This paper introduces a paradigm of context-dependent grammar (CDG) and an acquisition system that, through interactive teaching sessions, accumulates the CDG rules. The resulting context-sensitive rules are used by a stack-based, shift~reduce parser to compute unambiguous syntactic structures of sentences. The acquisition system and parser have been applied to the phrase structure and case analyses of 345 sentences, mainly from newswire stories, with 99 % accuracy. Extrapolation from our current grammar predicts that about 25 thousand CDG rule examples will be sufficient to train the system in phrase structure analysis of most news stories. Overall, this research concludes that CDG is a computationally and conceptually tractable approach for the construction of sentence grammar for large subsets of natural language text. 1.

this paper

The present paper describes the linguistic components of a platform for building spoken language dialogue systems. The platform is being designed and implemented within the EU-language engineering project REWARD. The linguistic components mainly consist of a suite of general unification based natural language processing utilities and a window based Sub Grammar Design Tool which provides non-experts access to the utilities. The Sub Grammar Design Tool enables nonexpert users to implement sub grammars for speech understanding within a minimum of effort. I. INTRODUCTION The three years EU-Language Engineering project REWARD ("Real World Applications of Robust Dialogue" LE1-2632) addresses the needs of organisations which do business over the telephone (i) to automate certain telephone services using spoken language dialogue technology and (ii) to automate the process of creating such services. The project brings together two technology suppliers, Vocalis Ltd. and CPK (Center for PersonK...

Due to recent development in the area of computational formalisms for linguistic representation, the task of designing a parser for a specified natural language is now shifted to the problem of designing its grammar in certain formal ways. This paper describes the results of a project whose aim was to design a formal grammar for modern Hebrew. Such a formal grammar has never been developed before. Since most of the work on grammatical formalisms was done without regarding Hebrew (and other Semitic languages as well), we had to choose a formalism that would best fit the specific needs of the language. This part of the project has been described elsewhere. In this paper we describe the details of the grammar we developed. The grammar deals with simple, subordinate and coordinate sentences as well as interrogative sentences. Some structures were thoroughly dealt with, among which are noun phrases, verb phrases, adjectival phrases, relative clauses, object and adjunct clauses; many types o...

This paper describes a suite of NLP-modules designed and implemented within two very different spoken language projects: (i) REWARD (Real World Applications of Robust Dialogue, LE1-2632) is an EU-funded project running for a 3 years period from Nov. 1995 - Oct. 1998. The project addresses the needs of organisations doing business over the telephone to automate telephone services using spoken language dialogue technology and to automate the process of creating such services. REWARD brings together two technical suppliers, Center for PersonKommunikation (CPK) at Aalborg University and Vocalis Ltd. in Cambridge U.K., along with four user organisations working in the area of telemarketing, marketing research, travel agency, and hardware maintenance. The user organisations are Taylor Nelson AGB (U.K.), NIPO (The Netherlands), DanTransport (Denmark), and MADE (Spain). (ii) IntelliMedia 2000 + was initiated in 1996 by the Institute of Electronic Systems, Aalborg University, and involves real-time computer processing and understanding of perceptual input from speech, textual and visual sources and implementation of an educational multimedia workbench (Chameleon) with demonstrators. The research team includes researchers from CPK, Laboratory for Image Analysis (LIA), Laboratory for Medical Informatics (MI), and Computer Science (CS).

A UNIFICATION-BASED TOOL FOR LEARNING OF TURKISH MORPHOLOGY Pembeci, Izzet M.S., Department of Computer Engineering Supervisor: Assist. Prof. Dr. Cem Boz¸sahin Co-Supervisor: Assoc. Prof. Dr. Deniz Zeyrek January 1998, ?? pages This thesis is about developing a teaching tool for the morphology of Turkish. The tool will provide a practising environment for non-native learners of Turkish and students of linguistics. Two-level morphology is chosen for modelling the morphology of Turkish. In this model, morphophonemic processes are described by two-level phonology rules and the ordering constraints for the morphemes are encoded by a word grammar. Keywords: Morphology, Morphotactics, Morphophonemics, Computer Aided Language Learning iii OZ T URKC¸E'N IN B IC¸ IMB IL IM IN IN B ILG ISAYAR YARDIMIYLA O GRET ILMES I Pembeci, Izzet Yuksek Lisans, Bilgisayar Muhendisligi Bolumu Tez Yoneticisi: Yrd. Do¸c. Dr. Cem Boz¸sahin Ortak Tez Yoneticisi: Do¸c. Dr. Deniz Zeyrek Ocak 1998, ?...

ing from the nature of the states involved, we can specify the change that an atomic program a effects by means of a two place transition relation R a on a set of states. This perspective gives rise to the study of so-called transition systems. The most general style of reasoning about programs and transition system is found in propositional dynamic logics (Pratt [ Pratt, 1976 ] [ Pratt, 1980 ] , Harel [ Harel, 1984 ] ) and in algebras of processes (Hennessy [ Hennessy, 1988 ] ). Processes and transition systems are studied from the perspective of modal logic in Stirling [ Stirling, 1987 ] and Van Benthem and Bergstra [ Benthem and Bergstra, 1993 ] . Dynamic semantics can be put to use to stipulate relational denotations for propositions. In this perspective, a state of information is a set of possible worlds, and a program updates a state of information by removing the worlds incompatible with the new information. Thus, the semantics of language is defined in terms of its potential to...

A WORD GRAMMAR OF TURKISH WITH MORPHOPHONEMIC RULES Oztaner, Serdar Murat M.S., Department of Computer Engineering Supervisor: Assist. Prof. Dr. Cem Boz¸sahin January 1996, 128 pages This thesis is about the computational morphological analysis and generation of Turkish word forms. Turkish morphological description is encoded using the two-level morphological model. This description consists of a phonological component that contains the two-level morphophonemic rules, and a lexicon component which lists lexical items (indivisible words and affixes) and encodes the morphotactic constraints. In the scope of the study, a generic word grammar in a tabular form expressing the ordering relationships among morphemes is designed and morphophonemic processes along with solutions to exceptional cases are formulated. Keywords: Morphology, Morphotactics, Morphophonemics, Turkish grammar iii OZ T URKC¸E'N IN S OZC UK YAPISI VE B IC¸ IMB IR IM SESB IR IM ETK ILES¸ IM I IC¸ IN KURA...

Attribute grammars are an elegant formalization of the augmented context-free grammars characteristic of most current natural language systems. This paper presents an extension of Earley's algorithln to Knuth's attribute grammars, considering the case of S-attributed grammars. For this case, we study the conditions on the underlying base grammar under which the extended algorithm may be guaranteed to terminate. Finite partitioning of attribute domains is proposed to guarantee the termination of the algorithm, without the need for any restrictions on the context-free base.

The "Voice Browser " activity within the W3C consortium addresses the need for standards for speech grammars, dialogue descriptions etc. in distributed systems. This paper discusses the consortium's recent speech grammar working draft specification. The W3C specification is based on the Java Speech Grammar Format (JSGF) defined by Sun Microsystems and is - with all good and bad qualities - characterized by traditions of formal language theory. In a constructive spirit, we suggest some possible improvements based on natural language theory. The suggestions concern compound feature-based semantic presentations, lexicon structure, ambiguity, and exhaustive parsing, 1.