Product Description
A variety of problems in machine learning and digital communication deal with complex but structured natural or artificial systems. In this book, Brendan Frey uses graphical models as an overarching framework to describe and solve problems of pattern classification, unsupervised learning, data compression, and channel coding. Using probabilistic structures such as Bayesian belief networks and Markov random fields, he is able to describe the relationships between random variables in these systems and to apply graph-based inference techniques to develop algorithms. Among the algorithms described are the wake-sleep algorithm for unsupervised learning, the iterative turbodecoding algorithm, the bits-back coding method, the Markov chain Monte Carlo technique, and variational inference.
From the Publisher
Series Editor for Adaptive Computation and Machine LearningThe goal of building systems that can adapt to their environments and learn from their experience has attracted researchers from many fields, including computer science, engineering, mathematics, physics, neuroscience, and cognitive science. Out of this research has come a wide variety of learning techniques that have the potential to transform many industrial and scientific fields. Recently, several research communities have begun to converge on a common set of issues surrounding supervised, unsupervised, and reinforcement learning problems. The MIT Press Series on Adaptive Computation and Machine Learning seeks to unify the many diverse strands of machine learning research and to foster high quality research and innovative applications.
This book by Brendan Frey is a perfect illustration of the convergence of these fields around a common set of mathematical and computational tools. Frey studies hierarchical probabilistic models, algorithms for learning and reasoning with them, and applications of these models to problems in supervised learning, unsupervised learning, data compression, and perhaps most surprisingly, error-control coding. The recent discovery that certain algorithms for decoding error-correcting codes are performing belief propagation over Bayesian belief networks is thrilling. Not only does it provide yet another connection between information theory and machine learning, but it also suggests new approaches to the design of improved error-correcting codes and to the design and analysis of learning algorithms. Frey provides the reader with a beautifully written introduction to graphical models and their associated algorithms. Then he takes us on a tour of the research frontier and of his recent work, which has advanced this frontier significantly. The book concludes with a thought-provoking assessment of future directions in this exciting research area.
With an Amazon.co.uk Gift Certificate, you can get them what they want (even if you don't know what that is). 
Amazon Home
Make sure you always get what you want with an Amazon.co.uk 
