This book provides the first broad yet thorough coverage of issues in morphological theory. It includes a wide array of techniques and systems in computational morphology (including discussion of their limitations), and describes some unusual applications.Sproat motivates the study of computational morphology by arguing that a computational natural language system, such as a parser or a generator, must incorporate a model of morphology. He discusses a range of applications for programs with knowledge of morphology, some of which are not generally found in the literature. Sproat then provides an overview of some of the basic descriptive facts about morphology and issues in theoretical morphology and (lexical) phonology, as well as psycholinguistic evidence for human processing of morphological structure. He take up the basic techniques that have been proposed for doing morphological processing and discusses at length various systems (such as DECOMP and KIMMO) that incorporate part or all of those techniques, pointing out the inadequacies of such systems from both a descriptive and a computational point of view. He concludes by touching on interesting peripheral areas such as the analysis of complex nominals in English, and on the main contributions of Rumelhart and McClelland's connectionism to the computational analysis of words.

Previous work on morphology has largely tended either to avoid precise computational details or to ignore linguistic generality. Computational Morphologyis the first book to present an integrated set of techniques for the rigorous description of morphological phenomena in English and similar languages. By taking account of all facets of morphological analysis, it provides a linguistically general and computationally practical dictionary system for use within an English parsing program. The authors covermorphographemics (variations in spelling as words are built from their component morphemes),morphotactics (the ways that different classes of morphemes can combine, and the types of words that result), andlexical redundancy (patterns of similarity and regularity among the lexical entries for words). They propose a precise rule-notation for each of these areas of linguistic description and present the algorithms for using these rules computationally to manipulate dictionary information. These mechanisms have been implemented in practical and publicly available software, which is described in detail, and appendixes contain a large number of computer-tested sets of rules and lexical entries for English. Graeme D. Ritchie is a Senior Lecturer in the Department of Artificial Intelligence at the University of Edinburgh, where Alan W. Black is currently a research student. Graham J. Russell is a Research Fellow at ISSCO (Institut Dalle Molle pour les etudes semantiques et cognitives) in Geneva, and Stephen G. Pulman is a Lecturer in the University of Cambridge Computer Laboratory and Director of SRI International's Cambridge Computer Science Research Centre.

TheInternationalSymposiumCreatingBrain-LikeIntelligencewasheldinFeb- ary 2007 in Germany. The symposium brought together notable scientists from di?erent backgrounds and with di?erent expertise related to the emerging ?eld of brain-like intelligence. Our understanding of the principles behind brain-like intelligence is still limited. After all, we have had to acknowledge that after tremendous advances in areas like neural networks, computational and arti?cial intelligence (a ?eld that had just celebrated its 50 year anniversary) and fuzzy systems, we are still not able to mimic even the lower-level sensory capabilities of humans or animals. We asked what the biggest obstacles are and how we could gain ground toward a scienti?c understanding of the autonomy, ?exibility, and robustness of intelligent biological systems as they strive to survive. New principles are usually found at the interfaces between existing disciplines, and traditional boundaries between disciplines have to be broken down to see how complex systems become simple and how the puzzle can be assembled. During the symposium we could identify some recurring themes that p- vaded many of the talks and discussions. The triad of structure, dynamics and environment,theroleoftheenvironmentasanactivepartnerinshapingsystems, adaptivity on all scales (learning, development, evolution) and the amalga- tion of an internal and external world in brain-like intelligence rate high among them. Each of us is rooted in a certain community which we have to serve with the results of our research. Looking beyond our ?elds and working at the interfaces between established areas of research requires e?ort and an active process.

Intelligence results from the interaction of the brain, body and environment. The question addressed in this book is, can we measure the contribution of the body and its' interaction with the environment? To answer this, we first present a comprehensive overview of the various ways in which a body reduces the amount of computation that the brain has to perform to solve a task. This chapter will broaden your understanding of how important inconspicuously appearing physical processes and physical properties of the body are with respect to our cognitive abilities. This form of contribution to intelligence is called Morphological Intelligence. The main contribution of this book to the field is a detailed discussion of how Morphological Intelligence can be measured from observations alone. The required mathematical framework is provided so that readers unfamiliar with information theory will be able to understand and apply the measures. Case studies from biomechanics and soft robotics illustrate how the presented quantifications can, for example, be used to measure the contribution of muscle physics to jumping and optimise the shape of a soft robotic hand. To summarise, this monograph presents various examples of how the physical properties of the body and the body’s interaction with the environment contribute to intelligence. Furthermore, it treats theoretical and practical aspects of Morphological Intelligence and demonstrates the value in two case studies.

By the late 1970s phonologists, and later morphologists, had departed from a linear approach for describing morphophonological operations to a nonlinear one. Computational models, however, remain faithful to the linear model, making it very difficult, if not impossible, to implement the morphology of languages whose morphology is nonconcatanative. Computational Nonlinear Morphology aims at presenting a computational system that counters the development in linguistics. It provides a detailed computational analysis of the complex morphophonological phenomena found in Semitic languages based on linguistically motivated models.

This is the first comprehensive overview of computational approaches to Arabic morphology. The subtitle aims to reflect that widely different computational approaches to the Arabic morphological system have been proposed. The book provides a showcase of the most advanced language technologies applied to one of the most vexing problems in linguistics. It covers knowledge-based and empirical-based approaches.

This book provides the first broad yet thorough coverage of issues in morphological theory. It includes a wide array of techniques and systems in computational morphology (including discussion of their limitations), and describes some unusual applications.Sproat motivates the study of computational morphology by arguing that a computational natural language system, such as a parser or a generator, must incorporate a model of morphology. He discusses a range of applications for programs with knowledge of morphology, some of which are not generally found in the literature. Sproat then provides an overview of some of the basic descriptive facts about morphology and issues in theoretical morphology and (lexical) phonology, as well as psycholinguistic evidence for human processing of morphological structure. He take up the basic techniques that have been proposed for doing morphological processing and discusses at length various systems (such as DECOMP and KIMMO) that incorporate part or all of those techniques, pointing out the inadequacies of such systems from both a descriptive and a computational point of view. He concludes by touching on interesting peripheral areas such as the analysis of complex nominals in English, and on the main contributions of Rumelhart and McClelland's connectionism to the computational analysis of words.