Investigation of End-To-End Speaker-Attributed ASR for Continuous Multi-Talker Recordings.
Naoyuki KandaXuankai ChangYashesh GaurXiao-Fei WangZhong MengZhuo ChenTakuya YoshiokaPublished in: CoRR (2020)
Keyphrases
- end to end
- automatic speech recognition
- speech recognition
- spontaneous speech
- acoustic features
- audio visual
- high bandwidth
- admission control
- multipath
- speaker verification
- congestion control
- wireless ad hoc networks
- internet protocol
- scalable video
- noisy environments
- speech signal
- content delivery
- ad hoc networks
- language model
- transport layer