BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:"The Power of Non-Convex Relaxations for Stochastic Discrete Optim
 ization"
DTSTART:20180502T110000
DTEND:20180502T120000
DTSTAMP:20260916T032121Z
UID:3a56ab7082ada63dbdb11be1c85703dede6154fd62a3672318b8deac
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Amin Karbasi\, University of Yale\nMany procedures in st
 atistics and artificial intelligence require solving non-convex problems. 
 Historically\, the focus has been to convexify the non-convex objectives. 
 In recent years\, however\, there has been significant progress to optimiz
 e non-convex functions directly. This direct approach has led to provably 
 good guarantees for specific problem instances such as latent variable mod
 els\, non-negative matrix factorization\, robust PCA\, matrix completion\,
  etc. Unfortunately\, there is no free lunch and it is well known that in 
 general finding the global optimum of a non-convex optimization problem is
  NP-hard. This computational barrier has mainly shifted the goal of non-co
 nvex optimization towards two directions: a) finding an approximate local 
 minimum by avoiding saddle points or b) characterizing general conditions 
 under which the underlying non-convex optimization is tractable.\n\nIn thi
 s talk\, I will consider a broad class of non-convex optimization problems
  that possess special combinatorial structures. More specifically\, I will
  focus on maximization of stochastic continuous submodular functions. Desp
 ite the apparent lack of convexity\, we will see that first order methods 
 can indeed provide strong approximation guarantees. We then see that by us
 ing stochastic continuous relaxation as an interface\, we can also provide
  tight approximation guarantees for maximizing a stochastic submodular set
  function.\nIn this talk\, I will not assume any particular background on 
 submodularity or optimization and will try to motivate and define all the 
 necessary concepts.\n\nBio: Amin Karbasi is currently an assistant profess
 or of Electrical Engineering\, Computer Science\, and Statsitics at Yale 
 University. He has been the recipient of AFOSR 2018 Young Investigator Awa
 rd\, Grainger Award 2017 from National Academy of Engineering for interdi
 sciplinary research\, Microsoft Azure research award 2017\, DARPA 2016 Yo
 ung Faculty Award\, Simons-Berkeley fellowship 2016\, Google Faculty Award
  2015\, and ETH fellowship 2013. His work has been recognized with a vari
 ety of paper awards\, including Medical Image Computing and Computer Assi
 sted Interventions Conference (MICCAI) 2017\, International Conference on
  Artificial Intelligence and Statistics (AISTAT) 2015\, IEEE ComSoc Data S
 torage 2013\, International Conference on Acoustics\, Speech\, and Signal
  Processing (ICASSP) 2011\, ACM SIGMETRICS 2010\, and IEEE International 
 Symposium on Information Theory (ISIT) 2010 (runner-up). His Ph.D. work r
 eceived the Patrick Denantes Memorial Prize 2013 from the School of Compu
 ter and Communication Sciences at EPFL\, Switzerland. 
LOCATION:BC 329 https://plan.epfl.ch/?room==BC%20329
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR