Recent research results in the area of parallel algorithms for problem solving, search, natural language parsing, and computer vision, are brought together in this book. The research reported demonstrates that substantial parallelism can be exploited in various machine intelligence and vision problems. The chapter authors are prominent researchers actively involved in the study of parallel algorithms for machine intelligence and vision. Extensive experimental studies are presented that will help the reader in assessing the usefulness of an approach to a specific problem. Intended for students and researchers actively involved in parallel algorithms design and in machine intelligence and vision, this book will serve as a valuable reference work as well as an introduction to several research directions in these areas.
Parallel processing for AI problems is of great current interest because of its potential for alleviating the computational demands of AI procedures. The articles in this book consider parallel processing for problems in several areas of artificial intelligence: image processing, knowledge representation in semantic networks, production rules, mechanization of logic, constraint satisfaction, parsing of natural language, data filtering and data mining. The publication is divided into six sections. The first addresses parallel computing for processing and understanding images. The second discusses parallel processing for semantic networks, which are widely used means for representing knowledge - methods which enable efficient and flexible processing of semantic networks are expected to have high utility for building large-scale knowledge-based systems. The third section explores the automatic parallel execution of production systems, which are used extensively in building rule-based expert systems - systems containing large numbers of rules are slow to execute and can significantly benefit from automatic parallel execution. The exploitation of parallelism for the mechanization of logic is dealt with in the fourth section. While sequential control aspects pose problems for the parallelization of production systems, logic has a purely declarative interpretation which does not demand a particular evaluation strategy. In this area, therefore, very large search spaces provide significant potential for parallelism. In particular, this is true for automated theorem proving. The fifth section considers the problem of constraint satisfaction, which is a useful abstraction of a number of important problems in AI and other fields of computer science. It also discusses the technique of consistent labeling as a preprocessing step in the constraint satisfaction problem. Section VI consists of two articles, each on a different, important topic. The first discusses parallel formulation for the Tree Adjoining Grammar (TAG), which is a powerful formalism for describing natural languages. The second examines the suitability of a parallel programming paradigm called Linda, for solving problems in artificial intelligence. Each of the areas discussed in the book holds many open problems, but it is believed that parallel processing will form a key ingredient in achieving at least partial solutions. It is hoped that the contributions, sourced from experts around the world, will inspire readers to take on these challenging areas of inquiry.
The third in an informal series of books about parallel processing for Artificial Intelligence, this volume is based on the assumption that the computational demands of many AI tasks can be better served by parallel architectures than by the currently popular workstations. However, no assumption is made about the kind of parallelism to be used. Transputers, Connection Machines, farms of workstations, Cellular Neural Networks, Crays, and other hardware paradigms of parallelism are used by the authors of this collection. The papers arise from the areas of parallel knowledge representation, neural modeling, parallel non-monotonic reasoning, search and partitioning, constraint satisfaction, theorem proving, parallel decision trees, parallel programming languages and low-level computer vision. The final paper is an experience report about applications of massive parallelism which can be said to capture the spirit of a whole period of computing history. This volume provides the reader with a snapshot of the state of the art in Parallel Processing for Artificial Intelligence.
Computer vision is one of the most complex and computationally intensive problem. Like any other computationally intensive problems, parallel pro cessing has been suggested as an approach to solving the problems in com puter vision. Computer vision employs algorithms from a wide range of areas such as image and signal processing, advanced mathematics, graph theory, databases and artificial intelligence. Hence, not only are the comput ing requirements for solving vision problems tremendous but they also demand computers that are efficient to solve problems exhibiting vastly dif ferent characteristics. With recent advances in VLSI design technology, Single Instruction Multiple Data (SIMD) massively parallel computers have been proposed and built. However, such architectures have been shown to be useful for solving a very limited subset of the problems in vision. Specifically, algorithms from low level vision that involve computations closely mimicking the architec ture and require simple control and computations are suitable for massively parallel SIMD computers. An Integrated Vision System (IVS) involves com putations from low to high level vision to be executed in a systematic fashion and repeatedly. The interaction between computations and information dependent nature of the computations suggests that architectural require ments for computer vision systems can not be satisfied by massively parallel SIMD computers.
This book covers parallel algorithms and architectures and VLSI chips for a range of problems in image processing, computer vision, pattern recognition and artificial intelligence. The specific problems addressed include vision and image processing tasks, Fast Fourier Transforms, Hough Transforms, Discrete Cosine Transforms, image compression, polygon matching, template matching, pattern matching, fuzzy expert systems and image rotation. The collection of papers gives the reader a good introduction to the state-of-the-art, while for an expert this serves as a good reference and a source of some new contributions in this field.
This contributed volume highlights two areas of fundamental interest in high-performance computing: core algorithms for important kernels and computationally demanding applications. The first few chapters explore algorithms, numerical techniques, and their parallel formulations for a variety of kernels that arise in applications. The rest of the volume focuses on state-of-the-art applications from diverse domains. By structuring the volume around these two areas, it presents a comprehensive view of the application landscape for high-performance computing, while also enabling readers to develop new applications using the kernels. Readers will learn how to choose the most suitable parallel algorithms for any given application, ensuring that theory and practicality are clearly connected. Applications using these techniques are illustrated in detail, including: Computational materials science and engineering Computational cardiovascular analysis Multiscale analysis of wind turbines and turbomachinery Weather forecasting Machine learning techniques Parallel Algorithms in Computational Science and Engineering will be an ideal reference for applied mathematicians, engineers, computer scientists, and other researchers who utilize high-performance computing in their work.
It has been widely recognized that artificial intelligence computations offer large potential for distributed and parallel processing. Unfortunately, not much is known about designing parallel AI algorithms and efficient, easy-to-use parallel computer architectures for AI applications. The field of parallel computation and computers for AI is in its infancy, but some significant ideas have appeared and initial practical experience has become available. The purpose of this book has been to collect in one volume contributions from several leading researchers and pioneers of AI that represent a sample of these ideas and experiences. This sample does not include all schools of thought nor contributions from all leading researchers, but it covers a relatively wide variety of views and topics and in this sense can be helpful in assessing the state ofthe art. We hope that the book will serve, at least, as a pointer to more specialized literature and that it will stimulate interest in the area of parallel AI processing. It has been a great pleasure and a privilege to cooperate with all contributors to this volume. They have my warmest thanks and gratitude. Mrs. Birgitta Knapp has assisted me in the editorial task and demonstrated a great deal of skill and patience. Janusz S. Kowalik vii INTRODUCTION Artificial intelligence (AI) computer programs can be very time-consuming.
This comprehensive book is primarily intended for researchers, computer vision specialists, and high-performance computing specialists who are interested in parallelizing computer vision techniques for the sake of accelerating the run-time of computer vision methods. This book covers different penalization methods on different parallel architectures such as multi-core CPUs and GPUs. It is also a valuable reference resource for researchers at all levels (e.g., undergraduate and postgraduate) who are seeking real-life examples of speeding up the computer vision methods’ run-time.
This integrated collection covers a range of parallelization platforms, concurrent programming frameworks and machine learning settings, with case studies.
