Emformer: Efficient Memory Transformer Based Acoustic Model For Low Latency Streaming Speech Recognition.
Yangyang ShiYongqiang WangChunyang WuChing-Feng YehJulian ChanFrank ZhangDuc LeMichael L. SeltzerPublished in: CoRR (2020)
Keyphrases
- speech recognition
- low latency
- stream processing
- hidden markov models
- highly efficient
- automatic speech recognition
- language model
- pattern recognition
- continuous query processing
- speech processing
- speech synthesis
- speech signal
- speech recognizer
- high bandwidth
- speech recognition technology
- real time
- noisy environments
- speaker identification
- speech recognition systems
- high speed
- high throughput
- main memory
- gaussian mixture model
- speaker adaptation