Inverted Alignments for End-to-End Automatic Speech Recognition.
Patrick DoetschMirko HannemannRalf SchlüterHermann NeyPublished in: IEEE J. Sel. Top. Signal Process. (2017)
Keyphrases
- end to end
- automatic speech recognition
- speech recognition
- speech signal
- hidden markov models
- word error rate
- conversational speech
- spoken words
- broadcast news
- spontaneous speech
- wireless ad hoc networks
- congestion control
- admission control
- word recognition
- internet protocol
- noisy environments
- recognition errors
- speech corpus
- speech retrieval
- scalable video
- pattern recognition
- error rate
- speech sounds