BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Non-stationary Network Modelling by Particle Filtering:  case of G
 ene Interaction Networks - Dr Ercan E Kuruoglu\, ISTI-CNR\, Pisa
DTSTART:20151102T150000Z
DTEND:20151102T160000Z
UID:TALK62264@talks.cam.ac.uk
CONTACT:Prof. Ramji Venkataramanan
DESCRIPTION:We have stayed  too long with traditional input-output system 
 models. The nature of modern data is different: they involve two-way inter
 acting variables in large quantities. MIMO models provide only a partial s
 olution and network modelling seems to be the way forward for dealing with
  data now vastly available in various applications ranging from social net
 works to finance and from genomics to mobile communications. Despite the e
 xplosion of research on big data\, the time varying dimension of the inter
 actions are widely ignored. \nIn this talk\, we address the problem of tim
 e varying network estimation with gene interaction networks as a case stud
 y. \n\nExisting methods used for gene regulatory network identification ar
 e mostly dedicated to inference of steady state networks which are prevale
 nt over all time instants. However\, the interactions among genes in a net
 work are not stationary during the life cycle of an organism. Moreover\, i
 nformation about the gene interactions in different stages of a life cycle
  is of high importance for biology in understanding of protein production\
 , human diseases and in designing personalized treatment plans. In the lit
 erature one can find a large amount of data measured at a single time inst
 ant. Unfortunately\, only a limited amount of sources present experimental
  data on temporal sequences for gene expression and most of available expe
 rimental data are measured sparsely or/and for a short time period. This l
 ack of experimental data significantly limits the success of inference on 
 network topology. We model the gene interactions over time with multivaria
 te linear regressions where the parameters of the regressive process are c
 hanging over time. We propose Particle Filtering for dynamic network infer
 ence and its potentials in time varying gene expression tracking are demon
 strated. The proposed model is able to track the interactions not only fro
 m the one step past but also interactions with a delay of n-time steps whi
 ch is a realistic scenario for gene interactions in general. Smoothness an
 d sparsity of network changes with time are also discussed. We would like 
 to stress that the method is easily extendable to model nonlinear interact
 ions. Moreover\, the proposed methodology is applicable in any type of tim
 e varying network including various other biological processes where\nvari
 ables evolve in relation to each other.
LOCATION:LR5\, Department of Engineering
END:VEVENT
END:VCALENDAR