Time-varying dynamic Bayesian network reconstruction with information sharing
Add to your list(s)
Download to your calendar using vCal
If you have a question about this talk, please contact Konstantina Palla.
Network reconstruction is a well-studied topic, with applications in systems biology, pathway
medicine and social science among other areas. In many of the most interesting problems,
the network structure does not remain static over time, but changes as a result of environmental
factors or outside interventions. Thus we require network reconstruction techniques that can deal
with time-varying networks. In this talk, I present a technique based on dynamic Bayesian networks
with change points. The network structure, parameters, and the changepoints are all inferred from
the data using reversible-jump MCMC . In addition, we use information sharing priors to leverage
the commonalities in network structure between adjacent time epochs. The model is applied to
two real-world examples of changing networks; a network of muscle development genes in
drosophila during morphogenesis, and the IRMA synthetic biology gene network that can be
turned on or off by changing the growth medium. I demonstrate that the model improves on
previous approaches both in terms of change point detection and network reconstruction.
This talk is part of the Machine Learning Reading Group @ CUED series.
This talk is included in these lists:
Note that ex-directory lists are not shown.
|