Atnaujinkite slapukų nuostatas

Circuits of the Mind [Minkštas viršelis]

4.43/5 (14 ratings by Goodreads)
(Gordon McKay Professor of Computer Science and Applied Mathematics, Harvard University)
  • Formatas: Paperback / softback, 256 pages, aukštis x plotis x storis: 156x234x14 mm, weight: 358 g
  • Išleidimo metai: 11-Jan-2001
  • Leidėjas: Oxford University Press Inc
  • ISBN-10: 0195126688
  • ISBN-13: 9780195126686
Kitos knygos pagal šią temą:
  • Formatas: Paperback / softback, 256 pages, aukštis x plotis x storis: 156x234x14 mm, weight: 358 g
  • Išleidimo metai: 11-Jan-2001
  • Leidėjas: Oxford University Press Inc
  • ISBN-10: 0195126688
  • ISBN-13: 9780195126686
Kitos knygos pagal šią temą:
In this groundbreaking work, computer scientist Leslie G. Valiant details a promising new computational approach to studying the intricate workings of the human brain. Focusing on the brain's enigmatic ability to access a massive store of accumulated information very quickly during reasoning processes, the author asks how such feats are possible given the extreme constraints imposed by the brain's finite number of neurons, their limited speed of communication, and their restricted interconnectivity. Valiant proposes a "neuroidal model" that serves as a vehicle to explore these fascinating questions.
While embracing the now classic theories of McCulloch and Pitts, the neuroidal model also accommodates state information in the neurons, more flexible timing mechanisms, a variety of assumptions about interconnectivity, and the possibility that different areas perform different functions. Programmable so that a wide range of algorithmic theories can be described and evaluated, the model provides a concrete computational language and a unified framework in which diverse cognitive phenomena--such as memory, learning, and reasoning--can be systematically and concurrently analyzed.
Included in this volume is a new preface that highlights some remarkable points of agreement between the neuroidal model and findings in neurobiology made since that model's original publication. Experiments have produced strong evidence for the theory's predictions about the existence of strong synapses in cortex and about the use of precise timing mechanisms within and between neurons. The theory also provides a quantitative explanation of how randomly placed neurons can be harnessed as resources for general purpose learning and memory--and is therefore synergistic with the striking recent discovery of neurogenesis in cortex.
Requiring no specialized knowledge, Circuits of the Mind, masterfully offers an exciting new approach to brain science for students and researchers in computer science, neurobiology, neuroscience, artificial intelligence, and cognitive science.

Recenzijos

The book is written in a clear style, with a sufficient number of figures illustrating the algorithms. . .This new insight into complex problems of the brain, as well as the proposed methodology, makes the book highly readable and interesting. * Computing Reviews * The author shows that the proposed neuroidal model supports the cognitive activities he identifies. It provides a good structure to explore the functions of the mind still further. * IIEEE Spectrum * Although there are many books today dealing with a simple neuronal model based on the weighted sum principle, this one rises above these others in providing an explanation of cognitive functions. * Choice * Delivers what its title promises, and more: an engaging, broad, thorough, and often deep, development of undergraduate complex analysis and related areas (non-Euclidean geometry, harmonic functions, etc.) from a geometric point of view. The style is lucid, informal, reader-friendly, and rich with helpful images (e.g., the complex derivative as an "amplitwist"). A truly unusual and notably creative look at a classical subject. * American Mathematical Monthly *

The Approach
1(8)
Biological Constraints
9(16)
Introduction
9(2)
The Neocortex
11(6)
Pyramidal Neurons
17(8)
Computational Laws
25(8)
Introduction
25(4)
Three Sources of Complexity
29(4)
Cognitive Functions
33(16)
Introduction
33(1)
Boolean Functions
34(4)
Learning
38(2)
The Nature of Concepts
40(3)
Experimental Psychology
43(6)
The Neuroidal Model
49(14)
Programmable Models
49(2)
Neuroids
51(9)
Timing
60(3)
Knowledge Representations
63(16)
Positive Knowledge Representations
63(3)
Vicinal Algorithms
66(3)
Frontier Properties and Storing New Items
69(4)
Frontier Properties and Associations
73(2)
Hashing
75(4)
Unsupervised Memorization
79(10)
An Algorithm
79(10)
Supervised Memorization
89(10)
Introduction
89(2)
A Simple Algorithm
91(3)
A Second Algorithm
94(5)
Supervised Inductive Learning
99(22)
Introduction
99(2)
Pac Learning
101(4)
Learning Conjunctions
105(5)
Learning Disjunctions
110(3)
Learning Linear Threshold Functions
113(8)
Correlational Learning
121(6)
An Algorithm
121(3)
Computing with Numerical Values
124(3)
Objects and Relational Expressions
127(18)
Multiple Object Scenes
127(2)
Relations
129(4)
Timed Conjunctions
133(3)
Memorizing Expressions Containing Relations
136(4)
Memorizing New Relations
140(1)
Discussion
141(4)
Systems Questions
145(12)
Introduction
145(1)
General Organizational Principles
146(6)
Compatibility of Mechanisms
152(5)
Reasoning
157(22)
Introduction
157(3)
Reflex Reasoning
160(1)
Simple Reflex Reasoning
161(8)
Compound Reflex Reasoning
169(3)
Nonmonotonic Phenomena
172(7)
More Detailed Neural Models
179(24)
Implementing Vicinal Algorithms
179(3)
A Laminar Model
182(10)
A Columnar Model
192(4)
Sparser Random Graphs
196(3)
Another Columnar Model
199(4)
Afterword
203(6)
Notes 209(6)
Exercises 215(4)
References 219(11)
Index of Notation 230(5)
Index 235