Bøger i The Springer International Series in Engineering and Computer Science serien

Change of Representation and Inductive Bias One of the most important emerging concerns of machine learning researchers is the dependence of their learning programs on the underlying representations, especially on the languages used to describe hypotheses. The effectiveness of learning algorithms is very sensitive to this choice of language; choosing too large a language permits too many possible hypotheses for a program to consider, precluding effective learning, but choosing too small a language can prohibit a program from being able to find acceptable hypotheses. This dependence is not just a pitfall, however; it is also an opportunity. The work of Saul Amarel over the past two decades has demonstrated the effectiveness of representational shift as a problem-solving technique. An increasing number of machine learning researchers are building programs that learn to alter their language to improve their effectiveness. At the Fourth Machine Learning Workshop held in June, 1987, at the University of California at Irvine, it became clear that the both the machine learning community and the number of topics it addresses had grown so large that the representation issue could not be discussed in sufficient depth. A number of attendees were particularly interested in the related topics of constructive induction, problem reformulation, representation selection, and multiple levels of abstraction. Rob Holte, Larry Rendell, and I decided to hold a workshop in 1988 to discuss these topics. To keep this workshop small, we decided that participation be by invitation only.

Real-Time Systems Engineering and Applications is a well-structured collection of chapters pertaining to present and future developments in real-time systems engineering. After an overview of real-time processing, theoretical foundations are presented. The book then introduces useful modeling concepts and tools. This is followed by concentration on the more practical aspects of real-time engineering with a thorough overview of the present state of the art, both in hardware and software, including related concepts in robotics. Examples are given of novel real-time applications which illustrate the present state of the art. The book concludes with a focus on future developments, giving direction for new research activities and an educational curriculum covering the subject. This book can be used as a source for academic and industrial researchers as well as a textbook for computing and engineering courses covering the topic of real-time systems engineering.

Computational Architectures Integrating Neural and Symbolic Processes: A Perspective on the State of the Art focuses on a currently emerging body of research. With the reemergence of neural networks in the 1980s with their emphasis on overcoming some of the limitations of symbolic AI, there is clearly a need to support some form of high-level symbolic processing in connectionist networks. As argued by many researchers, on both the symbolic AI and connectionist sides, many cognitive tasks, e.g. language understanding and common sense reasoning, seem to require high-level symbolic capabilities. How these capabilities are realized in connectionist networks is a difficult question and it constitutes the focus of this book. Computational Architectures Integrating Neural and Symbolic Processes addresses the underlying architectural aspects of the integration of neural and symbolic processes. In order to provide a basis for a deeper understanding of existing divergent approaches and provide insight for further developments in this field, this book presents: (1) an examination of specific architectures (grouped together according to their approaches), their strengths and weaknesses, why they work, and what they predict, and (2) a critique/comparison of these approaches. Computational Architectures Integrating Neural and Symbolic Processes is of interest to researchers, graduate students, and interested laymen, in areas such as cognitive science, artificial intelligence, computer science, cognitive psychology, and neurocomputing, in keeping up-to-date with the newest research trends. It is a comprehensive, in-depth introduction to this new emerging field.

The rapid development of wireless digital communication technology has cre- ated capabilities that software systems are only beginning to exploit. The falling cost of both communication and of mobile computing devices (laptop computers, hand-held computers, etc. ) is making wireless computing affordable not only to business users but also to consumers. Mobile computing is not a "e;scaled-down"e; version of the established and we- studied field of distributed computing. The nature of wireless communication media and the mobility of computers combine to create fundamentally new problems in networking, operating systems, and information systems. Further- more, many of the applications envisioned for mobile computing place novel demands on software systems. Although mobile computing is still in its infancy, some basic concepts have been identified and several seminal experimental systems developed. This book includes a set of contributed papers that describe these concepts and sys- tems. Other papers describe applications that are currently being deployed and tested. The first chapter offers an introduction to the field of mobile computing, a survey of technical issues, and a summary of the papers that comprise sub- sequent chapters. We have chosen to reprint several key papers that appeared previously in conference proceedings. Many of the papers in this book are be- ing published here for the first time. Of these new papers, some are expanded versions of papers first presented at the NSF-sponsored Mobidata Workshop on Mobile and Wireless Information Systems, held at Rutgers University on Oct 31 and Nov 1, 1994.

Real-time systems are now used in a wide variety of applications. Conventionally, they were configured at design to perform a given set of tasks and could not readily adapt to dynamic situations. The concept of imprecise and approximate computation has emerged as a promising approach to providing scheduling flexibility and enhanced dependability in dynamic real-time systems. The concept can be utilized in a wide variety of applications, including signal processing, machine vision, databases, networking, etc. For those who wish to build dynamic real-time systems which must deal safely with resource unavailability while continuing to operate, leading to situations where computations may not be carried through to completion, the techniques of imprecise and approximate computation facilitate the generation of partial results that may enable the system to operate safely and avert catastrophe. Audience: Of special interest to researchers. May be used as a supplementary text in courses on real-time systems.

Privacy, Security and Trust within the Context of Pervasive Computing is an edited volume based on a post workshop at the second international conference on Pervasive Computing. The workshop was held April18-23, 2004, in Vienna, Austria.The goal of the workshop was not to focus on specific, even novel mechanisms, but rather on the interfaces between mechanisms in different technical and social problem spaces. An investigation of the interfaces between the notions of context, privacy, security, and trust will result in a deeper understanding of the "e;atomic"e; problems, leading to a more complete understanding of the social and technical issues in pervasive computing.

Broadband opamps for multi-channel communication systems have strong demands on linearity performance. When these opamps are integrated in deep sub-micron CMOS technologies, the signal-swing has to occupy a large part of the rather low supply voltage to maintain the signal-to-noise-ratio. To obtain opamps with low distortion it is necessary to do a thorough analysis of the nonlinear behaviour of such circuits and this is the main subject of Design Criteria for Low Distortion in Feedback Opamp Circuits. The biasing of each transistor in the circuit is a major issue and is addressed in this work. It is important to bias the transistor such that the distortion is low and stable in the entire range of its terminal voltages. This will ensure high linearity and robustness against variations in circuit conditions such as power supply voltage, bias current and process variations. Design Criteria for Low Distortion in Feedback Opamp Circuits is written for .

Silicon on Insulator is more than a technology, more than a job, and more than a venture in microelectronics; it is something different and refreshing in device physics. This book recalls the activity and enthu- siasm of our SOl groups. Many contributing students have since then disappeared from the SOl horizon. Some of them believed that SOl was the great love of their scientific lives; others just considered SOl as a fantastic LEGO game for adults. We thank them all for kindly letting us imagine that we were guiding them. This book was very necessary to many people. SOl engineers will certainly be happy: indeed, if the performance of their SOl components is not always outstanding, they can now safely incriminate the relations given in the book rather than their process. Martine, Gunter, and Y. S. Chang can contemplate at last the amount of work they did with the figures. Our SOl accomplices already know how much we borrowed from their expertise and would find it indecent to have their detailed contri- butions listed. Jean-Pierre and Dimitris incited the book, while sharing their experience in the reliability of floating bodies. Our families and friends now realize the SOl capability of dielectrically isolating us for about two years in a BOX. Our kids encouraged us to start writing. Our wives definitely gave us the courage to stop writing. They had a hard time fighting the symptoms of a rapidly developing SOl allergy.

CMOS Current Amplifiers presents design strategies for high performance current amplifiers based on CMOS technology. After an introduction to various architectures of operational amplifiers, the operating principles of the current amplifier are outlined. This book provides the reader with simple and compact design equations for use in a pencil and paper design and the following simulation step. Chapter 1 introduces the general aspects of current amplifiers. After a preliminary classification of operational amplifiers, ideal blocks and models are discussed for different architectures and a first high-level comparison is made between traditional amplifiers and current amplifiers. Analysis and examples of basic circuits, as well as signal processing applications involving current amplifiers, are also given. Non-idealities and second- order effects causing limitations in performance are then discussed and evaluated. Chapter 2 focuses on low-drive current amplifiers. Several design examples for current conveyors and class A current amplifiers are discussed in detail and design equations are presented for the main performance parameters, which allows a good trade-off between requirements. High-performance solutions for high bandwidth and low voltage capability are also considered, and, finally, current comparators with progressively enhanced performance are reported and analyzed critically. Chapter 3 deals with current amplifiers for off-chip loads. Several class AB current-mode output stages are discussed and design strategies which improve performance are presented. A detailed analysis of non-ideal effect is carried out with particular emphasis on linearity. Design examples are given and circuit arrangements for further developments are included. CMOS Current Amplifiers serves as an excellent reference for researchers and professionals of analog IC design, and may also be used as an advanced text on current amplifiers.

This book is a revision and extension of my 1995 Sourcebook of Control Systems Engineering. Because of the extensions and other modifications, it has been retitled Handbook of Control Systems Engineering, which it is intended to be for its prime audience: advanced undergraduate students, beginning graduate students, and practising engineers needing an understandable review of the field or recent developments which may prove useful. There are several differences between this edition and the first. * Two new chapters on aspects of nonlinear systems have been incorporated. In the first of these, selected material for nonlinear systems is concentrated on four aspects: showing the value of certain linear controllers, arguing the suitability of algebraic linearization, reviewing the semi-classical methods of harmonic balance, and introducing the nonlinear change of variable technique known as feedback linearization. In the second chapter, the topic of variable structure control, often with sliding mode, is introduced. * Another new chapter introduces discrete event systems, including several approaches to their analysis. * The chapters on robust control and intelligent control have been extensively revised. * Modest revisions and extensions have also been made to other chapters, often to incorporate extensions to nonlinear systems.

Neural Network Parallel Computing is the first book available to the professional market on neural network computing for optimization problems. This introductory book is not only for the novice reader, but for experts in a variety of areas including parallel computing, neural network computing, computer science, communications, graph theory, computer aided design for VLSI circuits, molecular biology, management science, and operations research. The goal of the book is to facilitate an understanding as to the uses of neural network models in real-world applications. Neural Network Parallel Computing presents a major breakthrough in science and a variety of engineering fields. The computational power of neural network computing is demonstrated by solving numerous problems such as N-queen, crossbar switch scheduling, four-coloring and k-colorability, graph planarization and channel routing, RNA secondary structure prediction, knight's tour, spare allocation, sorting and searching, and tiling. Neural Network Parallel Computing is an excellent reference for researchers in all areas covered by the book. Furthermore, the text may be used in a senior or graduate level course on the topic.

Since the early 1980s, CAD frameworks have received a great deal of attention, both in the research community and in the commercial arena. It is generally agreed that CAD framework technology promises much: advanced CAD frameworks can turn collections of individual tools into effective and user-friendly design environments. But how can this promise be fulfilled? CAD Frameworks: Principles and Architecture describes the design and construction of CAD frameworks. It presents principles for building integrated design environments and shows how a CAD framework can be based on these principles. It derives the architecture of a CAD framework in a systematic way, using well-defined primitives for representation. This architecture defines how the many different framework sub-topics, ranging from concurrency control to design flow management, relate to each other and come together into an overall system. The origin of this work is the research and development performed in the context of the Nelsis CAD Framework, which has been a working system for well over eight years, gaining functionality while evolving from one release to the next. The principles and concepts presented in this book have been field-tested in the Nelsis CAD Framework. CAD Frameworks: Principles and Architecture is primarily intended for EDA professionals, both in industry and in academia, but is also valuable outside the domain of electronic design. Many of the principles and concepts presented are also applicable to other design-oriented application domains, such as mechanical design or computer-aided software engineering (CASE). It is thus a valuable reference for all those involved in computer-aided design.

Load Balancing in Parallel Computers: Theory and Practice is about the essential software technique of load balancing in distributed memory message-passing parallel computers, also called multicomputers. Each processor has its own address space and has to communicate with other processors by message passing. In general, a direct, point-to-point interconnection network is used for the communications. Many commercial parallel computers are of this class, including the Intel Paragon, the Thinking Machine CM-5, and the IBM SP2. Load Balancing in Parallel Computers: Theory and Practice presents a comprehensive treatment of the subject using rigorous mathematical analyses and practical implementations. The focus is on nearest-neighbor load balancing methods in which every processor at every step is restricted to balancing its workload with its direct neighbours only. Nearest-neighbor methods are iterative in nature because a global balanced state can be reached through processors' successive local operations. Since nearest-neighbor methods have a relatively relaxed requirement for the spread of local load information across the system, they are flexible in terms of allowing one to control the balancing quality, effective for preserving communication locality, and can be easily scaled in parallel computers with a direct communication network. Load Balancing in Parallel Computers: Theory and Practice serves as an excellent reference source and may be used as a text for advanced courses on the subject.

This book is based on the author's Ph.D. dissertation[56]. The the- sis research was conducted while the author was a graduate student in the Department of Computer Science at Rutgers University. The book was pre- pared at the University of Massachusetts at Amherst where the author is currently an Assistant Professor in the Department of Computer and Infor- mation Science. Programs that learn concepts from examples are guided not only by the examples (and counterexamples) that they observe, but also by bias that determines which concept is to be considered as following best from the ob- servations. Selection of a concept represents an inductive leap because the concept then indicates the classification of instances that have not yet been observed by the learning program. Learning programs that make undesir- able inductive leaps do so due to undesirable bias. The research problem addressed here is to show how a learning program can learn a desirable inductive bias.

This book is concerned with advances in serial-data computa- tional architectures, and the CAD tools for their implementation in silicon. The bit-serial tradition at Edinburgh University (EU) stretches back some 6 years to the conception of the FIRST silicon compiler. FIRST owes much of its inspiration to Dick Lyon, then at Xerox P ARC, who proposed a 'structured-design' methodology for construction of signal processing systems from bit-serial building blocks. Based on an nMOS cell-library, FIRST automates much of Lyon's physical design process. More recently, we began to feel that FIRST should be able to exploit more modern technologies. Before this could be achieved, we were faced with a massive manual re-design task, i. e. the porting of FIRST cell-library to a new technology. As it was to avoid such tasks that FIRST was conceived in the first place, we decided to move the level of user-specification much nearer to the silicon level (while still hiding details of transistor circuit design, place and route etc. , from the user), and by so doing, enable the specification of more functionally powerful libraries in technology-free form. The results of this work are in evidence as advances in serial-data design techniques, and the SECOND silicon compiler, introduced later in this book. These achievements could not have been accomplished without help from various sources. We take this opportunity to thank Profs.

One suspects that the people who use computers for their livelihood are growing more "e;sophisticated"e; as the field of computer science evolves. This view might be defended by the expanding use of languages such as C and Lisp in contrast to the languages such as FORTRAN and COBOL. This hypothesis is false however - computer languages are not like natural languages where successive generations stick with the language of their ancestors. Computer programmers do not grow more sophisticated - programmers simply take the time to muddle through the increasingly complex language semantics in an attempt to write useful programs. Of course, these programmers are "e;sophisticated"e; in the same sense as are hackers of MockLisp, PostScript, and Tex - highly specialized and tedious languages. It is quite frustrating how this myth of sophistication is propagated by some industries, universities, and government agencies. When I was an undergraduate at MIT, I distinctly remember the convoluted questions on exams concerning dynamic scoping in Lisp - the emphasis was placed solely on a "e;hacker's"e; view of computation, i. e. , the control and manipulation of storage cells. No consideration was given to the logical structure of programs. Within the past five years, Ada and Common Lisp have become programming language standards, despite their complexity (note that dynamic scoping was dropped even from Common Lisp). Of course, most industries' selection of programming languages are primarily driven by the requirement for compatibility (with previous software) and performance.

From little more than a circuit-theoretical concept in 1965, computer-aided circuit simulation developed into an essential and routinely used design tool in less than ten years. In 1965 it was costly and time consuming to analyze circuits consisting of a half-dozen transistors. By 1975 circuits composed of hundreds of transistors were analyzed routinely. Today, simulation capabilities easily extend to thousands of transistors. Circuit designers use simulation as routinely as they used to use a slide rule and almost as easily as they now use hand-held calculators. However, just as with the slide rule or hand-held calculator, some designers are found to use circuit simulation more effectively than others. They ask better questions, do fewer analyses, and get better answers. In general, they are more effective in using circuit simulation as a design tool. Why? Certainly, design experience, skill, intuition, and even luck contribute to a designer's effectiveness. At the same time those who design and develop circuit simulation programs would like to believe that their programs are so easy and straightforward to use, so well debugged and so efficient that even their own grandmother could design effectively using their program.