An Ultra-Low Power Binarized Convolutional Neural Network-Based Speech Recognition Processor With On-Chip Self-Learning.
Shixuan ZhengPeng OuyangDandan SongXiudong LiLeibo LiuShaojun WeiShouyi YinPublished in: IEEE Trans. Circuits Syst. I Regul. Pap. (2019)
Keyphrases
- speech recognition
- ultra low power
- low power
- single chip
- high speed
- low cost
- power consumption
- language model
- hidden markov models
- neural network
- speech signal
- pattern recognition
- speech recognition systems
- speech recognizer
- speech synthesis
- speech processing
- speech understanding
- cmos image sensor
- noisy environments
- automatic speech recognition
- speech recognition errors
- speech recognition technology
- processor core
- ibm power processor
- speech recognizers
- cmos technology
- keyword spotting
- speaker identification
- parallel processing
- document images
- speaker diarization
- speaker independent
- maximum likelihood
- image sensor
- bayesian networks
- isolated word
- machine learning