Lightly Supervised vs. Semi-supervised Training of Acoustic Model on Luxembourgish for Low-resource Automatic Speech Recognition.
Karel VeselýCarlos SeguraIgor SzökeJordi LuqueJan CernockýPublished in: INTERSPEECH (2018)
Keyphrases
- supervised training
- automatic speech recognition
- speech recognition
- unsupervised learning
- conditional random fields
- hidden markov models
- recognition process
- training phase
- supervised learning
- speech signal
- conversational speech
- neural nets
- training algorithm
- speech retrieval
- broadcast news
- dependency parsing
- language model
- higher order
- learning algorithm
- information retrieval
- question answering
- graphical models
- feature space
- pattern recognition
- machine learning