Use of Global and Acoustic Features Associated with Contextual Factors to Adapt Language Models for Spontaneous Speech Recognition.
Shohei ToyamaDaisuke SaitoNobuaki MinematsuPublished in: INTERSPEECH (2017)
Keyphrases
- speech recognition
- language model
- acoustic features
- speech signal
- automatic speech recognition
- contextual factors
- language modeling
- n gram
- probabilistic model
- noisy environments
- word error rate
- test collection
- query expansion
- contextual information
- mel frequency cepstral coefficients
- speaker identification
- broadcast news
- information retrieval
- retrieval model
- information seeking
- query terms
- mixture model
- hidden markov models
- speaker verification
- information extraction
- classification accuracy
- context dependent
- noise reduction
- visual features
- out of vocabulary
- neural network