End-to-End Neural Speaker Diarization with Permutation-Free Objectives.
Yusuke FujitaNaoyuki KandaShota HoriguchiKenji NagamatsuShinji WatanabePublished in: CoRR (2019)
Keyphrases
- end to end
- speaker diarization
- network architecture
- speech recognition
- neural network
- congestion control
- admission control
- speaker verification
- machine learning
- broadcast news
- unsupervised learning
- bayesian information criterion
- transport layer
- speaker identification
- application layer
- non stationary
- hidden markov models