Text, Speech and Dialogue
14th International Conference, TSD 2011, Pilsen, Czech Republic, September 1-5, 2011, Proceedings
Published on 28. August 2011
XV, 444 pages
978-3-642-23538-2 (ISBN)
Description
This book constitutes the refereed proceedings of the 14th International Conference on Text, Speech and Dialogue, TSD 2011, held in Pilsen, Czech Republic, in September 2011.
The 53 papers presented together with 2 invited talks were carefully reviewed and selected from 110 submissions. The main topic of this year's conference was "integrating modern Web with speech and language technologies". This year the Third International Workshop on Balto-Slavonic Natural Language was affiliated to TSD. The present book contains 8 contributions from this workshop.
More details
Other editions
Additional editions
Ivan Habernal | Vaclav Matousek
Text, Speech and Dialogue
14th International Conference, TSD 2011, Pilsen, Czech Republic, September 1-5, 2011, Proceedings
Book
08/2011
1st Edition
Springer
€53.49
Shipment within 7-9 days
Content
- Title
- Preface
- Organization
- Table of Contents
- Invited Talks
- Dealing with Unexpected Words in Automatic Recognition of Speech
- Introduction
- Current ASR
- Unexpected Words in Human Communication
- Physiological Evidence for the Parallel Combination of the Sensory and the Prior Knowledge (the Context) in Human Recognition of Speech
- Psychophysical Evidence for the Parallel Prior Knowledge (Context) Channel in Human Recognition of Speech
- The Multiplication of Error in Parallel Processing Channels
- The Proposal
- How to Implement Our Scheme?
- Initial Results
- Next Steps
- Summary of More Recent Results
- Some Thoughts for the Future
- References
- A Cloud on the Horizon
- Conference Papers
- A Novel Lecture Browsing System Using Ranked Key Phrases and StreamGraphs
- Introduction
- Related Work and Motivation
- Data
- Key Phrases
- Candidate Extraction
- Ranking
- Evaluation
- Setup
- Evaluation Measure
- Results
- Integrated Browsing System
- Summary
- References
- Addressing Multimodality in Overt Aggression Detection
- Introduction
- Related Work
- Corpus of Multimodal Aggression
- Database Description
- Annotation
- Automatic Aggression Detection
- Audio Processing
- Video Processing
- Multimodal Processing
- Results
- Unimodal Differences and Consequences for Multimodal Classification
- Conclusions and Future Work
- References
- Analysis of Data Collected in Listening Tests for the Purpose of Evaluation of Concatenation Cost Functions
- Introduction
- Perceptual Data Collection
- Sentence Material
- Sentence Selection Considerations
- Selection Methods
- Listening Tests Subjects
- Listening Tests Procedure
- Data Analysis
- Analysis of Listeners' Answers
- Collection of ``facts''
- Distribution of ``facts'' in ?F0 × ?En Plane
- Distribution of ``facts'' in Test Stimuli Sets
- Conclusions and Future Work
- References
- Analysis of Inconsistencies in Cross-Lingual Automatic ToBI Tonal Accent Labeling
- Introduction
- Experimental Setup
- Processing of the Corpora
- The Classifiers
- Results
- Contrasting Automatic and Manual Labeling
- Conclusions and Future Work
- References
- Automatic Semantic Labeling of Medical Texts with Feature Structures
- Introduction
- Semantic Information Identified
- Data Preparation
- CRF Model of Label Assignment
- Conclusions
- References
- Automatic Switchboard Operator
- Introduction
- Dialogue System and the Dialogue
- Scheme of the Dialogue
- Key Algorithms
- Text Preprocessing
- Rule-Based Generation of Utterances
- The Speech Grammar and Its Complexity
- Experiences and Grammar Adjustments
- Statistics Collected during the Operation
- Out-of-Grammar Utterances
- Summary
- References
- Automatic Topic Identification for Large Scale Language Modeling Data Filtering
- Introduction
- Topic Identification
- Topic (Keyword) Tree
- Identification Algorithms
- Evaluation
- Language Modeling and ASR Experiments
- Conclusions and Future Work
- References
- Automatic Translation Error Analysis
- Introduction
- Method Description
- Word Alignment
- Detecting Lexical Errors
- Detecting Order Errors
- Error Summarization
- Experiments and Results
- Used Data
- Evaluation Results
- Related Work
- Future Work
- Conclusions
- References
- Automatic Word Sense Disambiguation and Construction Identification Based on Corpus Multilevel Annotation*
- Introduction
- Linguistic Data
- Toolkit for WSD and CxI
- Conditions for WSD and CxI
- Identification of Context Markers for Russian Nouns
- Construction Identification
- Conclusion
- References
- Bootstrapping Bilingual Lexicons from Comparable Corpora for Closely Related Languages
- Introduction
- Related Work
- Resources Used
- Building a Comparable Corpus
- Building a Seed Dictionary
- Building a Gold Standard
- Experimental Setup
- Adding Cognates to the Seed Dictionary
- Adding First Translation Candidates to the Seed Dictionary
- Combining Cognates and First Translation Candidates to Extend the Seed Dictionary
- Conclusions and Future Work
- References
- Combining Topic Specific Language Models
- Introduction
- Related Work
- Comparing Specific and General Models
- Bayesian Network Models
- Dynamic Bayesian Networks
- A Topic-Based Model
- A Cluster-Based Model
- The Dynamic Bayesian Network Tree Model
- Learning DBNT Language Models
- Selection of Component Language Models in Prediction
- Conclusion
- References
- Czech HMM-Based Speech Synthesis: Experiments with Model Adaptation
- Introduction
- System Overview
- Training Stage
- Adaptation
- Synthesis Stage
- Contextual Factors
- Combined Context-Related Clustering Questions
- Experiments and Results
- Experimental Data Description
- Quality Evaluation
- Voice Identity Evaluation
- Conclusion and Future Work
- References
- Effective Parsing Using Competing CFG Rules
- Introduction
- Syntactic Parser synt
- Grammar
- Chart and Forest of Values
- Pruning by Rule Levels
- Chart-Based Local Pruning
- Forest-Based Non-local Pruning
- Evaluation
- Conclusions
- References
- Efficiency of Speech Alignment for Semi-automated Subtitling in Dutch
- Introduction
- Segmentation of the Audio Stream
- Preprocessing of the Transcripts
- Speech Alignment
- Overview of the Complete System
- System Evaluation
- Qualitative Evaluation
- Quantitative Evaluation
- Conclusions
- References
- Evaluation of Hands-Free Large Vocabulary Continuous Speech Recognition by Blind Dereverberation Based on Spectral Subtraction by Multi-channel LMS Algorithm
- Introduction
- Dereverberation Based on Power Spectral Subtraction
- Compensation Parameter Estimation for Spectral Subtraction by Multi-channel LMS Algorithm
- Experiments
- Experimental Setup
- Experimental Results on LVCSR
- Effect Factor Analysis of Compensation Parameter Estimation
- Conclusions and Future Work
- References
- Fusion of Discriminative and Generative Scoring Criteria in GMM-Based Speaker Verification
- Introduction
- Joint Factor Analysis
- Support Vector Machine
- Experimental Setup
- Test Set
- Feature Extraction
- Factor Analysis Training
- Estimation of the Speaker Model
- Scoring
- Score Normalization
- Score Fusion
- Results
- Conclusions
- References
- Generalized Non-uniform Time Scaling Distribution Method for Natural-Sounding Speech Rate Change
- Introduction
- Time-Scale Modifications of Speech
- Generalized Non-linear Scaling Scheme
- Scheme Illustration
- Setting the Rates
- Conclusion
- References
- Grouping Alternating Schemata in Semantic Valence Dictionary of Polish Verbs
- Introduction
- Related Works
- Valence Dictionary
- Classification of Alternations
- Rules of Grouping Schemata
- Experiments
- Validation
- Conclusions
- References
- Hierarchical Dialogue System for Guide Robot in Shopping Mall Environments
- Introduction
- Hierarchical Dialogue Model Based on Robot Function Modes
- Conclusion
- References
- Identifying Concatenation Discontinuities by Hierarchical Divisive Clustering of Pitch Contours
- Introduction
- Perceptual Data Collection
- Test Material
- Listening Tests Subjects
- Listening Tests Procedure
- Listening Test Evaluation and Results
- Clustering Experiment
- Motivation
- Set of Observations
- Results
- Conclusions and Future Work
- References
- Identifying Verbal Collocations in Wikipedia Articles
- Introduction
- The Characteristics of Verb-Particle Constructions and Light Verb Constructions
- Related Work
- Experiments
- Background
- Methods for Detecting Verbal Collocations
- Results
- Discussion
- Conclusions
- References
- Initialization of fMLLR with Sufficient Statistics from Similar Speakers
- Introduction
- Adaptation
- Feature Maximum Likelihood Linear Regression (fMLLR)
- Sufficient Statistics of Closest Speakers
- Accumulation of Sufficient Statistics
- Selecting a Cohort of Speakers
- Estimating a New Model
- fMLLR Initialization through Sufficient Statistics
- Accumulation of Sufficient Statistics
- Selecting a Cohort of Speakers
- fMLLR Transform Estimation
- Experiments
- SpeechDat-East (SD-E) Corpus
- Adaptation Setup
- Results
- Conclusion
- References
- Intelligibility Rating with Automatic Speech Recognition, Prosodic, and Cepstral Evaluation
- Introduction
- Test Data and Subjective Evaluation
- Cepstral Analysis
- The Speech Recognition System
- Prosodic Features
- Support Vector Regression (SVR)
- Results and Discussion
- References
- Maximum Entropy Named Entity Recognition for Czech Language
- Introduction
- State of the Art
- Classifier
- Features
- Semantic Spaces
- Experiments
- Results
- Conclusion and Future Work
- References
- Mining Significant Words from Customer Opinions Written in Different Natural Languages
- Introduction
- Data Description
- Text Document Pre-processing and Representation
- Creating Dictionaries of Significant Words
- Experiments and Their Results
- Conclusions
- References
- On Positive and Unlabeled Learning for Text Classification
- Introduction
- Related Work
- The Proposed Technique
- The Rocchio Technique
- The Proposed Technique
- Experiments and Results
- Conclusion
- References
- Optimisation Approach to the Construction of the Polish Morphological Guesser
- Introduction
- General Approach
- Constructing a Tree from Reversed Suffixes
- Introducing a Notion of a Guessing Function
- Choosing the Right Function
- Computing the F-Measure
- Avoiding Exponential Growth
- Further Optimisations
- Testing and Results
- Evaluation
- References
- Prefix Recognition Experiments
- Introduction
- Methods
- Naive Method
- Squares
- Entropy Methods
- Economy Method
- Data
- Experiments
- Comparison of All Methods
- Conclusions and Plans
- References
- Question Classification by Weighted Combination of Lexical, Syntactic and Semantic Features
- Introduction
- Related Work
- Choosing the Classifier
- Features in Question Classification
- Lexical Features
- Syntactic Features
- Semantic Features
- Combining Features
- Conclusion
- References
- Recursive Decompounding in Afrikaans
- Introduction
- Development of the Algorithm
- Basic Principles
- Problem Areas
- The Decompounding Algorithm
- Performance Measures and Results
- Conclusions
- References
- Reliable Detection of ImportantWord Boundaries Using Prosodic Features
- Introduction
- Speech Database
- Prosodically Marked Boundaries
- Erlangen Prosody Module
- Slot Boundary Model
- Detecting Prosodically Marked Boundaries
- Detailed Analysis
- Recordings Containing No Classified Boundaries
- Recordings Containing at Least One Classified Boundaries
- PMBs between Two Content Words
- PMBs between a Content Word and a Non-content Word
- Summary and Outlook
- References
- Rule-Based Triphone Mapping for Acoustic Modeling in Automatic Speech Recognition
- Introduction
- Rule-Based Triphone Mapping
- Rules for Phonetic Similarity in Slovak
- Experiments
- Data
- Results
- Discussion
- References
- Semantic Relatedness for Named Entity Disambiguation Using a Small Wikipedia
- Introduction
- Related Work
- Experimental Settings
- Resources
- Methods
- Evaluation and Results
- Conclusions
- References
- Speaker-Clustered Acoustic Models Evaluated on GPU for On-line Subtitling of Parliament Meetings
- Introduction
- Methods
- Unsupervised Clustering
- Acoustic Models Fusion
- GPU Accelerated Acoustic Model Evaluation
- Train Data Description
- Annotated Data
- Unsupervised Data
- Experimental Setup
- Acoustic Processing
- Acoustic Modeling
- Unsupervised Speaker Clustering
- Tests Description
- Results
- Conclusion
- References
- Speaker Recognition from Coded Speech Using Support Vector Machines
- Introduction
- Impact of Speech Coding on Speaker Recognition
- SVMs and Speaker Recognition
- Aims of This Study
- Experiment Setup
- Speech Data
- Tested Codecs
- Classification
- Results
- Conclusions and Future Works
- References
- Statistical Analysis of Complementary Spectral Features of Emotional Speech in Czech and Slovak
- Introduction
- Rule-Based Abbreviation Expansion
- Statistical-Based Abbreviation Expansion
- Experiments and Results
- Conclusion and Future Work
- References
- The Role of Neural Network Size in TRAP/HATS Feature Extraction
- Introduction
- Probabilistic Features
- System Description
- TRAP/HATS Neural Network Architectures
- Experimental Results and Discussion
- Detailed Analysis
- Conclusions
- References
- Time Dimension in the Dolphin Nick Knowledge Base Using Transparent Intensional Logic
- Introduction
- Simple Temporal Aspect in TIL and Dolphin Nick
- Actions and Verbs
- Time Inference and Grammatical Tenses
- Example Questions and Answers
- Conclusions and Future Work
- References
- Towards Automatic Annotation of Sign Language Dictionary Corpora
- Introduction
- Related Work
- Data
- Hand Tracking
- Skin Color Segmentation
- Tracking
- Head Tracking
- Experiment
- Conclusion
- References
- Unsupervised Topic-Oriented Keyphrase Extraction and Its Application to Croatian
- Introduction
- Related Work
- Topic-Oriented Keyphrase Extraction
- Keyword Extraction
- Keyword-to-Keyphrase Expansion
- Evaluation
- Evaluation Methodology
- Results and Discussion
- Evaluating Keyword-to-Keyphrase Expansion
- Conclusion
- References
- Voice Assessment of Speakers with Laryngeal Cancer by Glottal Excitation Modeling Based on a 2-Mass Model
- Introduction
- Dataset
- Glottal Excitation
- Two Mass Model
- Model Optimization
- Results and Discussion
- Summary
- References
- Web Text Data Mining for Building Large Scale Language Modelling Corpus
- Introduction
- System Architecture
- Data Sources
- Text Preprocessing
- Text Cleaning
- Tokenization and Text Normalization
- Vocabulary-Based Token Substitution and Decapitalization
- Duplicity Detection
- Topic Identification
- Corpus Statistics
- Conclusion and Future Work
- References
- Web-Based System for Automatic Reading of Technical Documents for Vision Impaired Students
- Introduction
- System Overview
- Backend
- Frontend
- Text-to-Speech
- Automatic Reading of Formulas
- Reading of ``Inline'' Formulas Represented by a Plain Text
- Reading of Formulas Represented by MathML
- Conclusion and Future Work
- References
- Zanzibar OpenIVR: An Open-Source Framework for Development of Spoken Dialog Systems
- Introduction
- The Architecture of Zanzibar OpenIVR
- Components
- Integration Issues
- Dialog Management and Natural Language Understanding
- Related Work
- Evaluation and Discussion
- Conclusion and Further Work
- References
- Balto-Slavonic Natural Language Processing 2011 Workshop
- Building Support Tools for Russian-Language Information Extraction
- Introduction
- Background and Context of the Work
- Baseline English System
- AOT
- System Integration
- AOT Wrapper
- Information Extraction
- Current Work
- References
- Finding the Optimal Number of Clusters for Word Sense Disambiguation
- Introduction
- Selected Stopping Rules
- Datasets
- Experiments
- Conclusions and Further Works
- References
- hrWaC and slWac: Compiling Web Corpora for Croatian and Slovene
- Introduction
- Building the hrWaC and slWaC
- Collecting Seeds, Crawling, Physical Deduplication
- Content Extraction
- Language Identification, Filtering and PoS Tagging
- Corpus Comparison
- Conclusion
- References
- Question Classification for a Croatian QA System
- Introduction
- Related Work on Question Classification
- Question Classification for the Croatian Language
- Features
- QC Test Collection
- Evaluation
- Conclusion
- References
- Random Indexing Distributional Semantic Models for Croatian Language
- Introduction
- Distributional Semantic Models
- Building DSMs for Croatian Language
- Corpus
- Models
- Evaluation and Results
- Semantic Similarity Judgments
- Results
- Remarks
- Conclusion
- References
- Structure Annotation in the Polish Corpus of Suicide Notes
- Introduction
- Choosing Text-Encoding Standard
- Annotation Scheme
- Case Study
- Corpus Statistics and Availability
- Conclusions and Further Research
- References
- Unsupervised Russian POS Tagging with Appropriate Context
- Introduction
- CHMM for Russian POS Tagging
- Parameter Estimation and a Potential Issue
- Incorporating All Three Types of Context
- Experiments and Results
- Conclusion and Future Work
- References
- WCCL: A Morpho-syntactic Feature Toolkit
- Background
- JOSKIPI
- Implementations and Applications
- Limitations and Design Flaws
- Proposed Successor to JOSKIPI
- Toolkit Features
- WCCL Operators
- Case Study: Creating a Memory-Based Chunker
- Conclusion
- References
- Author Index
