Tensor completion for PDEs with uncertain coefficients and Bayesian Update

Files
Litvinenko_SIAMCSE.pdf (1.32 MB)

Type
Presentation

Authors
Litvinenko, Alexander

KAUST Department
Extreme Computing Research Center
Center for Uncertainty Quantification in Computational Science and Engineering (SRI-UQ)

Date
2017-03-05

Abstract
In this work, we tried to show connections between Bayesian update and tensor completion techniques. Usually, only a small/sparse vector/tensor of measurements is available. The typical measurement is a function of the solution. The solution of a stochastic PDE is a tensor, the measurement as well. The idea is to use completion techniques to compute all "missing" values of the measurement tensor and only then apply the Bayesian technique.

Acknowledgements
ECRC KAUST, SRI UQ KAUST

Conference/Event Name
SIAM CSE Conference

Additional Links
http://meetings.siam.org/sess/dsp_programsess.cfm?SESSIONCODE=61237

Permanent link to this record
http://hdl.handle.net/10754/623695
Collections
Presentations
Extreme Computing Research Center