BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Dynamic Network Tomography: Model\, Algorithm\, Theory\, and Appli
 cation - Prof. Eric Xing (CMU)
DTSTART:20100311T140000Z
DTEND:20100311T150000Z
UID:TALK23539@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:Across the sciences\, a fundamental setting for representing a
 nd interpreting information about entities\, the structure and organizatio
 n of communities\, and changes in these over time\, is a stochastic networ
 k that is topologically rewiring and semantically evolving over time\, or 
 over a genealogy. While there is a rich literature in modeling invariant n
 etworks\, until recently\, little has been done toward modeling the dynami
 c processes underlying rewiring networks\, and on recovering such networks
  when they are not observable. \n\nIn this talk\, I will present a number 
 of recent developments on analyzing what we refer to as the dynamic tomogr
 aphy of evolving networks. I will first present new formalisms for modelin
 g network evolution over time\; and then\, new algorithms for estimating t
 he structure of evolving networks underlying nonstationary time-series or 
 tree-series of nodal attributes\, along with theoretical results on the as
 ymptotic sparsistency of the proposed methods\; and finally\, Bayesian met
 hods for estimating and visualizing the trajectories of latent multi-funct
 ionality of nodal states in the evolving networks. \n\nI will show some pr
 omising empirical results on recovering and analyzing the latent evolving 
 social networks in the US Senate based voting history\, the evolving gene 
 network of fruit fly while aging\, and the gene network evolving along cel
 l lineage during breast cancer progression and reversal\, at a time resolu
 tion only limited by sample frequency. In all cases\, our methods reveal i
 nteresting dynamic patterns in the networks.\n
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR