6th HLF – Interviews with mathematics and computer science laureates: Sanjeev Arora

Video in TIB AV-Portal: 6th HLF – Interviews with mathematics and computer science laureates: Sanjeev Arora

Formal Metadata

6th HLF – Interviews with mathematics and computer science laureates: 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
Laureates at the 6th HLF sit down with Tom Geller, Tom Geller Productions, to discuss their career, mentoring and their experience at the Heidelberg Laureate Forum (HLF). These renowned scientists have been honored with most prestigious awards in mathematics and computer science: Abel Prize, ACM A.M. Turing Award, ACM Prize in Computing, Fields Medal and Nevanlinna Prize. 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

Mathematics Internet forum Computer Musical ensemble
NP-hard Algorithm Machine learning Scaling (geometry) Software developer Model theory Virtual machine Computational science Deep Web NP-complete Social class
Point (geometry) Trail Variety (linguistics) Multiplication sign Virtual machine Student's t-test Open set Mereology Field (computer science) Expected value Goodness of fit Mathematics Different (Kate Ryan album) Hypermedia YouTube Area Gradient Expert system Bit Total S.A. Cartesian coordinate system Numerical analysis Faculty (division) Computational complexity theory Integrated development environment Internet forum Universe (mathematics) File archiver Video game Quicksort
Area Group action Observational study Multiplication sign Moment (mathematics) Model theory Virtual machine Open set Cartesian coordinate system Field (computer science) Expected value Data model Mathematics Process (computing) Strategy game Statement (computer science) Video game Natural language Computational science Musical ensemble
Mathematics View (database) Distribution (mathematics) Internet forum Computer Videoconferencing
could you briefly summarize what your work is about the work for which you won the ACM prize so there's a class of problems and computer science called np-hard problems or np-complete problems and for which we know that we are unlikely to have any efficient algorithms assuming P is not equal to NP which is a famous mathematical conjecture so a natural question that arose after that discovery was whether you could compute approximate solutions to these problems and in my work I've I
and my co-author showed that computing even approximate solutions is np-hard also for many of these np-hard problems so even approximate solutions may not be possible and then I also did some work showing for some of those problems there are reasonable approximate solutions that you can compute and and you mentioned before we started the interview that you're now more interested or you've gotten into machine learning was it it and what is your work regarding machine learning in particular so machine learning currently for the last few years has involved computational problems that also appear to be np-hard and intractable but somehow we can solve quite well quite fast at very large scale and that's a big mystery so that's what got me interested in machine learning originally but now I'm interested in a broad swath of questions understanding deep nets and other models that are being studied in recent years and this new interest the the development of the interest from your earlier research was that made possible by new discoveries or
was it just sort of a natural switch for you it was a natural switch yes yeah I was intrigued that such seemingly intractable problems could be solved at such large scale in in a variety of applications so that's what got me interested in machine learning I'd like to go back a little bit since the HLF is so much about people starting their careers and find out sort of what led you up to the point where you led the where you won the prize where you were awarded the prize and what mentors and what examples you followed who was especially influential on you so yes I I think anybody who succeeds usually has a lot of mentors so it's it probably wouldn't be fair to just single out one person but at every stage in life from high school to undergrad I was an undergrad in two different institutions yeah but many faculty in those places who were I would say at least four or five who quite influential and encouraging and and then in grad school my advisors yeah and also as a faculty member you know colleagues at Princeton where I was faculty or other universities and yeah so I think the total number of people would be at least 10 or 15 or more I mean there are so many people who are very supportive and and give good guidance are there any particular qualities that you think made a mentor good for you no I think the I
think the most important part of getting I mean of a mentoring relationship is just being there and people who would make time in their busy lives to to listen to whatever question I had and give advice and that they were there I think that was probably the most important part and your own mentorship style since you're now teaching the students I assume mostly on mostly graduate students once you cried but a few undergrads every year yes and how do you approach that so I can maybe describe the research style first my research side is that I'm often interested in doing new things figuring out new things that I new areas that I did not know before and I think that's a big plus for the grad students because I'm not an expert either and so we're on an equal footing and often they you know they figure out things which I don't know and I think maybe that's very empowering do you see much difference between when you were a student and the students who are now in school as far as environment and what they're studying and the challenges they may have no I wouldn't say there's a big difference in the students but there's a big difference in the environment they are in so this archive and YouTube and social media and I think they have many more ways of keeping track of what's going on in the field then we did way back the pace is faster too so so I think that's the biggest change so going on to the HR left because of course this is also a mentorship situation is this your first one is my first one well do you have any expectations or hopes or fears even from this week no I just came with an open mind yeah I never been here people have spoken very highly of it so I'm looking forward to it are there any of the other laureates who you haven't met who you're you're looking forward to perhaps probably many of the laureates haven't met before yeah I would say maybe two-thirds at least yeah in your own field and as we
discussed machine learning as course a very very active field at the moment what do you think is the most exciting area at the moment so I think the most exciting area is that people are able to train very vast models based on large amounts of data effectively and [Music] so yeah the exciting thing is that that seems to be possible and then the more exciting thing is to figure out how to leverage that because there are many situations out there where we cannot leverage that so far and and finally something that I'm personally very interested in is to understand these large models you know how and why can we train them any particular any particular area of large data model that I'm referring to deep learning yeah for any particular application though I'm interested broadly uh I'm quite interested in natural language processing as an application area yeah is there anything else that you'd like
to say about about your time here or your award or your studies yeah I guess what I what I've learned in my career is that you can't go in with a preset expectation of what you're going to do and be flexible and open to working on new things and often that's where the action is not because often you don't have as a student or or actually actually in any stage of life you don't really have full information so if you go in with a preconceived notion of what you're going to work on for the next five years maybe that works in mathematics but in computer science I think that's usually not a good strategy so you have to be open to work on new new things new challenges it's funny that that echo is also your statement about what you expect from this week that you go in with an open mind yeah yeah I'm definitely of that character yeah well thank you very much [Music]


  344 ms - page object


AV-Portal 3.20.1 (bea96f1033d39fbe77f82542458e108105398441)