No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Conclusion
Abstract
Demand for MT has grown and will continue to grow steadily. The advent of the Internet and electronic commerce in particular have increased the demand for automatic translation of sufficient quality for determining the content of a Web page. A variety of authoring tools and document production techniques have also made linguistic information available in a variety of formats. This information will need to be translated if it is to have the widest possible distribution.
ResearchGate has not been able to resolve any citations for this publication.
This is a book notice that appeared in the journal Language
Unification-based grammar formalisms use feature structures to represent linguistic knowledge. The only operation defined on feature structures, unification, is information-combining and monotonic. Several authors have proposed nonmonotonic extensions of this formalism, as for a linguistically adequate description of certain natural language phenomena some kind of default reasoning seems essential. We argue that the effect of these proposals can be captured by means of one general, nonmonotonic, operation on feature structures, called default unification. We provide a formal semantics of the operation and demonstrate how some of the phenomena used to motivate nonmonotonic extensions of unification-based formalisms can be handled.
To translate collocations such as prendre une décision, prendre l'habitude, perdre l'habitude, avoir soif, most MT systems include bilingual rules as prendre ( — décision ) --> make, prendre ( — habitude ) --> get into, perdre ( — habitude ) --> get out of, soif (avoir — ) --> thirsty (be — ). These numerous rules, which are lexically context-sensitive, are costly, be they used in a rationalist or empiricist approach. This paper presents a solution where they are avoided. This solution is based on linguistics notions such as "predicative element", "aspectual variant", "diathetic variant" which allow us to use only the unavoidable simple bilingual rules such as décision --> decision, habitude -•> habit, thirsty --> soif. This solution, which has been successfully implemented in a rule-based transfer system, i.e. the Eurotra system which handles nine languages, requires a lot of monolingual lexical information, but we will show that this lexical information should be easily extracted from tagged corpora.
This textbook provides a thorough and precise account of all the major areas of English grammar. For practical reasons the author concentrates on Standard English and only selected aspects of its regional variation. The book is written for students who may have no previous knowledge of linguistics and little familiarity with 'traditional' grammar. All grammatical terms, whether traditional or more recent, are therefore carefully explained, and in the first three chapters the students is introduced to the theoretical concepts and methodological principles needed to follow the later descriptive chapters. Nevertheless, the book is more than a straightforward 'grammar of English'. Rodney Huddleston does not espouse any formalised contemporary model of syntax and morphology, but he adopts the framework of modern 'structural' linguistics, in a very broad understanding of that term. The grammatical categories postulated derive from a study of the combinational and contrastive relationships the words and other forms enter into, and Dr Huddleston pays particular attention to the problem of choosing between alternative analyses and justifying the analysis he proposes. In this sense his book is addressed to the student of linguistics, who will find Introduction to the Grammar of English a much needed foundation for more advanced work in theoretical linguistics.
This classic research monograph develops and illustrates the theory of linguistic structure known as Cognitive Grammar, and applies it to representative phenomena in English and other languages. Cognitive grammar views language as an integral facet of cognition and claims that grammatical structure cannot be understood or revealingly described independently of semantic considerations.
Semantic underspecification is an essential and pervasive property of natural language. This monograph provides a comprehensive survey of the various phenomena in the field of ambiguity and vagueness. The book discusses the major theories of semantic indefiniteness, which have been proposed in linguistics, philosophy and computer science. It argues for a view of indefiniteness as the potential for further contextual specification, and proposes a unified logical treatment of indefiniteness on this basis. The inherent inconsistency of natural language induced by irreducible imprecision is investigated, and treated in terms of a dynamic extension of the proposed logic.
The book is an extended edition of a German monograph and is addressed to advanced students and researchers in theoretical and computational linguistics, logic, philosophy of language, and NL- oriented AI. Although it makes extensive use of logical formalisms, it requires only some basic familiarity with standard predicate logic concepts since all technical terms are carefully explained.
Castel Ivano, originally built in 1375, is one of many beautiful and impressive castles strategically placed atop hills in Trentino's Valsugana in Northern Italy. It was in this castle on a series of brilliant sunny crisp November days in 1990 that an international group of computer scientists and cognitive scientists met at a workshop to discuss theoretical and applied issues concerning communi cation from an Artificial Intelligence and Cognitive Science perspective. About forty people, representing nine countries, participated in the workshop, either as speakers, discussants, or observers. The main motivationfor the workshop wasto address the questionofwhether and how current computational approaches to communication can or might be able to accommodate the range of complexities that characterize both human human and human-machine communication. The chapters in this book are based on the papers that were presented at the workshop. They are presented in an order that is determined primarily by the specificity of the topics they address. The initial chapters are more theoretical in nature with an emphasis on formal approaches to communication. The middle chapters focus on particular application issues, such as the generation ofmultimedia documents and the role of planning in building systems to support human-human or human-machine interaction. The final few chapters consider more general issues relating to com munication, such as the influence ofsocial structure on, and the role of affect in communication.
The now-classic Metaphors We Live By changed our understanding of metaphor and its role in language and the mind. Metaphor, the authors explain, is a fundamental mechanism of mind, one that allows us to use what we know about our physical and social experience to provide understanding of countless other subjects. Because such metaphors structure our most basic understandings of our experience, they are "metaphors we live by"--metaphors that can shape our perceptions and actions without our ever noticing them.
In this updated edition of Lakoff and Johnson's influential book, the authors supply an afterword surveying how their theory of metaphor has developed within the cognitive sciences to become central to the contemporary understanding of how we think and how we express our thoughts in language.
From the Publisher:
Eugene Charniak breaks new ground in artificial intelligence research by presenting statistical language processing from an artificial intelligence point of view in a text for researchers and scientists with a traditional computer science background.
New, exacting empirical methods are needed to break the deadlock in such areas of artificial intelligence as robotics, knowledge representation, machine learning, machine translation, and natural language processing (NLP). It is time, Charniak observes, to switch paradigms. This text introduces statistical language processing techniques -- word tagging, parsing with probabilistic context free grammars, grammar induction, syntactic disambiguation, semantic word classes, word-sense disambiguation -- along with the underlying mathematics and chapter exercises.
Charniak points out that as a method of attacking NLP problems, the statistical approach has several advantages. It is grounded in real text and therefore promises to produce usable results, and it offers an obvious way to approach learning: "one simply gathers statistics."
Language, Speech, and Communication
This work presents a detailed description of modern Spanish, drawing on conversational and written examples taken from both the old and the new worlds. It takes into account the growing prestige of Latin-American Spanish as well as Iberian Spanish.
A parsing algorithm which seems to be the most efficient general context-free algorithm known is described. It is similar to both Knuth's LR(k) algorithm and the familiar top-down algorithm. It has a time bound proportional to n3 (where n is the length of the string being parsed) in general; it has an n2 bound for unambiguous grammars; and it runs in linear time on a large class of grammars, which seems to include most practical context-free programming language grammars. In an empirical comparison it appears to be superior to the top-down and bottom-up algorithms studied by Griffiths and Petrick. © 1983, ACM. All rights reserved.
We provide a tripartite overview of HPSG, a constraint-based model-theoretic framework for grammatical representation widely used for both theoretical research and computational implementations of natural language grammars. We begin with a condensed survey of the descriptive machinery assumed in the theory, focusing on the feature-based encoding of grammatical properties and the form of constraints regulating the specification of values for these features in complex syntatic structures, with illustrations from various local and nonlocal dependencies exhibited in a range of grammatical constructions. The following section explores in some detail the logical and formal foundations of the descriptive apparatus, centered on the crucial ontological distinction between, on the one hand, formal objects embodying the properties of linguistic expressions, and on the other, descriptions of those objects. We conclude with a review of HPSG-based processing, including the motivations behind such work, the different computational approaches and systems, as well as current strands of research.
Preface. 1. The Problem of Cross-Language Information Retrieval G. Grefenstette. 2. On the Effective Use of Large Parallel Corpora in Cross-Language Text Retrieval M.W. Davis. 3. Statistical Methods for Cross-Language Information Retrieval L. Ballesteros, W.B. Croft. 4. Distributed Cross-Lingual Information Retrieval C. Fluhr, et al. 5. Automatic Cross-Language Information Retrieval Using Latent Semantic Indexing M.L. Littman, et al. 6. Mapping Vocabularies Using Latent Semantics D.A. Evans, et al. 7. Cross-Language Information Retrieval: A System for Comparable Corpus Querying E. Picchi, C. Peters. 8. A Language Conversion Front-End for Cross-Language Information Retrieval K. Yamabana, et al. 9. The SYSTRAN NLP Browser: An Application of Machine Translation Technology in Cross-Language Information Retrieval D.A. Gachot, et al. 10. A Weighted Boolean Model for Cross-Language Text Retrieval D. Hull. 11. Building a Large Multilingual Test Collection from Comparable News Documents P. Sheridan, et al. 12. Evaluating Cross-Language Text Filtering Effectiveness D.W. Oard, B.J. Dorr. References. Index.
In this paper, we present an algorithm for the automatic extraction of transla-tion patterns between two (Indo-)European languages. These consist of possibly discontiguous text fragments, with the bilingual relationship between the text frag-ments and the discontinuities between them made explicit. The patterns are ex-tracted from a bilingual parallel corpus aligned at the sentence level, without the need for linguistic analysis, and are used to build a translation memory database which is intended for use in a machine aided human translation (MAHT) setting, such as a translator's workbench (TWB). The patterns extracted could also form the basis for example-based machine translation (EBMT) without the need for complex linguistic or statistical processing. Given a TM database made up of our concept of translation patterns and a SL input string, relevant translation patterns combine to form TL translations as suggestions to the translator. We evaluate the accuracy of the translation patterns extracted along with the quality of translations produced.
Tangible User Interfaces (TUIs) are emerging as a new paradigm for facilitating user interaction with the digital world by providing intuitive means to link the physical and digital worlds. The MEMODULES project has the objective of developing, experimenting and evaluating the concept of tangible shortcuts (reminders), facilitating (a) the control of devices in the everyday life and also (b) the categorization of information in order to ease or improve information access and retrieval. The project aims at facilitating the user interaction with multimedia information by supporting both the creation and management of tangible links to digital content. Moreover, our research investigates the opportunity of a more complex, multi-sensorial combination of objects and multimedia information by combining multiple interaction modalities - such as voice and gesture - with interactive information visualizations. In order to address these issues we propose a user-oriented framework, called Memodules Framework, enabling end users to turn everyday objects into Memodules. The framework provides a single platform that combines end-user programming, tangible interaction, multimodality and personal information management issues. Memodules framework is built upon MemoML (Memodules Markup Language) and SMUIML (Synchronized Multimodal User Interaction Markup Language) models, which guarantee framework flexibility, extensibility and evolution over time.
We describe an interlingua-based approach to machine translation, in which a DRS represen- tation of the source text is used as the inter- lingua representation. A target DRS is then created and used to construct the target text. We describe several advantages of this level of representation. We also argue that problems of translation mismatch and divergence should properly bo viewed not as translation problems per se but rather as generation problems, al- though the source text can be used to guide the target generator. The system we have built relics exclusively on monolingual linguistic de- scriptions that are also, for the most part, bi- directional.