Next:
11 Mathematical Methods
Up:
11 Mathematical Methods
Previous:
10 Transmission and Stroage
Chapter 11Contents
11 Mathematical Methods
11.1 Overview
11.1.1 High-level Linguistic Methods
11.1.2 Statistical and Low-level Processing Methods
11.1.3 Future Directions
11.2 Statistical Modeling and Classification
11.2.1 Primitive Acoustic Features
11.2.2 Quantization
11.2.3 Maximum Likelihood and Related Rules
11.2.4 Class Conditional Density Functions
11.2.5 Hidden Markov Model Methodology
11.2.6 Syntax
11.2.7 Semantics
11.2.8 Performance
11.2.9 Future Directions
11.3 DSP Techniques
11.3.1 Feature Extraction
11.3.2 Dealing with Channel Effects
11.3.3 Vector Quantization
11.3.4 Future Directions
11.4 Parsing Techniques
11.4.1 Parsing Complexity
11.4.2 Derivation Trees
11.4.3 Unification-based Grammars
11.4.4 Parsing as Deduction
11.4.5 LR Parsing
11.4.6 Parsing by Finite State Transducers
11.4.7 Remarks
11.4.8 Future Directions
11.5 Connectionist Techniques
11.5.1 ANNs and Feature Extraction
11.5.2 ANNs and Pattern Sequence Matching
11.5.3 Hybrid HMM/ANN Approach
11.5.4 Language Modeling and Natural Language Processing
11.5.5 Future Directions
11.6 Finite State Technology
11.6.1 Future Directions
11.7 Optimization and Search in Speech and Language Processing
11.7.1 Dynamic Programming-based Search for Speech Recognition
11.7.2 Training/Learning as Optimization
11.7.3 Future Directions
11.8 References
Maintained by
Mike Noel
and
Wei Wei
