From Synapses to Rules

by Apolloni, Bruno; Kurfess, Franz

• ISBN13: 9780306474026
• ISBN10: 0306474026
• Format: Hardcover
• Copyright: 2002-12-01
• Publisher: Plenum Pub Corp

Summary

The book aims to propose a theoretical and applicatory framework for extracting formal rules from data. To this end recent approaches in relevant disciplines are examined that bring together two typical goals of conventional Artificial Intelligence and connectionism - respectively, deducing within an axiomatic shell formal rules about a phenomenon and inferring the actual behavior of it from examples - into a challenging inferential framework where we learn from data and understand what we have learned. The goal is to obtain a translation of the subsymbolic structure of the data - stored in the synapses of a neural network - into formal properties described by rules.To capture this journey from synapses to rules and then render it manageable for real world learning tasks, the contributions deal in depth with the following aspects: i. theoretical foundations of learning algorithms and soft computing; ii. intimate relationships between symbolic and subsymbolic reasoning methods; iii. integration of the related hosting architectures in both physiological and artificial brain.

Table of Contents

The Theoretical Bases of Learning	p. 1
The Statistical Bases on Learning	p. 5
PAC Meditation on Boolean Formulas	p. 41
Learning Regression Functions	p. 61
Cooperative Games in a Stochastic Environment	p. 75
If-Then-Else and Rule Extraction from Two Sets of Rules	p. 87
Extracting Interpretable Fuzzy Knowledge from Data	p. 109
Fuzzy Methods for Simplifying a Boolean Formula Inferred from Examples	p. 117
Physical Aspects of Learning	p. 135
On Mapping and Maps in the Central Nervous System	p. 139
Molecular Basis of Learning and Memory: Modelling Based on Receptor Mosaics	p. 165
Physiological and Logical Brain Functionalities: A Hypothesis for a Self-Referential Brain Activity	p. 197
Modeling of Spontaneous Bursting Activity Observed in In-Vitro Neural Networks	p. 219
The Importance of Data for Training Intelligent Devices	p. 229
Learning and Checking Confidence Regions for the Hazard Function of Biomedical Data	p. 251
Systems that Bridge the Gap	p. 273
Integrating Symbol-Oriented and Sub-Symbolic Reasoning Methods into Hybrid Systems	p. 275
From the Unconscious to the Conscious	p. 293
On Neural Networks, Connectionism and Brain-Like Learning	p. 315
Adaptive Computation in Data Structures and Webs	p. 343
IUANT: An Updating Method for Supervised Neural Structures	p. 363
Conclusions	p. 371
References	p. 372
Index	p. 385
Table of Contents provided by Blackwell. All Rights Reserved.