Empirically combining unnormalized NNLM and back-off N-gram for fast N-best rescoring in speech recognition.
Yongzhe ShiWei-Qiang ZhangMeng CaiJia LiuPublished in: EURASIP J. Audio Speech Music. Process. (2014)
Keyphrases
- speech recognition
- n gram
- language model
- language modeling
- language independent
- probabilistic model
- document retrieval
- automatic speech recognition
- query expansion
- speech processing
- test collection
- information retrieval
- speech signal
- variable length
- retrieval model
- text classification
- hidden markov models
- speech recognizer
- speech recognition systems
- noisy environments
- word segmentation
- speaker identification
- speech synthesis
- pattern recognition
- part of speech
- speech recognition technology
- artificial intelligence
- text retrieval
- question answering
- training data
- character n grams
- machine learning