 |
GetTextbooks.com
Compare Prices & Save up to 90% |
|
|
|
| Login | Sign up | My Wish List |
 | | Engineering of Mind: An Introduction to the Science of Intelligent Systems by James S. Albus ISBN-10: 9780471438540 ISBN-10: 0-471-43854-5 ISBN-13: 9780471438540 ISBN-13: 978-0-471-43854-0 Hardcover 2001-09-07 Wiley-Interscience Find Lowest Price |
Editorials | ||
Product Description Presenting a reference model architecture for the design of intelligent systems Engineering of Mind presents the foundations for a computational theory of intelligence. It discusses the main streams of investigation that will eventually converge in a scientific theory of mind and proposes an avenue of research that might best lead to the development of truly intelligent systems. This book presents a model of the brain as a hierarchy of massive parallel computational modules and data structures interconnected by information pathways. Using this as the basic model on which intelligent systems should be based, the authors propose a reference model architecture that accommodates concepts from artificial intelligence, control theory, image understanding, signal processing, and decision theory. Algorithms, procedures, and data embedded within this architecture would enable the analysis of situations, the formulation of plans, the choice of behaviors, and the computation of uncertainties. The computational power to implement the model can be achieved in practical systems in the foreseeable future through hierarchical and parallel distribution of computational tasks. The authors' reference model architecture is expressed in terms of the Real-time Control System (RCS) that has been developed primarily at the National Institute of Standards and Technology. Suitable for engineers, computer scientists, researchers, and students, Engineering of Mind blends current theory and practice to achieve a coherent model for the design of intelligent systems. | ||
Reviews | ||
Some interesting proposals for building intelligent machines The authors of this book are very ambitious in their goals, for in the book they attempt to outline in some detail how to design a machine that is not only intelligent but can function in more than one domain. A machine that could not only beat the world champion at chess but could also create food recipes would be an example of a machine that can think in more than one domain. Certainly one would view such a machine as intelligent, at least from a qualitative standpoint, without having an explicit way of measuring intelligence. The authors actually define intelligence as an ability to behave appropriately in an uncertain environment, i.e. an entity that is intelligent will engage in behavior that maximizes the likelihood of success in the achievement of its goals. With this definition, they hope to capture the intelligence of many different entities, both biological and otherwise. Intelligence is not to be defined as an either/or proposition, but instead is graded, and there are many examples of it in systems in the real world, both natural and man-made. Intelligence can therefore exist in degrees, these being determined by the computational power and memory capacity of the brain of the machine, the processes the machine uses for obtaining data from the environment, and the quality of the data stored in memory. Different levels of intelligence, the authors believe, produce different probabilities of success in achieving goals. They do not give explicit examples of how these probabilities vary with the levels however. It would be very interesting to see these examples worked out in detail, for it would give the reader a more quantitative (and useful) notion of "machine IQ." One area of discussion that particularly stands out in the book deals with knowledge representation. Instead of representing everything symbolically via logical theorems, expert system rules, or linguistic grammars, the authors explicitly reject the symbol manipulation systems and instead want to represent knowledge using images or maps. This is interesting because of the high computational demands placed on a machine in performing image processing. The authors are aware of this, and so propose using parallel computation in the image and map domains combined with doing analysis in time and frequency space. This is to be done in a "multilevel architecture of dynamic recursive loops." Most of the book is devoted to explaining how to engineer this architecture. Heavy use is made of control theory to do this. The authors recognize that much remains unknown about the nature of intelligent behavior, but that the concept of a "goal" is central to ascribing intelligence to an entity. The greater the intelligence of an entity the more ingenious the entity is in dealing with unexpected events or challenges, and predicting the future is an ability possessed by the most intelligent entities. The rather extreme view of intelligence as expressed in this last statement is of course a logical consequence of the author's assertions. Needless to say, no example of an entity that can predict the future is given in the book. Since goal seeking is an essential characteristic of an intelligent machine, then the machine has to be able to make plans in order to reach its goals. The authors outline discuss two approaches to implementing a planning architecture. One of these is computationally expensive and involves frequently replanning, in order to deal with unpredictability of the environment. The other approach is to use feedback from sensors in order that the planned actions can be modified as needed. The authors outline several different heuristics that could be used to search for plans, and settle on a notion of `hierarchical multiresolutional planning'. This type of planning involves partitioning the planning process into hierarchical layers so that the search space is effectively reduced at each level of the planning hierarchy. In order for the authors to convince the reader that their efforts will be fruitful in designing an intelligent machine, they devote over two-thirds of the book to the real-world construction of such a machine. This is done by first discussing a reference model architecture, called RCS (for Real-Time Control System), that expresses their computational model of intelligence. The believe that the RCS architecture will allow the eventual design of intelligent machines that can meet specified requirements. Again, their project is very ambitious, due to the many capabilities that the RCS architecture must have. Without observing the machine actually working it would be difficult to verify whether or not they authors have succeeded in their goals. They do however give detailed explanations of the architecture, including line drawings and relevant mathematics, making their approach seem highly plausible. In addition, they give examples involving unmanned military ground vehicles that illustrate the principles they have outlined. The last chapter of the book is a look toward the future, and the authors, like all others who work in artificial intelligence, feel obligated to address the anxiety felt by some regarding the development of intelligent machines. Hollywood and popular literature is replete with examples of malevolent machines bent on the destruction of humankind or at least taking employment opportunities away from humans. The authors though remain refreshingly optimistic and rightfully dismiss these Hollywood/literary fantasies. They argue well for the productivity gains and positive social impact of intelligent machines. One can confidently look forward to the presence of millions of artificial intelligent machines in the twenty-first century, some of the architectures of which may be similar to the ones that are delineated in this book. | ||
A Profoundly Important Book Engineering of Mind is a fresh and exciting vision of the emerging science of machine intelligence. It begins with a readable introduction to the scientific analysis of the mind and brain. It then presents an overview of current research in the neurosciences, artificial intelligence, control theory, and robotics, and suggests an engineering approach to the design of intelligent machines. Finally, it addresses the scientific, economic, military, and human implications of a creating a new species of intelligent creatures. This is a seminal work of great importance. | ||