Theoretical Foundations of Artificial General Intelligence

Theoretical Foundations of Artificial General Intelligence......Page 3
Contents......Page 5
1.1 The Matter of Artificial General Intelligence......Page 10
1.2 The Matter of Theoretical Foundation......Page 12
1.3 The Matter of Objective......Page 13
1.4 The Matter of Approach......Page 14
1.5 Challenges at the Heart of the Matter......Page 15
1.6 Summary......Page 16
Bibliography......Page 17
2.1.1 Naming the problem......Page 19
2.1.3 The State of the Science......Page 20
2.2 Existing Methods and Standards are not Sufficient......Page 21
2.2.2 Neuroscience......Page 22
2.2.3 Artificial intelligence......Page 23
2.3 CognitiveModeling: The Model Fit Imperative......Page 24
2.4 Artificial Intelligence and CognitiveModeling Can Help Each Other......Page 29
Bibliography......Page 32
3.1 Introduction......Page 33
3.2 More on Psychometric AGI......Page 36
3.2.1 Newell & the Neglected Route Toward General Machine Intelligence......Page 37
3.2.2 So, What is Psychometric AGI?......Page 38
3.2.3 Springboard to the Rest of the Present Paper......Page 41
3.3 Descartes’ Two Tests......Page 42
3.4 Piaget’s View of Thinking & The Magnet Test......Page 43
3.5 The LISA model......Page 46
3.6 Analogico-Deductive Reasoning in the Magnet Test......Page 47
3.7 Next Steps......Page 53
Bibliography......Page 54
4.1 Introduction......Page 57
4.2 Octopus Intelligence......Page 58
4.3 A “Ladder” of Intelligence......Page 61
4.4 Linguistic Grounding......Page 63
4.5 Implications of the Ladder for AGI......Page 65
4.6 Conclusion......Page 71
Bibliography......Page 72
5 One Decade of Universal Artificial Intelligence......Page 74
5.1 Introduction......Page 75
5.2 The AGI Problem......Page 77
5.3 Universal Artificial Intelligence......Page 81
5.4 Facets of Intelligence......Page 84
5.5 Social Questions......Page 86
5.6 State of the Art......Page 88
5.7 Discussion......Page 90
Bibliography......Page 91
6.1 Introduction: Decomposing the AGI Problem......Page 96
6.2.1 Overcoming the Curse of Dimensionality......Page 98
6.2.2 Spatiotemporal State Inference......Page 99
6.3.1 Deep Reinforcement Learning......Page 101
6.3.2 Actor-Critic Reinforcement Learning Themes in Cognitive Science......Page 104
6.4 Neuromorphic Devices Scaling AGI......Page 106
6.5 Conclusions and Outlook......Page 108
Bibliography......Page 109
7.1 Introduction......Page 110
7.2 Why the LIDA model may be suitable for AGI......Page 111
7.3 LIDA architecture......Page 112
7.4 Cognitive architectures, features and the LIDA model......Page 115
7.4.1 7.4.1 Ron Sun’s Desiderata [53......Page 116
7.4.2 Newell’s functional criteria (adapted from Lebiere and Anderson 2003)......Page 118
7.4.3 BICA table......Page 122
7.5 Discussion, Conclusions......Page 124
Bibliography......Page 125
8.1 Introduction......Page 129
8.2 Key Ingredients of the Integrative Human-Like Cognitive Architecture Diagram......Page 131
8.3 An Architecture Diagram for Human-Like General Intelligence......Page 134
8.4 Interpretation and Application of the Integrative Diagram......Page 142
8.5 Cognitive Synergy......Page 144
8.6 Why Is It So Hard to Measure Partial Progress Toward Human-Level AGI?......Page 146
8.7 Conclusion......Page 149
Bibliography......Page 150
9.1 Introduction......Page 151
9.2 The Nature of (General) Intelligence......Page 153
9.3 Constructionist AI: A Critical Look......Page 157
9.4 The Call for a New Methodology......Page 162
9.5 Towards a New Constructivist AI......Page 164
9.5.1 Temporal Grounding......Page 165
9.5.2 Feedback Loops......Page 167
9.5.3 Pan-Architectural Pattern Matching......Page 169
9.5.4 Transparent Operational Semantics......Page 170
9.6 Conclusions......Page 173
Bibliography......Page 175
10.1 Introduction......Page 178
10.2 The Gödel Machine Concept......Page 180
10.3 The Theoretical Foundations of Self-Reflective Systems......Page 183
10.3.1 Basic λ -calculus......Page 184
10.3.2 Constants, Conditionals, Side-effects, and Quoting......Page 188
10.4 Nested Meta-Circular Evaluators......Page 189
10.5 A Functional Self-Reflective System......Page 191
10.6 Discussion......Page 196
Bibliography......Page 199
11.1 Introduction......Page 201
11.2 Evaluating Flexibility......Page 204
11.2.1 The Testing Paradigm......Page 205
11.2.2 Combinatorial Difficulties of Superposition or Mixes......Page 206
11.2.3 “Occluding” Superpositions......Page 207
11.2.4 Counting Tests......Page 208
11.2.5 Binding Tests......Page 209
11.2.6 Binding and The Set-Cover Problem......Page 211
11.2.8 Scoring the Tests......Page 212
11.3 Evaluation of Flexibility......Page 213
11.3.2 Superpositions without loss......Page 215
11.3.3 Counting Tests......Page 216
11.3.6 Scoring Tests Together......Page 217
11.3.7 Conclusion from Tests......Page 218
11.4 Summary......Page 219
Bibliography......Page 220
12.1 Introduction......Page 222
12.2 Productivity and CognitiveMechanisms......Page 224
12.3 Cross-Domain Reasoning......Page 225
12.4 Basic Foundations of Theory Blending......Page 229
12.5 The Complex Plane: A Challenging Historical Example......Page 231
12.6 Outlook for Next Generation General Intelligent Systems......Page 236
Bibliography......Page 241
13.1 Introduction......Page 243
13.2 Emotion and affect......Page 246
13.3 Affective states emerging from cognitive modulation......Page 248
13.4 Higher-level emotions emerging from directing valenced affects......Page 252
13.5 Generating relevance: the motivational system......Page 254
13.6 Motive selection......Page 259
13.7 Putting it all together......Page 261
Bibliography......Page 263
14.1 Introduction......Page 265
14.2 Consciousness......Page 266
14.3 Machine Consciousness......Page 269
14.4 Agent’s Body......Page 272
14.5 Interactions with the Environment......Page 273
14.6 Time......Page 275
14.7 FreeWill......Page 276
14.8 Experience......Page 277
14.9 Creativity......Page 279
14.10 Conclusions......Page 281
Bibliography......Page 282
15.1 Introduction......Page 285
15.1.1 The Hard Problem of Consciousness......Page 286
15.1.2 A Problem within the Hard Problem......Page 288
15.1.3 An Outline of the Solution......Page 289
15.2.1 The Analysis Mechanism......Page 290
15.2.2 The Molecular Framework......Page 291
15.2.3 Explanation in General......Page 293
15.2.4 Explaining Subjective Concepts......Page 295
15.2.5 The “That Misses The Point” Objection......Page 298
15.3 The Real Meaning of Meaning......Page 299
15.3.1 Getting to the Bottom of Semantics......Page 300
15.3.2 Extreme Cognitive Semantics......Page 301
15.4 Some Falsifiable Predictions......Page 303
15.4.2 Prediction 2: New Qualia......Page 304
15.5 Conclusion......Page 305
Bibliography......Page 306
16.1 The problem of AI theory......Page 307
16.2 Nature and content of AI theories......Page 310
16.3 Desired properties of a theory......Page 314
16.4 Relations among the properties......Page 319
16.5 Issues on the properties......Page 320
16.6 Conclusion......Page 322
Bibliography......Page 323
Index......Page 326