|Physics Seminar : Using the Information in Experimental Observations to Develop and Test Quantitative Models
When quantitative, usually nonlinear, models of observed phenomena are developed they are typically incomplete as many unknown parameters remain. Further to use the model to forecast future behaviour, one must accurately estimate the full state of the system when predictions begin and measurements are usually only of a subset of the variables describing the system state.
We describe and illustrate tools for the exact description of transferring information from such sparse measurements to quantitative models when the observations are noisy, the model has errors, and there is uncertainty in the state of the system when observations begin. While the discussion is quite general, the ideas are illustrated by examples selected from neurobiology and geophysics. The important role of nonlinearity in the procedure is emphasized.
Henry D.I. Abarbanel, Distinguished Professor of Physics, Marine Physical Laboratory (Scripps Institution of Oceanography), Center for Theoretical Biological Physics, University of California)
Rm 2:15, Physics, UWA
: 6488 2738
Tue, 15 Mar 2011 15:45
Tue, 15 Mar 2011 16:45
Gay Hollister <[email protected]>
Thu, 21 Apr 2011 09:53
- Locations of venues on the Crawley and Nedlands campuses are
available via the Campus Maps website.
- Download this event as:
Mail this event: