Spoken Language Understanding Systems 2
HomeFull List of Titles
1: ICSLP'98 Proceedings
Keynote SpeechesText-To-Speech Synthesis 1Spoken Language Models and Dialog 1Prosody and Emotion 1Hidden Markov Model Techniques 1Speaker and Language Recognition 1Multimodal Spoken Language Processing 1Isolated Word RecognitionRobust Speech Processing in Adverse Environments 1Spoken Language Models and Dialog 2Articulatory Modelling 1Talking to Infants, Pets and LoversRobust Speech Processing in Adverse Environments 2Spoken Language Models and Dialog 3Speech Coding 1Articulatory Modelling 2Prosody and Emotion 2Neural Networks, Fuzzy and Evolutionary Methods 1Utterance Verification and Word Spotting 1 / Speaker Adaptation 1Text-To-Speech Synthesis 2Spoken Language Models and Dialog 4Human Speech Perception 1Robust Speech Processing in Adverse Environments 3Speech and Hearing Disorders 1Prosody and Emotion 3Spoken Language Understanding Systems 1Signal Processing and Speech Analysis 1Spoken Language Generation and Translation 1Spoken Language Models and Dialog 5Segmentation, Labelling and Speech Corpora 1Multimodal Spoken Language Processing 2Prosody and Emotion 4Neural Networks, Fuzzy and Evolutionary Methods 2Large Vocabulary Continuous Speech Recognition 1Speaker and Language Recognition 2Signal Processing and Speech Analysis 2Prosody and Emotion 5Robust Speech Processing in Adverse Environments 4Segmentation, Labelling and Speech Corpora 2Speech Technology Applications and Human-Machine Interface 1Large Vocabulary Continuous Speech Recognition 2Text-To-Speech Synthesis 3Language Acquisition 1Acoustic Phonetics 1Speaker Adaptation 2Speech Coding 2Hidden Markov Model Techniques 2Multilingual Perception and Recognition 1Large Vocabulary Continuous Speech Recognition 3Articulatory Modelling 3Language Acquisition 2Speaker and Language Recognition 3Text-To-Speech Synthesis 4Spoken Language Understanding Systems 4Human Speech Perception 2Large Vocabulary Continuous Speech Recognition 4Spoken Language Understanding Systems 2Signal Processing and Speech Analysis 3Human Speech Perception 3Speaker Adaptation 3Spoken Language Understanding Systems 3Multimodal Spoken Language Processing 3Acoustic Phonetics 2Large Vocabulary Continuous Speech Recognition 5Speech Coding 3Language Acquisition 3 / Multilingual Perception and Recognition 2Segmentation, Labelling and Speech Corpora 3Text-To-Speech Synthesis 5Spoken Language Generation and Translation 2Human Speech Perception 4Robust Speech Processing in Adverse Environments 5Text-To-Speech Synthesis 6Speech Technology Applications and Human-Machine Interface 2Prosody and Emotion 6Hidden Markov Model Techniques 3Speech and Hearing Disorders 2 / Speech Processing for the Speech and Hearing Impaired 1Human Speech ProductionSegmentation, Labelling and Speech Corpora 4Speaker and Language Recognition 4Speech Technology Applications and Human-Machine Interface 3Utterance Verification and Word Spotting 2Large Vocabulary Continuous Speech Recognition 6Neural Networks, Fuzzy and Evolutionary Methods 3Speech Processing for the Speech-Impaired and Hearing-Impaired 2Prosody and Emotion 72: SST Student Day
SST Student Day - Poster Session 1SST Student Day - Poster Session 2Author Index
A B C D E F G H I
J K L M N O P Q R
S T U V W X Y Z
Multimedia Files
Automatic Ambiguity Detection
Authors:
Richard Sproat, Bell Laboratories, Lucent Technologies (USA)
Jan P.H. van Santen, Bell Laboratories, Lucent Technologies (USA)
Page (NA) Paper number 41
Abstract:
Most work on sense disambiguation presumes that one knows beforehand --- e.g. from a thesaurus --- a set of polysemous terms. But published lists invariably give only partial coverage. For example, the English word tan has several obvious senses, but one may overlook the abbreviation for tangent. In this paper, we present an algorithm for identifying interesting polysemous terms and measuring their degree of polysemy, given an unlabeled corpus. The algorithm involves: (i) collecting all terms within a k-term window of the target term; (ii) computing the inter-term distances of the contextual terms, and reducing the multi-dimensional distance space to two dimensions using standard methods; (iii) converting the two-dimensional representation into radial coordinates and using isotonic/antitonic regression to compute the degree to which the distribution deviates from a single-peak model. The amount of deviation is the proposed polysemy index.
SL980041.PDF (From Author)
SL980041.PDF (Rasterized)
TOP
Empowering Knowledge Based Speech Understanding through Statistics
Authors:
Julia Fischer, Chair for Pattern Recognition, University of Erlangen-Nuremberg (Germany)
Juergen Haas, Chair for Pattern Recognition, University of Erlangen-Nuremberg (Germany)
Elmar Nöth, Chair for Pattern Recognition, University of Erlangen-Nuremberg (Germany)
Heinrich Niemann, Chair for Pattern Recognition, University of Erlangen-Nuremberg (Germany)
Frank Deinzer, Chair for Pattern Recognition, University of Erlangen-Nuremberg (Germany)
Page (NA) Paper number 369
Abstract:
In this paper we present an innovative approach to speech understanding which is based on a fine-grained knowledge representation automatically compiled from a semantic network and on iterative optimization. Besides allowing an efficient exploitation of parallelism, any-time capability is provided since after each iteration step a (sub-)optimal solution is always available. We apply this approach to a real-world task, which is a dialog system able to answer queries about the German train timetable. In order to speed up the search for the best interpretation of an utterance we make use of statistical methods, e.g. neural networks, n-grams, and classification trees, which are trained on application relevant utterances collected over the public telephone network. At the moment the real-time factor for interpreting the initial user's utterance is 0.7.
SL980369.PDF (From Author)
SL980369.PDF (Rasterized)
TOP
Concept-Driven Speech Understanding Incorporated with a Statistic Language Model
Authors:
Akito Nagai, Information Technology R&D Center, MITSUBISHI Electric Corporation (Japan)
Yasushi Ishikawa, Information Technology R&D Center, MITSUBISHI Electric Corporation (Japan)
Page (NA) Paper number 1023
Abstract:
We have proposed a method of concept-driven semantic interpretation based on general semantic knowledge of conceptual dependency. In our approach, a concept is a unit of semantic interpretation and an utterance is regarded as a sequence of concepts that convey an intention. However, a considerable number of accepted results were not syntactically meaningful. This is because the order in which linguistic features occurred in the sequence of concepts was not taken into account in constructing the whole meaning from the concepts: only semantic constraint was used to attain linguistic robustness. Therefore, we introduce a statistical language model which calculates the plausibility of a sequence of concepts from the points of view of the order in which shallow linguistic features occur. Experimental results of speech understanding for 1000-word-vocabulary spontaneous speech show that the proposed method significantly improves the system performance.
SL981023.PDF (From Author)
SL981023.PDF (Rasterized)
TOP
On The Limitations of Stochastic Conceptual Finite-State Language Models For Speech Understanding
Authors:
José Colás, Grupo de Tecnología del Habla - Departamento de Ingeniería Electrónica - E. T. S. I. Telecomunicación - Universidad Politécnica de Madrid (Spain)
Javier Ferreiros, Grupo de Tecnología del Habla - Departamento de Ingeniería Electrónica - E. T. S. I. Telecomunicación - Universidad Politécnica de Madrid (Spain)
Juan Manuel Montero, Grupo de Tecnología del Habla - Departamento de Ingeniería Electrónica - E. T. S. I. Telecomunicación - Universidad Politécnica de Madrid (Spain)
Julio Pastor, Grupo de Tecnología del Habla - Departamento de Ingeniería Electrónica - E. T. S. I. Telecomunicación - Universidad Politécnica de Madrid (Spain)
Ascensión Gallardo, Grupo de Tecnología del Habla - Departamento de Ingeniería Electrónica - E. T. S. I. Telecomunicación - Universidad Politécnica de Madrid (Spain)
José Manuel Pardo, Grupo de Tecnología del Habla - Departamento de Ingeniería Electrónica - E. T. S. I. Telecomunicación - Universidad Politécnica de Madrid (Spain)
Page (NA) Paper number 1095
Abstract:
In a limited domain task (e.g. airline reservation, database retrieval, etc) many robust understanding systems, designed for both speech and text input, have been implemented [1][3][5][6][10] based on the Stochastic Conceptual Finite-State paradigm (Semantic Network) or CHRONUS paradigm [11] (Conceptual Hidden Representation of Natural Unconstrained Speech), which establishes relations between conceptual entities through a probabilistic graph-like structure. The use of this kind of grammar to model semantic information presents limitations, which have been analysed during the implementation of a flexible architecture for a robust information retrieval system, based on the same paradigm [1]. We have tried to solve some of them by integrating a set of conceptual probabilistic and non-probabilistic grammars, which allow certain complexity in the functionality of the application, such as applying non-SQL functions to the results of SQL queries in order to retrieve information not explicitly included in the database, translating certain natural spoken sentences (that would produce difficult embedded queries and therefore more natural queries) without so many restrictions in the relative position of inter-concept relationship.
SL981095.PDF (From Author)
SL981095.PDF (Rasterized)
TOP
Towards Speech Understanding Across Multiple Languages
Authors:
Todd Ward, IBM (USA)
Salim Roukos, IBM (USA)
Chalapathy Neti, IBM (USA)
Jerome Gros, IBM (formerly of) (USA)
Mark Epstein, IBM (USA)
Satya Dharanipragada, IBM (USA)
Page (NA) Paper number 400
Abstract:
In this paper we describe our initial efforts in building a natural language understanding (NLU) system across multiple languages. The system allows users to switch languages seamlessly in a single session without requiring any switch in the speech recognition system. Context-dependence is maintained across utterances, even when the user changes languages. Towards this end we have begun building a universal speech recognizer for English and French languages. We experiment with a common phonology for both French and English with a novel mechanism to handle language dependent variations. Our best results so far show about 5% relative performance degradation for English relative to a unilingual English system and a 9% relative degradation in French relative to a unilingual French system. The NLU system uses the same statistical understanding algorithms for each language, making system development, maintenance and portability vastly superior to systems built customly for each language.
SL980400.PDF (From Author)
SL980400.PDF (Rasterized)
TOP
Automatic Detection of Sentence Boundaries and Disfluencies Based on Recognized Words
Authors:
Andreas Stolcke, SRI International (USA)
Elizabeth Shriberg, SRI International (USA)
Rebecca Bates, Boston University (USA)
Mari Ostendorf, Boston University (USA)
Dilek Hakkani, SRI International (USA)
Madelaine Plauche, SRI International (USA)
Gökhan Tür, SRI International (USA)
Yu Lu, SRI International (USA)
Page (NA) Paper number 59
Abstract:
We study the problem of detecting linguistic events at interword boundaries, such as sentence boundaries and disfluency locations, in speech transcribed by an automatic recognizer. Recovering such events is crucial to facilitate speech understanding and other natural language processing tasks. Our approach is based on a combination of prosodic cues modeled by decision trees, and word-based event N-gram language models. Several model combination approaches are investigated. The techniques are evaluated on conversational speech from the Switchboard corpus. Model combination is shown to give a significant win over individual knowledge sources.