BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Efficient Simulation and Inference for Stochastic Reaction Network
 s - Raul Fidel Tempone (King Abdullah University of Science and Technology
  (KAUST))
DTSTART:20160404T130000Z
DTEND:20160404T134500Z
UID:TALK65292@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Co-authors: CHRISTIAN BAYER (WIAS\, BERLIN)\, CHIHEB BEN HAMMO
 UDA (KAUST\, THUWAL)\, ALVARO MORAES (ARAMCO\, DAMMAM)\, FABRIZIO RUGGERI 
 (IMATI\, MILAN)\, PEDRO VILANOVA (KAUST\, THUWAL)&nbsp\;<br><br>Stochastic
  Reaction Networks (SRNs)\, that are intended to describe the time evoluti
 on of interacting particle systems where one particle interacts with the o
 thers through a finite set of reaction channels. SRNs have been mainly dev
 eloped to model biochemical reactions but they also have applications in n
 eural networks\, virus kinetics\, and dynamics of social networks\, among 
 others.&nbsp\;<br><br>This talk is focused on novel fast simulation algori
 thms and statistical inference methods for SRNs.&nbsp\;<br><br>Regarding s
 imulation\, our novel Multi-level Monte Carlo (MLMC) hybrid methods provid
 e accurate estimates of expected values of a given observable at a prescri
 bed final time. They control the global approximation error up to a user-s
 elected accuracy and up to a certain confidence level\, with near optimal 
 computational work.&nbsp\;<br><br>With respect to statistical inference\, 
 we first present a multi-scale approach\, where we introduce a determinist
 ic systematic way of using up-scaled likelihoods for parameter estimation.
  In a second approach\, we derive a new forward-reverse representation for
  simulating stochastic bridges between consecutive observations. This allo
 ws us to use the well-known EM Algorithm to infer the reaction rates.&nbsp
 \;<br>
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR