David MacKay
.


.
Environment and Energy 
.
Information Theory 
·Textbook 
·Gallager codes & MN codes 
·Theory of Sparse Graph codes 
·Regular Gallager codes 
·Irregular Gallager codes 
·Turbo codes  
·Insertions and deletions 
·Quantum Codes 
·Cryptanalysis and Coding 
·Discussion documents 
·More Message-Passing 
.
Probabilistic Data Modelling 
·PhD thesis « 
·Bayes for neural networks 
·ICA & Latent Variable Models 
·Bayes FAQ 
·Bayesian computations 
·Monte Carlo methods 
·Variational methods 
·Gaussian processes 
·Language, Proteins, Density nets 
·Images, Radar, Neuroscience 
·Materials science 
·Decision theory, Fair Elections 
.
Human-computer Interfaces 
.
Computational Neuroscience 
.
Evolution and Molecular Biology  
.
Miscellaneous work 
.
Errata 
.
Software and Files 
.
Recommendations 
.
Any Problems? 


Search :


.

PhD Thesis (Caltech 1991) and related papers:

My thesis consisted of four published papers, one unpublished chapter, and `summary' and `postscript' chapters. If you are only interested in one published chapter then you might prefer to download the single `Neural Computation' files listed after each chapter below. The thesis version is to be preferred, however, as it contains slightly more information. The whole thing, postscript | djvu | pdf | Also available from Caltech library |
Contents
pp. i-vi (30K); |DJVU|
Chapter 1: Summary
pp. 01-06 (40K); |DJVU|
Chapter 2: Bayesian Interpolation
pp. 07-17 (86K); |DJVU|
pp. 18-25 (68K); |DJVU|
pp. 26-33 (91K); |DJVU|
or as published in Neural Computation 4 3 415-447 (188K). |DJVU|
Chapter 3: A Practical Bayesian Framework for Backprop Networks
pp. 34-43 (94K); |DJVU|
pp. 44-52 (71K); |DJVU|
or as published in Neural Computation 4 3 448-472 (144K).| |DJVU|
Chapter 4: Information-based objective functions for active data selection
pp. 53-64 (122K); |DJVU|
or as published in Neural Computation 4 4 589-603 (125K). |DJVU|
Chapter 5: The evidence framework applied to classification networks
pp. 65-69 (85K); |DJVU|
pp. 70-70 (195K); |DJVU|
pp. 71-78 (191K); |DJVU|
or as published in Neural Computation 4 5 698-714 (456K). |DJVU|
Chapter 6: Inferring input-dependent noise levels
pp. 79-81 (34K); |DJVU|
Chapter 7: Postscript
pp. 82-86 (39K); |DJVU|
Bibliography
pp. 87-92 (30K). |DJVU|
The Inference Group is supported by the Gatsby Foundation
and by a partnership award from IBM Zurich Research Laboratory
David J.C. MacKay
Site last modified Wed Nov 12 17:38:44 GMT 2014