Natural and artificial intelligence : processor systems compared to the human brain

목차 (Table of Contents)

• CONTENTS
• INTRODUCTION = 27
• Integration of existing theories = 32
• Integration of servo-mechanism and symbolic processing models: = 32
• Integration of neuroscience and artificial intelligence: = 34
• Integration of behavioral and cognilive models: = 40
• Integration of procedural and logical processing = 43
• Integration of pure logic and common sense reasoning: = 44
• Other integrations: = 46
• Main features of the model: = 46
• Its all-or-none characteristic = 46
• Coding rule: "One activated wire per period per cable" = 47
• A manifold of cooperating devices = 49
• Apparent continuity through discrete operations = 50
• Management of instantiations = 51
• Retrieval of information = 51
• A single instruction multiple data machine (SIMD) = 53
• Representation of knowledge in logic programming: = 54
• Representation of knowledge in the central nervous system = 56
• Summary of the new features of the model: = 58
• Summary of the computer architecture of the brain model = 60
• Summary of the operationas of the brain model: = 60
• Plan of the book = 61
• Discussion of the main characteristics of the model = 63
• Rhythmic processing = 63
• Symbol instantiations assume phasic processing = 63
• Concrete description of rhythmic operations = 63
• 1) The NASA space shuttle = 63
• 2) A robot which learns to move its arm = 64
• 3) Studies of the equilibrium mechanism in man: = 65
• Definition of rhythmic: = 65
• Principles of rhythmic motor control: = 66
• Synchronization of hierarchical operations = 66
• Discussion of rhythmic processing = 66
• Symbolic processing = 68
• Discussion of the symbolic processing model = 70
• Additive database = 75
• Database machines = 75
• Additive database = 76
• Discussion of the size of the additive memory = 80
• Hierarchy of integrated mechanisms = 82
• The many types of hierarchy in the model = 82
• The hierarchical organization of cognitive processing = 83
• Is there a top area in the brain? = 84
• Self-programming method = 86
• Discussion of self-learned rules of behavior = 90
• Associated research areas = 94
• Link with neuroanatomy = 94
• Link with neuroscience = 94
• Link with knowledge based systems = 95
• Revived hypotheses: = 95
• The pure model and its real implementation = 96
• From the theory of knowledge to computer knowledge networks = 99
• The knowledge network discovered by Plato = 99
• Knowledge processing described by Aristotle = 100
• Knowledge and quantitative science = 102
• The stimulus response model: = 103
• Intentionaliy = 103
• Discovery of brain wiring = 104
• History of logical machines = 104
• Conventional computers: = 105
• First experiments with pointer organization: = 106
• To what extent can we understand the brain? = 107
• What is the meaning of simple in a scientific explanation? = 107
• Can we understand our own brain? = 108
• CHAPTER ONE : HARDWARE = 111
• Computers compared with the brain = 111
• Man made devices and naturally generated objects = 111
• Computer design principles missing in brain = 113
• Numbers: = 113
• Arithmetic processor: = 113
• Logical unit: = 113
• Ordered chains of information such as lists: = 113
• Fast instruction: = 114
• Coded temporal strings: = 114
• Registers for data or addresses: = 114
• Instruction code: = 114
• External progromming: = 114
• Stages in memory: = 114
• Program preparation: = 114
• Reliable circuits: = 114
• Brain hardware = 115
• Processing of approximate data: = 115
• Very large database: = 115
• Content addressability of memory: = 115
• Noise elimination: = 115
• Fast response to threatening stimuli: = 115
• One program for all brain modules = 115
• Introduction to computer architecture = 116
• Logical impulse = 117
• Gate = 117
• Gate properties = 117
• Gate with many potential states = 117
• Analog gale = 118
• Theoretical cell = 119
• Transmission circuits = 119
• AND / OR circuits = 121
• Theoretical cells = 121
• Classifiers = 122
• Content Addressable Memory = 123
• Dynamically addressable memory = 123
• 1) Current computer storage: = 124
• 2) Content addressable memory(CAM): = 124
• 3) Recording Content Addressable Memory: = 124
• Connectionnist hardware = 124
• Cable definition = 124
• Connectionnist hardware = 125
• Connectionnist memory = 125
• A simplified message flow processor = 125
• Inversion of memory location in physical pointers = 125
• Memory is better characterized by pointers than bits = 126
• Arithmetic logic unit(ALU) = 128
• Distinction between data and instructions = 128
• An ALU processes production rules = 128
• RCAM as a simplified ALU = 129
• Finite State Machine = 129
• Various definitions of "state" = 129
• Mark passing processors = 129
• Instantiation, marks and address registers = 130
• Description of a Finite Stale Machine(FSM) = 130
• Transition network = 131
• Information Retrieval = 131
• Database organization with inverted files = 131
• Searching for documents with inverted files = 131
• Addition of documents to inverted files = 131
• Concepts and production rules as documents = 132
• RCAM compared to inverted file = 132
• Retrieval with fuzzy sets = 133
• Searching multiple responses in a RCAM = 134
• An Additive Database : A memory which is never erased = 134
• Almost infinite storage for each entity: = 134
• Almost unlimited number of entities: = 135
• No restrictions on the use of pointers: = 135
• No garbage collection requirements: = 135
• No interruption for memory reorganization: = 135
• 24 hours availability: = 135
• Addition of redundant information: = 135
• Flexibility of infomation: = 136
• Retrieval time independent of database size: = 136
• Association with modifiable memory: = 136
• Very large database construction: = 136
• History of significant operactions: = 136
• Backtracking for restart: = 136
• Memory reclaim in business database: = 137
• Flow processor = 137
• Programs as transformation of state = 137
• Packets = 137
• Processing of packets = 137
• Database machines = 139
• Directed Network Processor Systems(DNPS) = 139
• Processor architecture with RCAMs = 139
• Speed comparison of DNPS and conventional computers = 141
• Silicon brain model = 143
• RCAM simulated by microprocessors = 143
• Cables simulated by wires with lemporal patterns = 145
• General purpose and specialized processors = 145
• Bidirectional tree machines = 146
• The grandmother neuron = 146
• A processing algorithm for behavior control and memorization = 148
• Implementation in RCAM hardware = 149
• Pattern Processors = 149
• RCAM specifications = 150
• Comparison of feature extractors(FE) and RCAM = 150
• Comparison of RCAM and perceptron = 152
• Comparison of RCAMs with associative memories = 153
• Pattern generalization = 153
• Event discrimination = 154
• RCAM design = 155
• Unrecordable events = 155
• Confusion risk = 155
• Quantified model of an RCAM = 155
• Hypotheses about RCAM size = 157
• Size of a brain area = 160
• DNPS design = 162
• Combination of controllers = 162
• The number of controllers in hierarchical models = 162
• Brain storage requirements = 163