Christopher Bishop, assistant director for Microsoft Research at Cambridge, U.K., has just published a new text book titled, “Pattern Recognition and Machine Learning.” The textbook as claimed “provides a comprehensive introduction to the fields of pattern recognition and machine learning.” Its target audience is advanced undergraduates and graduate students which means that you need a good background on computer science and mathematics to fully understand the material in the book.
This is the most recent book on the subject and so the author was able to cover some of the most recent advances in machine learning including probabilistic graphical models and inference methods taking exclusively a Bayesian perspective. The book is 738 pages long and it includes over 400 exercises for the students. It starts with an introduction to probability, decision and information theory followed by a discussion of probability distributions, regression and classification. It then moves on to a brief introduction of neural networks and sparse kernel machines. After introducing graphical models the author moves on to present the Expectation Maximization (EM) algorithm and then different exact and approximate inference methods. The book concludes with a discussion of methods for handling sequential data.
This is not Bishop’s first book. He has previously published three books on artificial Neural Networks. He is a well known research scientist with a number of publications in prominent scientific conferences and journals. Bishop holds a B.A. in physics from Oxford and a PhD in Theoretical Physics from the University of Edinburgh. Currently, he is the assistant director and head of the machine learning and perception group at Microsoft Research at Cambridge, U.K.
The book is published by Springer and it is currently available for purchase. Other than a textbook, I expect that it will also be a very good reference book and a great addition to anyone’s collection.