BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:www.tcs.tifr.res.in/event/997
DTSTAMP:20230914T125946Z
SUMMARY:Scalable and Practical Discrete Optimization for Big Data
DESCRIPTION:Speaker: Rishabh Iyer (Microsoft\nRedmond\, WA)\n\nAbstract: \n
 Abstract: Data has been growing at an unprecedented rate in the last few 
 years. While this massive data is a blessing to data science by helping i
 mprove predictive accuracy\, it is also a curse since humans are unable to
  consume this large amount of data. Moreover\, majority of this data is 
 plagued with redundancy. In this talk\, I will present a powerful modeling
  abstraction which formulates these problems as a special class of combi
 natorial optimization called submodular optimization. I shall consider tw
 o applications of this paradigm\, one in summarizing massive data for hum
 an consumption\, and another in making machine learning models and traini
 ng processes more efficient. I shall also present a unified discrete grad
 ient based optimization framework for solving a large class of these opti
 mization problems\, which not only obtains good theoretical guarantees bu
 t is also easy to implement and scales well with large datasets. In additi
 on to describing the underlying algorithmic advances\, I will describe it
 s impact on several several concrete applications including visual data 
 summarization\, data subset selection\, data partitioning and diversified
  active learning. Finally\, I will describe a comprehensive\, highly opti
 mized\ndiscrete optimization package I developed (with my colleagues at UW
 ) which implements most state-of-the-art submodular optimization algorit
 hms\, and includes several implementation techniques to scale to large da
 tasets.\nBio: Rishabh Iyer is currently a Research Scientist at Microsoft\
 , where he works on several problems around Online Learning\, Contextual 
 Bandits\, Reinforcement Learning and Discrete Optimization with applicati
 ons to Computational Advertisement and Computer Vision. During his time a
 t Microsoft\, several of his algorithms and innovations have been shipped
  in Bing Ads platform substantially improving the efficiency of the syste
 m and Revenue. He finished his PostDoc and PhD from the University of Wa
 shington\, Seattle. His work has received best paper awards at ICML and N
 IPS. He also won the Microsoft PhD fellowship and Facebook PhD Fellowship
  (declined in the favor of Microsoft) in 2014\, along with the Yang Outst
 anding Doctoral Student Award from University of Washington. He has been 
 a visitor at Microsoft Research\, Redmond and Simon Fraser University.\nHe
  has worked on several aspects of Machine Learning including Discrete and
  Convex optimization\, deep learning\, video/image summarization\, divers
 ified active learning and data subset selection\, online learning etc.\n
URL:https://www.tcs.tifr.res.in/web/events/997
DTSTART;TZID=Asia/Kolkata:20190924T143000
DTEND;TZID=Asia/Kolkata:20190924T153000
LOCATION:A-201 (STCS Seminar Room)
END:VEVENT
END:VCALENDAR