Unsupervised cross-lingual speaker adaptation for HMM-based speech synthesis using two-pass decision tree construction.
Matthew GibsonTeemu HirsimäkiReima KarhilaMikko KurimoWilliam ByrnePublished in: ICASSP (2010)
Keyphrases
- cross lingual
- speech synthesis
- decision tree construction
- speech recognition
- hidden markov models
- language modeling
- language model
- vocal tract
- machine learning
- decision trees
- machine translation
- language independent
- cross language
- text to speech
- text classification
- automatic speech recognition
- unsupervised learning
- semi supervised
- parallel corpus
- supervised learning
- n gram
- transfer learning
- speech signal
- noisy environments
- data mining
- document clustering
- query translation
- parallel corpora
- pattern recognition
- image processing
- document retrieval
- active learning
- computer vision
- neural network