Title: Approximate Bayesian inference for Potts models
Authors: Yanan Fan - University of New South Wales (Australia) [presenting]
Abstract: Markov random fields play an important role in image analysis. A well-known problem is that for data on a large lattice, computation of the normalising constants quickly becomes intractable. We propose two approaches to overcome this problem on a regular lattice. In both cases, some form of approximation is used, with computational efficiency being the main trade-off.