- Page 1: COMPUTATION AND NEURAL SYSTEMS SERI
- Page 4 and 5: Library of Congress Cataloging-in-P
- Page 6 and 7: viPrefacenew professional society h
- Page 8 and 9: viiiPrefaceprocessing model, we hav
- Page 10 and 11: xPrefaceThere are, however, several
- Page 12 and 13: xiiContents4.3 The Hopfield Memory
- Page 15: HIntroduction toANS TechnologyWhen
- Page 19 and 20: Introduction to ANS TechnologyOutpu
- Page 21 and 22: Introduction to ANS Technology(b)Fi
- Page 23 and 24: 1.1 Elementary NeurophysiologyCell
- Page 25 and 26: 1.1 Elementary Neurophysiology 11Pr
- Page 27 and 28: 1.1 Elementary Neurophysiology 13Fi
- Page 29 and 30: 1.1 Elementary Neurophysiology 15wh
- Page 31 and 32: 1 .2 From Neurons to ANS 1 7Exercis
- Page 33 and 34: 1.2 From Neurons to ANS 19Notice th
- Page 35 and 36: 1 .2 From Neurons to ANS 21units, t
- Page 38 and 39: 24 Introduction to ANS Technologya
- Page 40 and 41: 26 Introduction to ANS TechnologyOu
- Page 42 and 43: 28 Introduction to ANS TechnologyIn
- Page 44 and 45: 30 Introduction to ANS Technologyle
- Page 46 and 47: 32 Introduction to ANS TechnologyFi
- Page 48 and 49: 34 Introduction to ANS Technologyne
- Page 50 and 51: 36 Introduction to ANS TechnologyN
- Page 52 and 53: 38 Introduction to ANS Technologyou
- Page 54 and 55: 40 Introduction to ANS TechnologySy
- Page 56 and 57: 42 Introduction to ANS Technology[5
- Page 59 and 60: C H A P T E RAdaline and MadalineSi
- Page 61 and 62: 2.1 Review of Signal Processing 471
- Page 63 and 64: 2.1 Review of Signal Processing 49A
- Page 65 and 66: which describes a typical square wa
- Page 67 and 68:
2.1 Review of Signal Processing 53W
- Page 69 and 70:
2.2 Adaline and the Adaptive Linear
- Page 71 and 72:
2.2 Adaline and the Adaptive Linear
- Page 73 and 74:
2.2 Adaline and the Adaptive Linear
- Page 75 and 76:
2.2 Adaline and the Adaptive Linear
- Page 77 and 78:
2.2 Adaline and the Adaptive Linear
- Page 79 and 80:
2.2 Adaline and the Adaptive Linear
- Page 81 and 82:
2.2 Adaline and the Adaptive Linear
- Page 83 and 84:
2.3 Applications of Adaptive Signal
- Page 85 and 86:
2.3 Applications of Adaptive Signal
- Page 87 and 88:
2.4 The Madaline 73= -1.5Figure 2.1
- Page 89 and 90:
2.4 The Madaline 75with the least c
- Page 91 and 92:
2.4 The Madaline 77right of the top
- Page 93 and 94:
2.5 Simulating the Adaline 79Retina
- Page 95 and 96:
2.5 Simulating the Adalinerecord la
- Page 97 and 98:
2.5 Simulating the Adaline S3Adalin
- Page 99 and 100:
2.5 Simulating the Adaline 85mented
- Page 101:
Bibliography 87[8] Rodney Winter an
- Page 104 and 105:
90 BackpropagationOutput read in pa
- Page 106 and 107:
92 Backpropagationan inordinate amo
- Page 108 and 109:
94 BackpropagationFigure 3.3 serves
- Page 110 and 111:
96 Backpropagation1. Apply an input
- Page 112 and 113:
98 Backpropagationwhere we have use
- Page 114 and 115:
1 00 Backpropagationmethod. Moreove
- Page 116 and 117:
1 02 Backpropagation1. Apply the in
- Page 118 and 119:
104 Backpropagationexample, the lim
- Page 120 and 121:
106 Backpropagation-- E,wFigure 3.6
- Page 122 and 123:
108 BackpropagationFigure 3.7This B
- Page 124 and 125:
110 BackpropagationTraining the Net
- Page 126 and 127:
112 BackpropagationAutomatic Paint
- Page 128 and 129:
114 Backpropagationfor the network
- Page 130 and 131:
116 Backpropagation3.5.2 BPN Specia
- Page 132 and 133:
118 Backpropagationof the Adaline s
- Page 134 and 135:
120 BackpropagationThe final routin
- Page 136 and 137:
122 BackpropagationHere again, to i
- Page 138 and 139:
124 Backpropagationyour modificatio
- Page 141 and 142:
HThe BAM and theHopfield MemoryThe
- Page 143 and 144:
4.1 Associative-Memory Definitions
- Page 145 and 146:
4.2 The BAM 131All the 6ij terms in
- Page 147 and 148:
.^Although we consistently begin wi
- Page 149 and 150:
4.2 The BAM 135For our first trial,
- Page 151 and 152:
4.2 The BAM 137is fairly easy to ap
- Page 153 and 154:
4.2 The BAM 139term valley to refer
- Page 155 and 156:
4.3 The Hopfield Memory 141where a
- Page 157 and 158:
4.3 The Hopfield Memory 143Figure 4
- Page 159 and 160:
4.3 The Hopfield Memory 145A =50(a)
- Page 161 and 162:
4.3 The Hopfield Memory 147In Eq. (
- Page 163 and 164:
4.3 The Hopfield Memory 149The TSP
- Page 165 and 166:
4.3 The Hopfield Memory 1512. Energ
- Page 167 and 168:
4.3 The Hopfield Memory 153causes a
- Page 169 and 170:
4.3 The Hopfield Memory 155where Aw
- Page 171 and 172:
4.4 Simulating the BAM 1571 2 3 4 5
- Page 173 and 174:
4.4 Simulating the BAM 159weight_pt
- Page 175 and 176:
4.4 Simulating the BAM 161The disad
- Page 177 and 178:
4.4 Simulating the BAM 163for refer
- Page 179 and 180:
4.4 Simulating the BAM 165of signal
- Page 181 and 182:
Bibliography 167Suggested ReadingsI
- Page 183 and 184:
C H A P T E RSimulated AnnealingThe
- Page 185 and 186:
5.1 Information Theory and Statisti
- Page 187 and 188:
5.1 Information Theory and Statisti
- Page 189 and 190:
5.1 Information Theory and Statisti
- Page 191 and 192:
5.1 Information Theory and Statisti
- Page 193 and 194:
5.2 The Boltzmann Machine 179where
- Page 195 and 196:
5.2 The Boltzmann Machine 1811. For
- Page 197 and 198:
5.2 The Boltzmann Machine 183popula
- Page 199 and 200:
5.2 The Boltzmann Machine 185Hf, on
- Page 201 and 202:
5.2 The Boltzmann Machine 187Thenan
- Page 203 and 204:
5.3 The Boltzmann Simulator 189Fort
- Page 205 and 206:
5.3 The Boltzmann Simulator 191(b)F
- Page 207 and 208:
5.3 The Boltzmann Simulator 193We w
- Page 209 and 210:
5.3 The Boltzmann Simulator 195ANNE
- Page 211 and 212:
5.3 The Boltzmann Simulator 197appl
- Page 213 and 214:
5.3 The Boltzmann Simulator 199Befo
- Page 215 and 216:
5.3 The Boltzmann Simulator 201Fina
- Page 217 and 218:
5.3 The Boltzmann Simulator 203begi
- Page 219 and 220:
5.3 The Boltzmann Simulator 205begi
- Page 221 and 222:
5.4 Using the Boltzmann Simulator 2
- Page 223 and 224:
5.4 Using the Boltzmann Simulator 2
- Page 225 and 226:
Suggested Readings 211All that rema
- Page 227 and 228:
C H A P T E RThe Counterpropagation
- Page 229 and 230:
6.1 CPN Building Blocks 215y' Outpu
- Page 231 and 232:
6.1 CRN Building Blocks 217The vect
- Page 233 and 234:
6.1 CPN Building Blocks 219(a)(b)Fi
- Page 235 and 236:
6.1 CPN Building Blocks 221Initial
- Page 237 and 238:
6.1 CPN Building Blocks 223Aw = cc(
- Page 239 and 240:
6.1 CPN Building Blocks 225'! '2 '
- Page 241 and 242:
6.1 CRN Building Blocks 227where x'
- Page 243 and 244:
6.1 CPN Building Blocks 229f(w)w(a)
- Page 245 and 246:
6.1 CPN Building Blocks 231y' Outpu
- Page 247 and 248:
6.1 CRN Building Blocks 233UCRUCS\(
- Page 249 and 250:
6.2 CPN Data Processing 2356.2 CPN
- Page 251 and 252:
6.2 CPN Data Processing 237The comp
- Page 253 and 254:
6.2 CPN Data Processing 239those ou
- Page 255 and 256:
6.2 CRN Data Processing 241Region o
- Page 257 and 258:
6.2 CRN Data Processing 243x' Outpu
- Page 259 and 260:
,6.3 An Image-Classification Exampl
- Page 261 and 262:
6.4 The CPN Simulator 247(a)(b)Figu
- Page 263 and 264:
6.4 The CPN Simulator 249outsweight
- Page 265 and 266:
6.4 The CPN Simulator 251our simula
- Page 267 and 268:
6.4 The CRN Simulator 253Before we
- Page 269 and 270:
6.4 The CRN Simulator 255learned, w
- Page 271 and 272:
6.4 The CPN Simulator 257doj = rand
- Page 273 and 274:
6.4 The CRN Simulator 259VFigure 6.
- Page 275 and 276:
Programming Exercises 261code for t
- Page 277 and 278:
C H A P T E RSelf-Organizing MapsTh
- Page 279 and 280:
7.1 SOM Data Processing 265topology
- Page 281 and 282:
7.1 SOM Data Processing 267(a)(b)Fi
- Page 283 and 284:
7.1 SOM Data Processing 269(0,0) (0
- Page 285 and 286:
7.1 SOM Data Processing 271(a)(b)Fi
- Page 287 and 288:
7.1 SOM Data Processing 273to assoc
- Page 289 and 290:
7.2 Applications of Self-Organizing
- Page 291 and 292:
7.2 Applications of Self-Organizing
- Page 293 and 294:
7.3 Simulating the SOM 279The resul
- Page 295 and 296:
7.3 Simulating the SOM 281model of
- Page 297 and 298:
7.3 Simulating the SOM 283domag = p
- Page 299 and 300:
7.3 Simulating the SOM 285row = (W-
- Page 301 and 302:
7.3 Simulating the SOM 287Figure 7.
- Page 303 and 304:
Bibliography 289to the number of tr
- Page 305 and 306:
H A P T E RAdaptiveResonance Theory
- Page 307 and 308:
8.1 ART Network Description 293equi
- Page 309 and 310:
8.1 ART Network Description 2951 0
- Page 311 and 312:
8.1 ART Network Description 2978.1.
- Page 313 and 314:
8.2 ART1 299Exercise 8.2: Show that
- Page 315 and 316:
8.2 ART1 301is inactive, G is not i
- Page 317 and 318:
8.2 ART1 303To all F 2 unitsFrom F,
- Page 319 and 320:
8.2 ART1 305from the F 2 layer. Sin
- Page 321 and 322:
8.2 ART1 307on connections from tho
- Page 323 and 324:
8.2 ART1 309Recall that |S| = |I| w
- Page 325 and 326:
8.2 ART1 311on FI and N be the numb
- Page 327 and 328:
8.2 ART1 313Since L = 3 and M = 5,
- Page 329 and 330:
8.2 ART1 315In this case, the equil
- Page 331 and 332:
8.3 ART2 317Orienting , Attentional
- Page 333 and 334:
8.3 ART2 319of keeping the activati
- Page 335 and 336:
8.3 ART2 321Notice that, after the
- Page 337 and 338:
8.3 ART2 323not want a reset in thi
- Page 339 and 340:
8.3 ART2 3254. Propagate forward to
- Page 341 and 342:
8.4 The ART1 Simulator 327and the t
- Page 343 and 344:
8.4 The ART1 Simulator 329rho : flo
- Page 345 and 346:
8.4 The ART1 Simulator 331Notice th
- Page 347 and 348:
8.4 The ART1 Simulator 333for i = 1
- Page 349 and 350:
8.4 The ART1 Simulator 335• We up
- Page 351 and 352:
Suggested Readings 337Programming E
- Page 353:
Bibliography 339[11] Stephen Grossb
- Page 356 and 357:
342 Spatiotemporal Pattern Classifi
- Page 358 and 359:
344 Spatiotemporal Pattern Classifi
- Page 360 and 361:
346 Spatiotemporal Pattern Classifi
- Page 362 and 363:
348 Spatiotemporal Pattern Classifi
- Page 364 and 365:
350 Spatiotemporal Pattern Classifi
- Page 366 and 367:
352 Spatiotemporal Pattern Classifi
- Page 368 and 369:
354 Spatiotemporal Pattern Classifi
- Page 370 and 371:
356 Spatiotemporal Pattern Classifi
- Page 372 and 373:
358 Spatiotemporal Pattern Classifi
- Page 374 and 375:
360 Spatiotemporal Pattern Classifi
- Page 376 and 377:
362 Spatiotemporal Pattern Classifi
- Page 378 and 379:
364 Spatiotemporal Pattern Classifi
- Page 380 and 381:
366 Spatiotemporal Pattern Classifi
- Page 382 and 383:
368 Spatiotemporal Pattern Classifi
- Page 384 and 385:
370 Programming Exercisesthe networ
- Page 387 and 388:
C H A P T E RThe NeocognitronANS ar
- Page 389 and 390:
The Neocognitron 375The visual cort
- Page 391 and 392:
10.1 Neocognitron Architecture 377S
- Page 393 and 394:
10.1 Neocognitron Architecture 379(
- Page 395 and 396:
10.2 Neocognitron Data Processing 3
- Page 397 and 398:
10.2 Neocognitron Data Processing 3
- Page 399 and 400:
10.2 Neocognitron Data Processing 3
- Page 401 and 402:
10.2 Neocognitron Data Processing 3
- Page 403 and 404:
10.3 Performance of the Neocognitro
- Page 405 and 406:
10.4 Addition of Lateral Inhibition
- Page 407:
Bibliography 393References at the e
- Page 410 and 411:
396 Indexsummary, 101, 160training
- Page 412 and 413:
398 Indexdifferential, 359, 361lear
- Page 414 and 415:
400 Indexassociator (A) units, 22,