Variable-rate hierarchical CPC leads to acoustic unit discovery in speech.
Santiago CuervoAdrian LancuckiRicard MarxerPawel RychlikowskiJan ChorowskiPublished in: NeurIPS (2022)
Keyphrases
- speech recognition systems
- acoustic features
- speech recognition
- prosodic features
- speech sounds
- emotional speech
- text to speech
- acoustic signal
- automatic speech recognition
- speaker independent
- speaker verification
- hierarchical model
- audio visual
- knowledge discovery
- coarse to fine
- speech synthesis
- hidden markov models
- discovery process
- speech signal
- emotion recognition
- recognition engine
- formant frequencies
- endpoint detection
- acoustic models
- neural network
- underwater vehicles
- causal discovery
- pattern discovery
- hierarchical clustering
- hierarchical structure
- probabilistic model
- pattern recognition