Emitting Word Timings with HMM-Free End-to-End System in Automatic Speech Recognition.
Xianzhao ChenHao NiYi HeKang WangZejun MaZongxia XiePublished in: Interspeech (2021)
Keyphrases
- end to end
- automatic speech recognition
- word error rate
- hidden markov models
- word recognition
- speech recognition
- compound words
- conversational speech
- speech signal
- speech recognizers
- recognition errors
- spontaneous speech
- speech recognizer
- speech recognition systems
- broadcast news
- noisy environments
- admission control
- handwritten words
- congestion control
- speech retrieval
- n gram
- handwriting recognition
- speaker adaptation
- pattern recognition
- word level
- acoustic features
- non stationary
- transport layer
- speaker independent
- language model
- multiresolution
- neural network
- speech sounds