Connectionist Models of Behaviour and Cognition II: Proceedings of the 11th Neural Computation and Psychology Workshop (Progress in Neural Processing)

CONTENTS......Page 8
Preface......Page 6
Embedded Cognition......Page 12
1. Introduction......Page 14
2. Method......Page 16
3. Results......Page 21
References......Page 24
1. Introduction......Page 26
2.1. Simulated robotic set-up......Page 28
2.2. Architecture and functioning of neural network model......Page 29
2.3. Learning phases......Page 32
3. Results......Page 34
References......Page 36
1. Introduction......Page 38
2. Motivation......Page 39
3.1. The neural network model......Page 41
3.2. The learning procedure......Page 42
3.3. Functional comparison with the MNS......Page 43
4. Analyzing the dynamics of the neural network......Page 44
5. Discussion......Page 46
References......Page 47
1. Introduction......Page 50
2. The Selective Attention for Action Model (SAAM)......Page 52
2.1.2. Hand Network......Page 54
3.1. Experiment......Page 57
3.2. Simulations......Page 59
4. Conclusion and Outlook......Page 60
References......Page 61
Memory......Page 62
1. Introduction......Page 64
2. Methods......Page 67
3. Results......Page 69
4. Conclusions......Page 74
References......Page 75
1. Introduction......Page 78
2. Pyramidal neuron model......Page 79
3. Cortical column model......Page 81
4. Cortico-hippocampal system model......Page 82
4.1. Flow of inputs to the hippocampal system......Page 83
4.2. Competition within the DG and CA3......Page 84
5. Accounting for psychological phenomena......Page 85
6. Computer implementation of the proposed model......Page 86
References......Page 89
1. Introduction......Page 90
2. A simple model of structural plasticity, synaptic consolidation, and cortico-hippocampal interplay......Page 91
3. On the function and benefits of structural plasticity for memory storage......Page 94
4. Structural plasticity and catastrophic forgetting......Page 96
5. Retrograde amnesia and Ribot gradients......Page 97
6. Discussion......Page 98
References......Page 99
1. Introduction......Page 102
2.1. The missing-letter task......Page 103
3. Modeling context maintenance and semantic-STM in Schizophrenia......Page 104
3.1. Model for contextual disambiguation......Page 106
3.2. Results for contextual disambiguation......Page 107
3.4. Results for semantic STM......Page 108
4. Discussion......Page 109
Acknowledgments......Page 110
References......Page 112
1. Introduction......Page 114
2. Network Dynamics and Training......Page 115
3. Measuring Performance......Page 116
4. Network Topology......Page 117
5. Background and Motivation......Page 118
6. Results and Discussion......Page 119
6.1. Verification......Page 120
6.2. How do different pattern sets affect memory recall?......Page 121
7. Conclusion......Page 123
References......Page 124
Categorisation......Page 126
1. Introduction......Page 128
3. Clustering......Page 129
4. GEMINI......Page 131
5.1. Image retrieval by hierarchical neural networks......Page 133
5.2. Class space as input space......Page 135
5.3. Re-projection space as input space......Page 136
6. Conclusion......Page 138
References......Page 139
1. Introduction......Page 140
2.1. Type I mirror effects......Page 141
2.2. Type II mirror effects......Page 142
3. Our model......Page 143
4. Results......Page 146
4.1. Type I mirror effect......Page 147
5. Conclusions and further work......Page 148
References......Page 150
1. Introduction......Page 152
2.1. Gabor Filters......Page 153
2.2. Principal Component Analysis......Page 155
2.3. Curvilinear Component Analysis......Page 156
2.5. Encoding Face......Page 157
3. Experiments and Results......Page 158
References......Page 161
1. Introduction......Page 164
2. The Model......Page 166
2.1. Stimuli......Page 167
3.1. Effect of labels on cortical representations......Page 168
3.2. Vocabulary spurt and overextension of labels......Page 169
3.3. Statistical effects on word learning......Page 171
3.4. Effect of word knowledge on object familiarization......Page 173
References......Page 174
Temporal Aspects of Cognition......Page 176
1. Introduction......Page 178
2.1. Method......Page 179
3. Serial memory model......Page 181
3.1. Single interval representations by synfire chains......Page 182
3.2.1. Stochastic framework......Page 184
3.2.2. Results......Page 185
3.2.3. Approximation......Page 187
4. Discussion......Page 188
References......Page 189
1. Expectation and surprise in music......Page 190
2.1. Entropy and information......Page 191
2.3. Predictive information and the Wundt curve......Page 192
3. Information dynamics in Markov chains......Page 194
4. Learning and subjective information......Page 195
5. An analysis of minimalist music......Page 196
6. Conclusions and future work......Page 199
References......Page 201
Concepts and High-Level Cognition......Page 202
2. Concepts......Page 204
2.1. Historical theories......Page 205
3.1.1. The self-organizing map......Page 206
3.1.2. Models using self-organization......Page 207
3.2.2. Models using Bayesian inference......Page 208
3.3. Combining self-organizing and Bayesian models......Page 209
4. Experiment in the color domain......Page 210
4.1. Experimental setup......Page 211
4.2. Results......Page 212
5. Conclusions and discussion......Page 213
References......Page 214
1. Introduction......Page 216
2.1.2. Procedure......Page 217
2.1.4. Method of Analysis......Page 218
3. Experiment 2......Page 219
3.1.4. Design......Page 220
3.2. Results and Discussion......Page 221
4.1. Introduction to Dynamic Fields......Page 222
4.2. SPAM-Ling: Integrating Non-Linguistic and Linguistic Spatial Cognition......Page 223
4.4. Spatial Working Memory Drift in the Dynamic Field Theory......Page 224
4.6. Simulations......Page 225
5. Conclusion......Page 226
References......Page 227
1. Introduction......Page 228
2. Linear RAAM......Page 229
2.2. Terminal test......Page 231
2.3. Training the network......Page 232
3.2. Generation of word features......Page 233
3.3. Levels of systematicity......Page 235
References......Page 238
1. Introduction......Page 240
2. Methods......Page 245
3. Results and Discussion......Page 249
Appendix......Page 250
References......Page 251
1. Introduction......Page 252
2.1. Model architecture......Page 253
3.1. Decision conflict and retrospective confidence judgments......Page 255
3.2. Cognitive control through within-trial dynamic threshold setting......Page 257
3.3. Conflict-modulated learning of stimulus-response associations......Page 259
4. General discussion......Page 261
References......Page 262
1. Introduction......Page 264
2. Factors and independent processes......Page 265
3. A connectionist network for identifying ARMA-IPA models......Page 266
3.1. Identification of the AR predictive matrix......Page 267
3.2. Whitening......Page 269
3.4. Remark on ARMA processes of higher order......Page 270
4. Mapping the network onto the hippocampal region......Page 271
4.1.1. Co-learning in the parallel systems......Page 272
5. Discussion......Page 274
References......Page 275
1. What use are models?......Page 276
3.1. Sufficiency......Page 278
3.3. Existence proof......Page 279
4. Models as theories, theories as explanations......Page 280
5. A proposal......Page 282
6. Conclusions......Page 283
References......Page 284
Language, Learning and Development......Page 286
1. Introduction......Page 288
2. Role of Child-Directed Speech in Child Language Acquisition......Page 289
3. Related Computational Work......Page 290
4. Experimental Method......Page 291
5.1. Child-Directed Speech (CDS) Training Data......Page 292
6.1. One-Word Utterances from CDS Extra-Linguistic Data......Page 293
6.2. Emulating Alison’s One-Word Child Language......Page 296
7. Conclusion and Future Work......Page 297
References......Page 298
1. Introduction......Page 300
2. The Dual-path model......Page 301
3. Strong systematicity......Page 302
4. Recursive productivity......Page 304
5. The problem of auxiliary fronting in polar interrogatives......Page 306
6. Question learning in the Dual-path model......Page 307
References......Page 311
1. Shallow orthographies and models of reading......Page 312
2. Properties of Italian orthography and phonology......Page 314
3.1. Architecture and representation......Page 316
3.3. Training and testing......Page 317
4. Results......Page 318
4.1. Morphological effect......Page 319
4.2. Neighborhood effect......Page 320
Acknowledgments......Page 321
References......Page 322
1. Introduction......Page 324
1.2. Constructivist cognitive development......Page 326
1.3. The Constructivist neural network model......Page 327
2. Simulations......Page 328
3.1. Learning......Page 329
3.2. Overgeneralization and U-shaped learning......Page 330
3.3. Generalization to novel verbs......Page 331
3.4. Emergent modularity......Page 332
4. Discussion......Page 333
References......Page 334
1. Introduction......Page 336
2. Method......Page 338
2.1. The First Correction and Second Correction: An overview......Page 339
2.3. Second Correction: The role of frequent words missing from the CDI......Page 340
3.1. First Correction: Underestimation due to the absence of idiosyncratic words......Page 341
3.2. Second Correction: Evaluation of the role of omission of frequent words on CDIs......Page 344
4. Discussion......Page 346
References......Page 347
1. Theories of “embodied” and “disembodied” cognition......Page 348
2.1. Sensory/motor activity for conceptual tasks......Page 349
2.3. Sensory/motor impairment and conceptual tasks......Page 350
3. Computational models of embodied cognition......Page 352
4.2. Training and testing......Page 354
4.3. Results and discussion......Page 355
5. Discussion: Deciding between models......Page 356
References......Page 358
1.1. Age-related changes in plasticity......Page 360
1.2. Evidence for competition as a mechanism that mediates plasticity......Page 361
1.3. Computational modeling of changes in functional plasticity......Page 362
2.1. Basic model......Page 363
2.2. Simulation details......Page 364
2.3. Design......Page 366
3. Results......Page 367
4. Discussion......Page 370
References......Page 371
1. Introduction......Page 372
2. Auto-Teaching Neural Networks......Page 373
3. Simulations with Dynamic Environments......Page 375
4.1. Simulations With Full Set of Sensory Information......Page 377
4.2. Simulations With Limited Sensory Information......Page 378
4.3. Simulations With Multiple Learning Rates......Page 380
5. Discussion and Conclusions......Page 382
References......Page 383
Sensory Processing and Attention......Page 384
1. Introduction......Page 386
2.1. Spectral analysis......Page 388
2.2. Spatial analysis......Page 390
4.3. Procedure......Page 391
5.1. Network......Page 392
5.4. Results......Page 393
6. Conclusion......Page 394
References......Page 395
1. Introduction......Page 398
2. The Conditions for Conditioning......Page 400
3. Sustained Inattentional Blindness......Page 401
4. Building a Model......Page 403
4.1. Training......Page 404
4.3. Results and Analysis......Page 405
5. Experiment 2: Varying Similarity......Page 406
5.1. Results......Page 407
References......Page 408
1. Introduction......Page 412
1.1. Human visual search......Page 413
1.2. Modelling human search......Page 414
1.3. Linking the model to fMRI......Page 416
2. sSoTS architecture......Page 417
3.1. Extraction of activation maps for top down inhibition and excitation......Page 418
3.2. Comparison of fMRI data with model bold responses......Page 419
4. Results......Page 420
5. Conclusions......Page 423
References......Page 424
Author Index......Page 426