Mostrando recursos 38.041 - 38.060 de 79.809

  1. Conversational Telephone Speech Recognition

    J. L. Gauvain; L. Lamel; H. Schwenk; G. Adda; L. Chen; F. Lefèvre
    This paper describes the development of a speech recognition system for the processing of telephone conversations, starting with a state-of-the-art broadcast news transcription system. We identify major changes and improvements in acoustic and language modeling, as well as decoding, which are required to achieve state-of-theart performance on conversational speech. Some major changes on the acoustic side include the use of speaker normalization (VTLN), the need to cope with channel variability, and the need for efficient speaker adaptation and better pronunciation modeling. On the linguistic side the primary challenge is to cope with the limited amount of language model training data....

  2. Information structure in Turkish: the word order–prosody interface

    Selcuk Issever
    This study investigates the linguistic realization of information structure (IS) in Turkish. Following Vallduvi and Engdahl [Linguistics 34 (1996) 459], Hoffman (Hoffman, B., 1995. The Computational Analysis of the Syntax and Interpretation of ‘‘Free’’ Word Order in Turkish. Unpublished PhD dissertation, University of Pennsylvania), and Kilicaslan (Kilicaslan, Y., 1994. Information Packaging in Turkish. Unpublished MSc dissertation, University of Edinburgh), it is assumed that IS has a tripartite structure, consisting of topic, tail, and focus. The main claim of this paper is that syntax and phonology, by means of word order and prosody, are both responsible for the realization of the...

  3. Published in Advances in Artificial Life, Floreano, D., Nicoud, J-D. and Mondada, F.

    Th European Conference; Daniel Livingstone; Colin Fyfe
    Some recent Artificial Life models have attempted to explain the origin of linguistic diversity with varying conclusions and explanations. We posit, contrary to some existing Artificial Life work, that linguistic diversity should naturally emerge in spatially organised populations of language learners, and this is supported by our experimental work and by recent literature.

  4. Constraints on the Generation of Referring Expressions, with Special Reference to Hindi

    Rashmi Prasad; Ellen F. Prince; K. Joshi
    WITH SPECIAL REFERENCE TO HINDI Rashmi Prasad Supervisor: Ellen F. Prince This dissertation makes a progress towards the generation of referring expressions in Hindi. We first make a proposal to exploit a combination of Gricean implicatures (Grice, 1975) and Centering theory constraints (Grosz et al., 1995) to formulate a generation algorithm for referring expressions whose domain of application is defined in terms of the Centering Transitions. The formulated algorithm is an abstraction over the cross-linguistic variability observed across languages. To set the language-specific parameters of the algorithm, in particular the parameter that decides the relative salience of the discourse entities...

  5. The Distributional Structure of Grammatical . . .

    Toben H. Mintz; Elissa L. Newport; Thomas G. Bever
    We present a series of three analyses of young children's linguistic input to determine the distributional information it could plausibly offer to the process of grammatical category learning. Each analysis was conducted on four separate corpora from the CHILDES database (MacWhinney, 2000) of speech directed to children under 2;5. We show that, in accord with other findings, a distributional analysis which categorizes words based on their co-occurrence patterns with surrounding words successfully categorizes the majority of nouns and verbs. In Analyses 2 and 3, we attempt to make our analyses more closely relevant to natural language acquisition by adopting more...

  6. HPSG-Based Annotation Scheme for Corpora Development and Parsing Evaluation

    Kiril Iv. Simov
    This paper proposes a formal framework for development and exploitation of a corpus, based on the HPSG linguistic theory. The formal representation of the annotation scheme facilitates the annotation process and ensures the quality of the corpus and its usage in different application scenarios. Also, evaluation over HPSG annotation scheme is discussed. The advantages of the approach are presented in comparison with other related works.

  7. Interdisciplinary Research to Advance Digital Imagery Indexing and . . .

    James Z. Wang; Jia Li; Ching-chih Chen
    This paper provides an introduction of our NSF-funded research project on advancing digital imagery technologies for Asian art and cultural heritages. This international collaborative research project aims at developing technologies related to the preservation, retrieval, and dissemination of digital imagery. Researchers in the US, China, and South Korea will collectively investigate and develop technologies for acquiring, browsing, managing, and searching large collections of high quality art images. One of the main research questions the team of US researchers focuses on is the problem of automatic indexing and retrieval of digital art images. Building on the foundation of a successful image...

  8. Automatic Distinction of Arguments and Modifiers:

    The Case Of; Paola Merlo; Matthias Leybold
    The automatic distinction of arguments and modifiers is a necessary step for the automatic acquisition of subcategorisation frames and argument structure. In this work, we report on supervised learning experiments to learn this distinction for the difficult case of prepositional phrases attached to the verb. We develop statistical indicators of linguistic diagnostics for argumenthood, and we approximate them with counts extracted from an annotated corpus. We reach an accuracy of 86.5%,over a baseline of 74% , showing that this novel method is promising in solving this difficult problem.

  9. Intelligent Methods In Dosing Control Of Water

    Treatment Esko Juuso; Esko Juuso; Katja Viirret; Marjatta Piironen
    Water treatment includes many complex phenomena, such as coagulation and flocculation. These reactions are hard or even impossible to control satisfyingly by conventional methods. That is the main reason why the applicability of the intelligent methods to the control of wastewater purification chemicals is discussed in this paper. There exist research articles concerning this subject. According to these articles the field test results have been promising. Linguistic equations method, which is chosen as the method in our case, is a new intelligent method in this line of process industry. For this research, data were recorded from a real water purification...

  10. A Direct Linguistic Induction Method For Systems

    Juan Moreno Garca; E. U. Ingeniería; Técnica Industrial
    The aim of this paper is presented a method for obtaining a linguistic model that reflects the behavior of a combined well-known data. The method is based on the technique of successive division of the input space (as CART[1] and ID3[4]). The obtained rules have linguistic variables as antecedent and consequent. So, incorporating the concept of linguistic intervals (as disjunctions of linguistic labels) the methods ID3 and CART are generalized for working directly with linguistic variables defined a priori.

  11. An RSRL Formalization of Serbo-Croatian Second Position CLitic Placement

    Gerald Penn
    This paper has three main purposes. The rst is to generalize the work by Reape 1994, Kathol 1995 and others on domain-based treatments of free verb-phrase-level and clause-level word order to a framework that can express the embedding of word-order domains inside larger domains, and that can deal with multiple kinds of word-order domains in the same grammar. The second is to provide a dierent formalization of wordorder domains in the logic of typed feature structures that allows them to account not only for cases of word order that are under-determined by syntactic principles, but also for cases where principles...

  12. Machine learning for modeling Dutch pronunciation variation

    Veronique Hoste; Steven Gillis; Walter Daelemans
    This paper describes the use of rule induction techniques for the automatic extraction of phonemic knowledge and rules from pairs of pronunciation lexicons. This extracted knowledge allows the adaptation of speech processing systems to regional variants of a language. As a case study, we apply the approach to Northern Dutch and Flemish (the variant of Dutch spoken in Flanders, a part of Belgium) , based on Celex and Fonilex, pronunciation lexicons for Northern Dutch and Flemish, respectively. In our study, we compare two rule induction techniques, TransformationBased Error-Driven Learning (TBEDL) (Brill, 1995) and C5.0 (Quinlan, 1993), and evaluate the extracted...

  13. Genesis-II: A Language Generation Module for Conversational Systems

    Stephanie Sene; Lauren M. Baptist; Lauren M. Baptist
    Language generation is a fundamental component of dialogue systems. Over the past year, we have developed a new generation module for conversational systems developed at MIT using the Galaxy architecture. In this thesis, we discuss how our system, Genesis-II, bridges non-linguistic and linguistic generation with an original framework that is simple, yet powerful. We have created a set of mechanisms that allow domain experts to rapidly develop template-like rules for simple domains, while giving them the power to carefully construct linguistic rules for complex domains. In particular, Genesis-II provides domain experts with exible tools for overcoming dicult linguistic challenges, like...

  14. Author Retains Full Rights AD Talking Out Both Sides of Your Mouth: Streamlining Communication via Metaphor

    Giac (gcih Gold Certification
    As Security is a relatively new field, we are still learning how to communicate what we know with those outside of it. When communicating with non-experts, we often fall back on simplification and analogy to make our points understood. These techniques are fundamentally based on metaphor. This paper explains how metaphors are classically used, drawing on works in the field of linguistic philosophy, communication theory and neuro-linguistic programming. It then explores classic metaphors used within the Security commun...

  15. Analysis of intonation in news presentation on television

    Emma Rodero
    The effectiveness on the television communication is sustained on the handling of voice, which is the base of fundamental expression of the audio-visual contents. Consequently, the audience attention to the television message as well as understanding and assimilating of it, will depend on a correct and expressive use of voice. These reasons guarantee the necessity to make an empirical investigation about the prosody form in which the news broadcasters on television use voice. Therefore, the actual communication is based on the results of a study in order to determine the intonation in broadcasters on television. Overview There are not many...

  16. stability

    Paul E. Hemeren; Sofia Kasviki; Barbara Gawronska
    Lexicalization of natural actions and cross-linguistic

  17. To appear in N. Schilling-Estes (ed.), GURT’06 Unendangered Dialects, Endangered People

    William Labov
    The topic that I will deal with here is a difficult one, especially in a forum devoted to the struggle to save endangered languages and support endangered dialects. 1 The other papers in this volume are concerned with the problem of how to preserve linguistic and cultural diversity throughout the world. Nothing that I will present here should be taken to diminish or undercut the importance of that program. But this report will deal with another side of diversity. I will be looking at social factors that lead dialects to diverge, develop and flourish, and with forms of cultural diversity...

  18. ISCA Archive Towards super-human speech recogniton

    Mukund Padmanabhan; Michael Picheny
    Research in speech recognition has been underway for decades, and a great deal of progress has been made in reducing the word error rate. However, recent studies still demonstrate that machine performance is still quite far from human performance across a wide variety of tasks, ranging from high-bandwidth digit recognition to large vocabulary telephony speech. In addition, for most speech recognition tasks, obtaining good performance relies on tuning to a particular domain or environment. For instance, a system trained on the Switchboard corpus is unlikely to provide close to optimal performance on a small vocabulary task such as telephone digits....

  19. ISCA Archive A BRIEF OVERVIEW OF DECODING TECHNIQUES FOR LARGE VOCABULARY CONTINUOUS SPEECH RECOGNITION

    Xavier L. Aubert
    A number of decoding strategies for large vocabulary speech recognition are examined from the viewpoint of their search space representation. Different design solutions are compared with respect to the integration of linguistic and acoustic constraints, as implied by M-gram LMs and cross-word phonetic contexts. This study is articulated along two main axes, namely, the network expansion and the search algorithm itself. Three broad classes of decoding methods are reviewed: the use of weighted finite state transducers for static network expansion, the time-synchronous dynamic-expansion search and the asynchronous stack decoding. 1.

  20. Backchannels revisited from a multimodal perspective

    Roxane Bertr; Gaëlle Ferré; Robert Espesser; Stéphane Rauzy
    In this study, we analyze the role of several linguistic cues (prosodic units, pitch contours, discourse markers, morphological categories, and gaze direction) in French turntaking face-to-face interactions. Specifically, we investigate vocal and gestural backchannel signals (BCs) produced by a recipient to show his active listening. We show that some particular pitch contours and discursive markers play a systematic role in inducing both gestural and vocal BCs. Conversely, morphological categories and gestural cues rather play a role for gestural BCs. Index Terms: dialogues, vocal and gestural backchannel signals, French, multimodal levels

Aviso de cookies: Usamos cookies propias y de terceros para mejorar nuestros servicios, para análisis estadístico y para mostrarle publicidad. Si continua navegando consideramos que acepta su uso en los términos establecidos en la Política de cookies.