculty a few yeah consoling or something it's just oh no yeah faculty were busy and I don't think I interacted with any Indian faculty there I mean I'm sure I talked to one or two but yeah but yeah the people I did research where on earth oh okay help me with your decision what's at MIT is to which direction your
education you have the banquet before you the big catalog yeah how do you select ah so that was actually interesting so in the year because I hadn't I have to defer for a year so I had more time to further my education and I guess at the back of my mind I thought going anyway so I started reading on my own and I had access to more books at the library and and I started getting more interested in math now in retrospect now I realized I had no idea what math was I mean I'm talking about linear algebra basic algebra which is very different from research mathematics but that sounded kind of cool too so I was a little bit less sure of doing physics at that point and then I come to MIT and I realized that there is this major called mathematical computer science and I looked into this field a little bit probably over the summer or something or maybe soon as I write and that sounded pretty exciting as well you know things like cryptography and algorithms that sounded all pretty cool as well so so I ended up in that that major and I actually took a physics course and I sort of was less interested in that were you prepared to jump into a computer specialty youdescribe India at that time is barely even having email and so forth you may not have had much access to computers before you got there so how was it that you are embracing this oh so and remember at IIT I was studying for your science for two-year self-service I had taken a few courses which I actually didn't like so much because those are programming courses but this math end of computer science which is what I ended up doing seemed much more exciting who are your key professors as you develop this specialty at MIT yeah so I was lucky enough to end up in the research group of Professor Tom Leighton and he had some grad students who were mentoring me so that was fantastic so I'm probably within a couple of weeks after that I arrived at MIT I was in this group with this and they were part of a larger group of maybe 10 to 12 faculty maybe some in this field that's not a good computer science so what questions were they pursuing the your seniors so to speak be my professor is the graduate students were there specific directions of inquiry where where are we now and we're in the late eighties no yeah okay they did these yeah so in terms of the state of the field in terms of their kinds of research interests what are they're pursuing with you're learning from them at this point so this group that I was a member of was interested in parallel computing so we just saw that makes interesting field right now so you have multiple processors in the computer instead of one like thousands and you want to harness them to do computations faster which is not as easy as it sounds you know you would imagine that sure I mean it may be I guess you can suppose there's a task you know which one person can do well now if you give it to a thousand people right I mean it's like too many cooks boiling broth right so it's not clear how to make that all work it appears how's the divided job a thousand so that that was the insurer here that that's made interested at this problem of this oh I should have mentioned that getting into the research group was also a form of employments so which I I needed to have a campus job as part of my scholarship okay so this was the campus job I got helping out with the research and whatever they ain't done you are intellectually engaged I mean this is not just standing by somebody who's doing the real work and you're feeding something I mean you you are challenged from the this group intellectually it's in stages I mean the way it works and this happens to me from the other side now you know that you take on all kinds of students and some are more talented than others and then the ones who are talented get involved more and given interesting problems and become wildly guess again that you were talented and so yeah so you got involved in interesting process after a year and after a year was it yeah and so we had a paper within a couple months which I mean so you're you're published in your second year there yeah diamond-like after my first year that summer okay what was that it was in parallel computing yeah connecting a parallel computers and speeding up communication amongst them what was another specialist but what was the insight that the paper rested on the inside was that if you connect up the processors randomly then that's actually a very good interconnection network so you imagine that based upon our you know how you design the city's transportation infrastructure or something that you take a lot of planning to decide okay this world snakes here and whatever this bridge there but we discovered that I mean this was part of that group but I mean I helped to push it further what we discovered was that randomly interconnected networks actually really good at communicating fast yeah I mean not in hindsight but yeah if you first tell somebody that fact right seems very conversant right as opposed to a fully design by human those the MIT undergraduate process require a thesis at the end of the for you not know so you you took the courses you developed your interests fortunately in this this advanced world
so I need to get you graduated I have decided on your next step right so then yeah I decided I like this field theoretical computer science that then a few courses in it as an undergrad done research so I decided to pursue it in grad school and that's like to grad schools and chose Berkeley what would the choice of schools and did the choice involve different emphases did the various schools present themselves was particularly good in one direction or another that you would be interested in and so I applied in this discipline the mathematical end of spear science say I understand what difference would it have made to about to Berkeley as opposed to maybe stated MIT or yeah that was an option yeah that was that was probably my two top choices yeah I visited there and I really liked new solid California yeah and I love their mighty intellectually what I thought it was a little lacking socially so it was a so um you would have advanced in either place in short but it was a social decision to choose between two intellectually equal opportunities yeah yeah so where people make decisions very often as long as it's comfortable if you talk to Berkeley right what you did yeah do you think that that though in the end affected the direction of your research for having gone to Berkeley what doing there yeah I mean these are the unknowns of life almost certainly yeah I mean you go to a different place you you have to do you do end up doing other things I think one thing I realized what that are not thing in retrospective it happens is that going to a new place is terrifying no in general any kind of change is terrifying you lose the comforts of whatever you're used to but it helps you grow and I mean I had already done that I'd I'd done in high school midway through high school because my father had to move and then went to IIT and then MIT and I could see I think that it's terrifying but it's that leads to growth right yeah it allows you to change because nobody knows you in the new place yeah you can do whatever you can present yourself are so sort of be a beneficial decision in that sense yeah so you have to figure out things I knew what the fact of DRL whereas at MIT I knew already what the master so there's yeah you have the equivalent of a large range of possibilities and you have a faculty with whom you can follow at various aspects how are you making those decisions now what next intellectually for you so that was a little bit less clear yeah so towards the end of my undergrad career had become interested in the P versus NP problem and in fact that was my goal in grad school to work on that I had started working with some faculty at MIT who this is not in parallel computing now so this is one topics related to famous Pearce's NP problem so I was trying to pursue that at Berkeley and now in retrospect I mean there's still basically no progress that problem so well what is the agenda you do a labor what is the court challenge in P versus NP so in P versus NP the core question is is whether or not there is any advantage to getting lucky so a problem is many possible solutions and you're looking for the optimal solution so an NP is the set of problems where when you have an optimal optimal solution before you you can verify ok the solution you're looking for the correct solution you can verify it so so I won't give a technical
definition but that side yeah yeah so it's kind of like you know your math problems it's difficult to solve them but when somebody gives you a solution it's much easier to verify that it's correct so that sort of is an example of an NP problem all right so so you're trying to show that so the question the P versus NP question is can all such problems be solved efficiently and and you would guess that no the answer is no clearly no because there's many possible solutions and you may it's it's kind of like looking for a needle in a haystack the correct solution and so once you find the needle you'll recognize it's there but how do you find it in the haystack so but the question is is that you know that there's no shortcut to finding the right haystack and it's a famous question in mathematics at this point and people believe that answer is no that computers don't have a shortcut for finding the needle in the haystack but nobody is able to show that it's a difficult problem that's a very basic question of that broadly can one divide these inquiries into theoretical and applied interest as you develop careers you're now in graduate school are these questions that fascinate for their own sake or in your mind is there some goal of application for it right so if the P versus NP conjecture is false namely that P equals NP which means that there is an efficient way to finally get needles in haystacks then as I said this without great practical implications for example there would be no cryptography because you'll be able to break crypto systems efficiently on the other hand most interesting problems of optimization in real life would be solvable efficiently so that's a consequence so so it has many practical consequences in this conjecture sports right but to show that the conjecture is true they're somewhat for your application okay so cryptography might exist or something but there are somewhat fewer applications it's more of a philosophical question it's a little bit intellectually descended from girdles and worried girdle theorem which plays sharp limitations on formal mathematics so these the P versus NP question is is descended from that because it actually turns out pertains to the existence of efficient ear and prune procedures is it fair to say just as a depression for what you just described that your own interest in let's even call it intellectual temperament is more toward the philosophical are you happy to live within that kind of problem-solving without so you mean like the abstract method so so certainly in that phase of life yes I was very interested in that I mean I've had the faces and so that was the question does that lead to well it does lead to dissipation somehow what what is that dissertation going to be okay so not so I tried this inquiry for a year I did Oh and I wasn't getting very far and of course I became aware of all the other smart people talked to them you know at conferences and so on and right realized they were all stuck to exactly and then my in the start of my first year I start reading some other new
research that had come out which people had told me was very important and interesting and so that is what I really managed to make some contributions to and became my thesis which was not P versus NP actually okay so again to layman what is the core element or challenge addressed by your thesis okay so so it's inspired by P versus NP so suppose you assume that P is different from NP there's no efficient way to get optimum solutions to problems to these large class of problems can you compute efficiently approximate solutions so you just want to get within 10% of optimal define it appropriately something like that you know can you do that efficiently so this is a question of approximation and thanks to some work that had happened there was just stuff happening in those your to up leading up to that point they were the right techniques that that had become available to address these questions and in fact what I showed with colleagues was that actually for many interesting problems computing approximate solutions is no easier than appealing optimal solutions so therefore if you believe in peace being different from NP then you should not expect to have a fish approximation of applause for these problems for these so this is essential is what you demonstrated that's what I demonstrate yeah so it's a consequence of P not being MP that you also wouldn't have good approximation algorithms I want to put it another way if you had a good approximation algorithm you would actually have an exact algorithm I just so now you're at the cusp of finishing the formal formal education of an academic career the beginnings do see yourself because you are broadly in the computer computational world do you see yourself as having the alternative of an industrial or an academic future or is the direction of your research and interest always going to keep you in the academic world yeah so this kind of inquiry mathematical and of computer science algorithms complexity yeah it is done in an industrial research lab so at the time that would have been IBM or Bell Labs two main ones and some smaller labs but I was more interested in academia there wasn't any storm of the soul about where shall I go no yes you were you were headed in an academic frankly where are you going to go now that you have your dissertation that have a series of problems that interest you what next so I ended up here that was the best job offer I had that year and I ended up at Princeton so yeah tell me what Chris's virtues were as you looked to a career here Princeton even back then had a very strong research group in this area including tartan was Turing Award winner and and Yahoo that later on some years after that won the Turing award and many other very highly regarded researchers so this was certainly a very good job for me and [Music] but even by that point I sort of was aware that I don't stick with one field of inquiry for very long okay I'm interested in all kinds of different things so I guess that only emphases in India of being a generalist yes normal kinds of things sort of stuck stuck with means it doesn't stick with all Indian so I don't think it's then you know that did with you it did with me and what was the nature of an invitation to join a faculty a community of researchers here wasn't the did you have to demonstrate again at least at that point a particular line of inquiry that interested you or they were essentially just interested let's call it in your curiosity yeah so to this day to I mean we hire people we hire people not you know people to do with certain projects and that was true back then too so you try to hire the best people and of course it that would mean that they haven't demonstrated something they will have done some some some accomplishments but they're not held to that as now you have to follow the following that's right right and especially in computer science it's it changes completely every decade so not completely but a lot so so characterize it the culture of computer science in are we now in the early 90s wherever I'm 1994 I tell me that with 94 yeah a moment in time yeah what is the broad culture what are the exciting questions being asked or at least exciting to you at this point ok so maybe I'll answer the first one so remember this is the time when the Internet is taking off the the web exists and yeah everybody I think there was a browser at Netscape or was it that's now called Mozilla and we are starting the Internet boom around that 1994 maybe 95 96 was when it took off and that had a big impact on pure science before that computer science was very much geared towards big iron computing for industry servers and so on and with the with the web taking off I think it became much much very different I mean there was much more interest in information retrieval information processing understanding networks so networking became a very big area so yeah computer science as a discipline shifted a lot where do you find your places so me [Music] so I became more interested in algorithms rather than computational complexity we're still interested in complexity but my focus was shifting and
I became interested in these approximation album so for my dissertation work had shown that for a lot of problems you couldn't do better than a certain approximation ratio for a bunch of problems but what is the correct Meishan ratio for all these interesting problems that you can achieve so so I became interested in actually designing algorithms for approximation algorithms for for many of these problems I had some success with Traveling Salesman problem which is a class of problem and various other problems along those lines graph partitioning some years later say oh and I'm becoming I'm switchings in retrospect from is there more mathematical computational complexity to more mathematical design of algorithms in my first seven eight years at Princeton that was a shift I that happened are you finding a particular intellectual companionship are you doing projects together with others or mostly developing individual questions there's always the authorization matters yeah there's always a conversation but I wasn't closely collaborating with any other Princeton colleagues in this new endeavor but there were other colleagues you know the universities that I did collaborated with investments what is the result of that those years of inquiry in terms of conclusions you came to the parties into a new dimension I mean I think you mostly stayed with that broader inquiry of algorithms that's right so that was the comfortable place for you that was tailoring the right one that's right but what were the ongoing challenges within that the new questions you you are asking along the way are asking I mean you're yeah we're in full career that's right so right so looking back I think I'm interested in understanding the power of algorithms and lately I've been interested in the power of machine learning albums what's possible with that but yeah that's been the maybe the common thread if you will and now given the boundaries of the disciplines of sub disciplines of computer science that inquiry crosses some boundaries but but yeah that's the unifying thread maybe do you I mean again I don't want to get epic here because everything is a step at a time but do you see the implications of greater understanding of algorithm affecting a wide range of kinds of inquiry even beyond formally computational mathematics and so forth what is your oh yeah right so actually that was the other shift maybe that I forgot to mention in the 1990s that suddenly the water discovered that that science is moving in the big data era as well whether it's biology genomics particle physics astronomy so all these Sciences were moving into the big data era and they realized that they needed add for those new algorithms so that was another big shift which didn't impact my research until more recently but but suddenly it was informing the development of algorithms within computer science right point so so indeed yeah design of algorithms broadly conceived is is a big part of your science so who was coming to you these days I mean are you are you getting interest from colleagues and fields you would not have expected initially - yeah so because I'm doing machine learning especially in the last few years six oh six seven years so that's maybe even more directly relevant to to the Big Data era say I have collaborations with colleagues and other departments at this point the intellectual culture in which you happily find yourself appraise them is it less boxed in and more a broader conversation among field so yeah that's my romantic notion of Princeton that it maybe is less just absolutely field driven is that true suddenly yeah easy to find other people but I'm not sure if it's it's probably true for many major universities I'm guessing so yeah I think there's a there's a lot of cross-disciplinary work including in our department right I'm not sure if it's more so than other than your voice needs but yeah maybe I all talking about it is a steer yeah it's definitely valued a lot and it happens a lot I got some sense from the the reading about you that I did before this that we're quite interested in teaching it's not one of those necessary evils in order to get your research time but it seems to be part of how do you think can you describe yourself as a little bit as the teacher and how this relates to your research yeah so teaching is is valued a lot at Princeton firstly and yeah I think I was always sort of a teacher it took me a long time to realize that but I was teaching looking back a lot of my classmates all along and probably from a fairly early age yeah so and actually I was talking to some Indian colleagues at some conference dinner once you know notice that you know four or five of us we all Indians and and I raises hypothesis that maybe at least Indians who come here are sort of on the achievement track that maybe all of them were teachers from an early age and they all confirmed that that was the case really yeah it's culturally maybe bit different in the u.s. that the Nerds were maybe it may be similar to hungry or something the Nerds were looked up to because it was assumed that they would be successful and I get those civil service jobs so they were they were among the respected kids in the class and rather there was the underground that there were in America right and there was no shame in asking them for
help and so they were actually considered very useful friends to have yeah so yeah so instead of it clicked in my mind only a few years ago that this is very different in my upbringing or my background then it is in my American politics are you have you found yourself that we're doing papers together with some of your all the time yeah all the almost exclusively yeah I mean all of my people almost all my cake was a grad student clock a quarter yeah I don't know what it was last time when I had a paper which had no grad student quarter yeah so but they also to finish teaching things that yeah so I think I think I just absorbed from a very early age that explaining to other students helped me understand much better which is again something all teachers know but I think I'd sort of result in a road eh so teaching is very specifically about algorithms and again based on something that I've read about you it seems to me you're you're trying to find new ways to teach about how there is a rhythms and so forth are you are you finding that you're teaching now the subjects that you're so interested in has changed in terms of how you engage your students and the problems or is this yeah there's an evolution based upon my own research interests I'm more interested in machine learning and big data problems so so that's only affected what I teach and what I find interesting and what I think I should know but beyond that I think throughout my career it's been kind of fun to figure out where the discipline should go in terms of teaching because it's a fascinating discipline so within algorithms and theoretical computer science what is interesting to teach today you know bigger better then every time assuming that to be wrong of course that one of the things that students ask of their professors but you give me the advanced levels is where should I go in my research I mean of course the very good students are ones who have their own passions and there are curiosity as you did inserted in your own career but guidance you know where do you where do you think the most interesting significant developments will be where will the money come in to support my research do you get those questions and what are your answers so yeah lots of questions like that yeah undergrads usually undergrads yeah I mean grad students have done that kind of introspection a lot already before coming to grad school usually but undergrad sale they combine a lot to figure out those kinds of Demond testicles I do give guys yeah yeah I hear enough yourself no I mean to the extent that it pertains to research and how to get into research and so on yeah sure so are there big fields of it some one of the earlier generations that era particularly we're at the cusp of a revolution but you - I mean the world before you became enters the intellectually in these questions was very different from now are we on the cusp of explorations and various fields that your students would feel excited to join is there going to be directions you know I'm not asking for prophecy I mean they're talking about the next ten years research that you think is particularly exciting that they may pick up so I think the cross-disciplinary work that I've done lately is more applying machine learning ideas to other disciplines there's a different kind of process than reward which some of my colleagues to where they are completely embedded in both fields and in particular in some other field of differ different from clear science neuroscience or biology and they have joint appointments and so on so that's a different kind of work which I have so far less experience so I mean I'm not so deeply embedded in other disciplines the people from other disciplines may come with a specific problem related to machine learning or analyzing data maybe are you are you finding some of your students embedding themselves in two fields is that how does that happened none of my students has done that I mean yeah sort of more in my mode today I was sort of curious and interested in collaborating with people in other disciplines but they're not embedded in all right last question will really be at this point in your in your work are you developing earlier ideas and enquiries are you beginning to wonder about new things and new directions in terms of your orientation oh yeah I've had a complete not complete but very drastic change of inquiry in the last six years or so as I indicated I'm interested in the power of algorithms for machine learning and artificial intelligence that's building as an interest that's yeah yeah that's what I'm embedded in quickly at this point and yeah so that involves a lot of retooling Julie in terms of learning first principles or just yeah yeah basically yeah I mean most of it is new I mean much of this didn't even exist when I was in grad school but that's true in computer science yeah you if you don't change your fields you get absolutely very fast because fields do shift a lot but even so I think yeah maybe I was among the early ones who made the switch from theoretical computer science into machine learning now there may be more people you're still designing algorithms but of a very different nature and the metrics of success are different you know what's a successful algorithm I mean it's a little bit more of a technical discussion but it's very different some of the issues are very different so so it's a different mode of thinking so thank you
you you