WaveTract: A hybrid generative model for speech synthesis.
Brunó Bence EnglertCsaba ZainkóGéza NémethPublished in: SpeD (2019)
Keyphrases
- generative model
- speech synthesis
- speech recognition
- probabilistic model
- text to speech
- bayesian framework
- vocal tract
- discriminative learning
- em algorithm
- prosodic features
- prior knowledge
- posterior probability
- semi supervised
- markov chain monte carlo
- computer vision
- discriminative models
- latent dirichlet allocation
- topic models
- learning algorithm
- language model
- hidden markov models
- active learning
- bayesian networks
- similarity measure
- generative and discriminative models