SMS scnews item created by Lamiae Azizi at Fri 22 May 2020 1020
Type: Seminar
Distribution: World
Expiry: 26 Jun 2020
Calendar1: 25 Jun 2020 1600-1700
CalLoc1: Zoom
CalTitle1: DARE seminar
Auth: lamiae@plamiae.pc (assumed)

DARE Seminar: Prof Kerrie Mengersen -- Dynamic Bayesian Network Inferencing for Nonhomogenous Complex Systems

Dear all, 

Join us for a DARE seminar with Distinguished Professor Kerrie Mengersen, who
will be presenting a talk on Dynamic Bayesian Network Inferencing for Nonhomogenous
Complex Systems over Zoom. 

About Professor Kerrie Mengersen: 

Distinguished Professor Kerrie Mengersen is a statistician and Director of the Centre
for Data Science at QUT.  She is an elected Fellow of the Australian Academy of Science,
the Australian Social Sciences Academy, and the Queensland Academy of Arts and
Sciences.  As an ARC Laureate Fellow, she works on the development and application of
methods for using diverse types of data to learn about complex systems and problems in
health, the environment and industry.  She is currently working with international teams
to use data science to gain insights into national and international dynamics of
COVID-19.  

About her talk: 

Dynamic Bayesian networks (DBNs) provide a versatile method for predictive,
whole of systems modelling to support decision makers in managing natural systems
subject to anthropogenic disturbances.  However, DBNs typically assume a homogeneous
Markov chain which we show can limit the dynamics that can be modelled especially for
complex ecosystems that are susceptible to regime change (i.e.  change in state
transition probabilities).  Such regime changes can occur as a result of exogenous
inputs and/or because of past system states; the latter is known as path dependence.  

In this presentation, I will describe a method that we* developed for non-homogeneous
DBN inference to capture the dynamics of potentially path-dependent ecosystems.  The
method enables dynamic updates of DBN parameters at each time slice in computing
posterior marginal probabilities given evidence for forward inference.  

We demonstrated the methods on a seagrass dredging case-study and showed that the
incorporation of path dependence enables conditional absorption into and release from
the zero state in line with ecological observations.  The model can help managers to
develop practical ways to manage the marked effects of dredging on high value seagrass
ecosystems.  

*This work was led by Paul Wu and is joint with Julian Caley, Gary Kendrick and Kathryn
McMahon.  The foundation methodology was published in the Journal of the Royal
Statistical Society Series C, 67, 417-434.  

Zoom link:
https://uni-sydney.zoom.us/j/98918599620