Latent semantic mapping principles & applications /
Latent semantic mapping (LSM) is a generalization of latent semantic analysis (LSA), a paradigm originally developed to capture hidden word patterns in a text document corpus. In information retrieval, LSA enables retrieval on the basis of conceptual content, instead of merely matching words between...
| Main Author: | |
|---|---|
| Format: | Electronic |
| Language: | English |
| Published: |
San Rafael, Calif. (1537 Fourth Street, San Rafael, CA 94901 USA) :
Morgan & Claypool Publishers,
c2007.
|
| Edition: | 1st ed. |
| Series: | Synthesis lectures on speech and audio processing (Online),
#3. |
| Subjects: | |
| Online Access: | Abstract with links to resource |
Table of Contents:
- Principles
- Introduction
- Motivation
- From LSA to LSM
- Organization
- Latent semantic mapping
- Co-occurrence matrix
- Vector representation
- Interpretation
- LSM feature space
- Closeness measures
- LSM framework extension
- Salient characteristics
- Computational effort
- Off-line cost
- Online cost
- Possible shortcuts
- Probabilistic extensions
- Dual probability model
- Probabilistic latent semantic analysis
- Inherent limitations
- Applications
- Junk e-mail filtering
- Conventional approaches
- LSM-based filtering
- Performance
- Semantic classification
- Underlying issues
- Semantic inference
- Caveats
- Language modeling
- N-gram limitations
- MultiSpan language modeling
- Smoothing
- Pronunciation modeling
- Grapheme-to-phoneme conversion
- Pronunciation by latent analogy
- Speaker verification
- The task
- LSM-based speaker verification
- TTS unit selection
- Concatenative synthesis
- LSM-based unit selection
- LSM-based boundary training
- Perspectives
- Discussion
- Inherent tradeoffs
- General applicability
- Conclusion
- Summary
- Perspectives.