Bayesian networks for pattern classification, data compression, and channel coding

grantor: University of TorontoPattern classification, data compression, and channel coding are tasks that usually must deal with complex but structured natural or artificial systems. Patterns that we wish to classify are a consequence of a causal physical process. Images that we wish to compress are also a consequence of a causal physical process. Noisy outputs from a telephone line are corrupted versions of a signal produced by a structured man-made telephone modem. Not only are these tasks characterized by complex structure, but they also contain random elements. Graphical models such as Bayesian networks provide a way to describe the relationships between random variables in a stochastic system. In this thesis, I use Bayesian networks as an overarching framework to describe and solve problems in the areas of pattern classification, data compression, and channel coding. Results on the classification of handwritten digits show that Bayesian network pattern classifiers outperform other standard methods, such as the k-nearest neighbor method. When Bayesian networks are used as source models for data compression, an exponentially large number of codewords are associated with each input pattern. It turns out that the code can still be used efficiently, if a new technique called "bits-back coding" is used. Several new error-correcting decoding algorithms are instances of "probability propagation" in various Bayesian networks. These new schemes are rapidly closing the gap between the performances of practical channel coding systems and Shannon's 50-year-old channel coding limit. The Bayesian network framework exposes the similarities between these codes and leads the way to a new class of "trellis-constraint codes" which also operate close to Shannon's limit.Ph.D