Inverse Problems and Uncertainty Quantification

Marco Iglesias, Andrew M. StuartCaltechAUTHORS 2014

Abstract

Quantifying uncertainty in the solution 
\nof inverse problems is an exciting area of 
\nresearch in the mathematical sciences, one 
\nthat raises significant challenges at the 
\ninterfaces between analysis, computation, 
\nprobability, and statistics. The reach in 
\nterms of applicability is enormous, with 
\ndiverse problems arising in the physical, 
\nbiological, and social sciences, such as 
\nweather prediction, epidemiology, and traffic flow.
\n
\nLoosely speaking, inverse problems confront mathematical models with data so 
\nthat we can deduce the inputs needed to 
\nrun the models; knowledge of these inputs 
\ncan then be used to make predictions, and 
\neven to devise control strategies based 
\non the predictions. Both the models and 
\nthe data are typically uncertain, as are the 
\nresulting deductions and predictions; as 
\na consequence, any decisions or control 
\nstrategies based on the predictions will be 
\ngreatly improved if the uncertainty is made 
\nquantitative.
BibTeX
@inproceedings{iglesias2014inverse,
  author = {Marco Iglesias and Andrew M. Stuart},
  title = {{Inverse Problems and Uncertainty Quantification}},
  booktitle = {CaltechAUTHORS},
  year = {2014},
}

Loading related content...

Figure