View Table of ContentsCreate a BookmarkAdd Book to My BookshelfPurchase This Book Online
Skip to Book Content
Book cover image

Index

Neural Smithing: Supervised Learning in Feedforward Artificial Neural Networks
Russell D. Reed and Robert J. Marks II
Copyright © 1999 Massachusetts Institute of Technology

Previous Section Next Section

Index

H

Hard-limiters, 134
Hard weight sharing, 269
Hebbian learning, 30
Hessian matrix
approximation of, 129-132
gradient-based methods and, 163, 166-167, 170-174, 182
ill-conditioned, 127-128, 179
learning rate and, 72, 81-84
node splitting and, 214-215
problem size and, 157
properties of, 127-132
pruning and, 223-225
quadratic function and, 179
search then converge method and, 147
Hidden layers
definition of, 2, 31
equivalent solutions and, 194
error surface and, 119, 124, 126
representational capabilities and, 32-47
weight-space symmetries and, 119
Hidden nodes
approximation error and, 37
derivative calculation for, 54, 56
nonrandom initialization and, 106, 107, 108
random initialization and, 102, 103-104
Hint functions, 275
Holdout method, 257-258
Homogeneously linear separability, 20
Hooke-Jeeves pattern search, 160-161
Hopfield networks, 49, 183
Hyperplanes, 15-18
capacity of, 20-23
nonrandom initialization and, 107
random initialization and, 102, 103
representational capability and, 42-44
stair-steps and, 114

Top of current section
Previous Section Next Section
Books24x7.com, Inc. © 1999-2001  –  Feedback