The Role of Correlation Decay in Filtering Compressed Signal Dynamics
Wednesday, May 20, 2015 - 2:00pm - 2:50pm
We consider the problem of filtering compressed high-dimensional signal measurements. We show that the structure of the measurement matrix through which the signal is compressed and an associated correlation decay property can be exploited to develop particle filters that can avoid the curse of dimensionality. For the simplest sequential Monte Carlo algorithm of this type we prove an approximation error bound that is uniform both in time and in the model dimensions. This is joint work with Patrick Rebeschini.