This is the first book that renders a thorough discussion of systems science. It draws on material from an extensive collection of external sources, including several other books and a special library collection complete with videotape empirical evidence of applicability of the theory to a wide variety of circumstances. This is essential because systems science must be responsive to diverse human situations of the widest difficulty, and it must fill the void that the specific sciences cannot fill, because these sciences are insensitive to the necessities of reconciling disparate views of multiple observers, and incorporating local conditions in hypotheses that precede inductive explorations.

Machine Learning: An Artificial Intelligence Approach, Volume III presents a sample of machine learning research representative of the period between 1986 and 1989. The book is organized into six parts. Part One introduces some general issues in the field of machine learning. Part Two presents some new developments in the area of empirical learning methods, such as flexible learning concepts, the Protos learning apprentice system, and the WITT system, which implements a form of conceptual clustering. Part Three gives an account of various analytical learning methods and how analytic learning can be applied to various specific problems. Part Four describes efforts to integrate different learning strategies. These include the UNIMEM system, which empirically discovers similarities among examples; and the DISCIPLE multistrategy system, which is capable of learning with imperfect background knowledge. Part Five provides an overview of research in the area of subsymbolic learning methods. Part Six presents two types of formal approaches to machine learning. The first is an improvement over Mitchell's version space method; the second technique deals with the learning problem faced by a robot in an unfamiliar, deterministic, finite-state environment.

Advances in the quality and accessibility of computer graphics has provided new pictorial displays and the tools with which to control them. These new display technologies have focused interest on how to design the static and dynamic images they produce to ensure effective communication. This book, based on the conference on Spatial Displays and Spatial instruments held at the Asilomar Conference Centre in 1987, focuses on the geometry of this communication. It is intended to be a source book of theoretical analysis, experimental demonstrations and practical examples from a range of contributors interested in pictorial communication, from medical artists to astronauts. The book offers the theoretical background and practical guidance needed by designers of contemporary 2D and 3D graphical computer interfaces. Its major contribution lies in its outlining of the elements of human perception and motor control which underlie the geometric design of head-mounted graphics for virtual reality displays.

Fuzzy knowledge and fuzzy systems affect our lives today as systems enter the world of commerce. Fuzzy systems are incorporated in domestic appliances (washing machine, air conditioning, microwave, telephone) and in transport systems (a pilotless helicopter has recently completed a test flight). Future applications are expected to have dramatic implications for the demand for labor, among other things. It was with such thoughts in mind that this first international survey of future applications of fuzzy logic has been undertaken. The results are likely to be predictive for a decade beyond the millenium. The predictive element is combined with a bibliography which serves as an historical anchor as well as being both extensive and extremely useful. Analysis and Evaluation of Fuzzy Systems is thus a milestone in the development of fuzzy logic and applications of three representative subsystems: Fuzzy Control, Fuzzy Pattern Recognition and Fuzzy Communications.

Now, for the first time, publication of the landmark work inbackpropagation! Scientists, engineers, statisticians, operationsresearchers, and other investigators involved in neural networkshave long sought direct access to Paul Werbos's groundbreaking,much-cited 1974 Harvard doctoral thesis, The Roots ofBackpropagation, which laid the foundation of backpropagation. Now,with the publication of its full text, these practitioners can gostraight to the original material and gain a deeper, practicalunderstanding of this unique mathematical approach to socialstudies and related fields. In addition, Werbos has provided threemore recent research papers, which were inspired by his originalwork, and a new guide to the field. Originally written for readerswho lacked any knowledge of neural nets, The Roots ofBackpropagation firmly established both its historical andcontinuing significance as it: * Demonstrates the ongoing value and new potential ofbackpropagation * Creates a wealth of sound mathematical tools useful acrossdisciplines * Sets the stage for the emerging area of fast automaticdifferentiation * Describes new designs for forecasting and control which exploitbackpropagation * Unifies concepts from Freud, Jung, biologists, and others into anew mathematical picture of the human mind and how it works * Certifies the viability of Deutsch's model of nationalism as apredictive tool--as well as the utility of extensions of thiscentral paradigm "What a delight it was to see Paul Werbos rediscover Freud'sversion of 'back-propagation.' Freud was adamant (in The Projectfor a Scientific Psychology) that selective learning could onlytake place if the presynaptic neuron was as influenced as is thepostsynaptic neuron during excitation. Such activation of bothsides of the contact barrier (Freud's name for the synapse) wasaccomplished by reducing synaptic resistance by the absorption of'energy' at the synaptic membranes. Not bad for 1895! But Werbos1993 is even better." --Karl H. Pribram Professor Emeritus,Stanford University