Tous nos rayons

Déjà client ? Identifiez-vous

Mot de passe oublié ?

Nouveau client ?

CRÉER VOTRE COMPTE
Graphical models for machine learning and digital communication
Ajouter à une liste

Librairie Eyrolles - Paris 5e
Indisponible

Graphical models for machine learning and digital communication

Graphical models for machine learning and digital communication

Brendan J. Frey

220 pages, parution le 25/08/1998

Résumé

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.

Summary of contents

  • Probabilistic inference in graphical models
  • pattern classification
  • unsupervised learning
  • data compression
  • channel coding
  • future research directions

Table of contents


Series Foreword
Preface
1 Introduction
1.1 A probabilistic perspective
1.2 Graphical models: Factor graphs, Markov random fields and Bayesian belief networks
1.3 Organization of this book
2 Probabilistic Inference in Graphical Models
2.1 Exact inference using probability propagation (the sum-product algorithm)
2.2 Monte Carlo inference: Gibbs sampling and slice sampling
2.3 Variational inference
2.4 Helmholtz machines
3 Pattern Classification
3.1 Bayesian networks for pattern classification
3.2 Autoregressive networks
3.3 Estimating latent variable models using the EM algorithm
3.4 Multiple-cause networks
3.5 Classification of handwritten digits
4 Unsupervised Learning
4.1 Extracting structure from images using the wake-sleep algorithm
4.2 Simultaneous extraction of continuous and categorical structure
4.3 Nonlinear Gaussian Bayesian networks (NLGBNs)
5 Data Compression
5.1 Fast compression with Bayesian networks
5.2 Communicating extra information through the codeword choice
5.3 Relationship to maximum likelihood estimation
5.4 The "bits-back" coding algorithm
5.5 Experimental results
5.6 Integrating over model parameters using bits-back coding
6 Channel Coding
6.1 Review: Simplifying the playing field
6.2 Graphical models for error correction: Turbocodes, low-density parity-check codes and more
6.3 "A code by any other network would not decode as sweetly"
6.4 Trellis-contrained codes (TCCs)
6.5 Decoding complexity of iterative decoders
6.6 Parallel iterative decoding
6.7 Speeding up iterative decoding by detecting variables early
7 Future Research Directions
7.1 Modularity and abstraction
7.2 Faster inference and learning
7.3 Scaling up to the brain
7.4 Improving model structures
7.5 Iterative decoding
7.6 Iterative decoding in the real world
7.7 Unification
References
Index

L'auteur - Brendan J. Frey

Brendan J. Frey

is a Beckman Fellow, Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign.

Caractéristiques techniques

  PAPIER
Éditeur(s) The MIT Press
Auteur(s) Brendan J. Frey
Parution 25/08/1998
Nb. de pages 220
Format 15,2 x 23
EAN13 9780262062022

Avantages Eyrolles.com

Livraison à partir de 0,01 en France métropolitaine
Paiement en ligne SÉCURISÉ
Livraison dans le monde
Retour sous 15 jours
+ d'un million et demi de livres disponibles
satisfait ou remboursé
Satisfait ou remboursé
Paiement sécurisé
modes de paiement
Paiement à l'expédition
partout dans le monde
Livraison partout dans le monde
Service clients sav@commande.eyrolles.com
librairie française
Librairie française depuis 1925
Recevez nos newsletters
Vous serez régulièrement informé(e) de toutes nos actualités.
Inscription