Incorporating Knowledge Sources into Statistical Speech Recognition

Front Cover
Springer Science & Business Media, Feb 27, 2009 - Technology & Engineering - 196 pages
0 Reviews

Incorporating Knowledge Sources into Statistical Speech Recognition addresses the problem of developing efficient automatic speech recognition (ASR) systems, which maintain a balance between utilizing a wide knowledge of speech variability, while keeping the training / recognition effort feasible and improving speech recognition performance. The book provides an efficient general framework to incorporate additional knowledge sources into state-of-the-art statistical ASR systems. It can be applied to many existing ASR problems with their respective model-based likelihood functions in flexible ways.

 

What people are saying - Write a review

We haven't found any reviews in the usual places.

Contents

Introduction and Book Overview
1
12 Approaches to Speech Recognition
4
122 Corpusbased Approaches
6
13 Stateoftheart ASR Performance
7
14 Studies on Incorporating Knowledge Sources
10
142 Existing Ways of Incorporating Knowledge Sources
12
143 Major Challenges to Overcome
15
15 Book Outline
16
Speech Recognition Using GFIKS
79
411 Causal Relationship Between Information Sources
80
412 Inference
81
414 Training and Recognition Issues
82
42 Applying GFIKS at the HMM Phoneticunit Level
83
422 Inference
85
424 Deleted Interpolation
86
43 Experiments with Various Knowledge Sources
87

Statistical Speech Recognition
19
22 Theory of Hidden Markov Models
22
222 General form of an HMM
23
223 Principle Cases of HMM
25
23 Pattern Recognition for HMMBased ASR Systems
35
231 Frontend Feature Extraction
36
232 HMMBased Acoustic Model
43
233 Pronunciation Lexicon
49
234 Language Model
50
235 Search Algorithm
51
Graphical Framework to Incorporate Knowledge Sources
54
31 Graphical Model Representation
56
312 Graphical Model
59
313 Junction Tree Algorithm
63
32 Procedure of GFIKS
68
321 Causal Relationship between Information Sources
70
322 Direct Inference on Bayesian Network
71
323 Junction Tree Decomposition
72
324 Junction Tree Inference
75
332 Different Levels of Incorporation
76
432 Incorporating Knowledge at the HMM Phoneticunit Level
116
44 Experiments Summary and Discussion
132
Conclusions and Future Directions
138
512 Application Issues
140
513 Experimental Issues
141
A Roadmap to a Spoken Language Dialog System
142
Speech Materials
145
A2 TIMIT AcousticPhonetic Speech Corpus
146
A3 Wall Street Journal Corpus
148
A4 ATR Basic Travel Expression Corpus
150
ATR Software Tools
152
B3 SSS Data Generating Tools
155
B5 Language Model Training Tools
157
Composition of Bayesian Widephonetic Context
163
C2 Variants of Bayesian Widephonetic Context Model
164
Statistical Significance Testing
168
D2 The Use of the Sign Test for ASR
172
References
175
Index
189
Copyright

Other editions - View all

Common terms and phrases