BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:www.tcs.tifr.res.in/event/300
DTSTAMP:20230914T125918Z
SUMMARY:An Online Network Tomography Algorithm
DESCRIPTION:Speaker: Gugan  Thoppe\n\nAbstract: \nNetwork tomography is the
  science of inferring spatially localized network behavior using only metr
 ics that are practically feasible to measure. Mathematically\, given a mat
 rix $A\,$ the goal of network tomography is to estimate the statistics of 
 $X\,$ a vector of mutually independent random variables\, from the measure
 ment model $Y = AX.$ The challenge in these problems stems from the fact t
 hat $A$ is usually an ill-posed matrix and hence non-invertible. In this t
 alk\, using the stochastic approximation variant of the Kaczmarz's (SAK) a
 lgortihm\, we will see an online scheme for estimating the expected value 
 of $X$ using only IID samples of the components of the random vector $Y.$
   Importantly\, we will prove that\, starting from the same initial point
 \, the SAK algorithm\, when only samples of components of $Y$ are availabl
 e\, and the usual Kaczmarz algorithm\, when $EY$ is exactly known\, conver
 ge to precisely the same point (this result is joint work with Prof. Vivek
  Borkar and Prof. D. Manjunath (IIT Mumbai)).\n
URL:https://www.tcs.tifr.res.in/web/events/300
DTSTART;TZID=Asia/Kolkata:20120824T150000
DTEND;TZID=Asia/Kolkata:20120824T163000
LOCATION:A-212 (STCS Seminar Room)
END:VEVENT
END:VCALENDAR