Towards end-to-end speech recognition for Chinese Mandarin using long short-term memory recurrent neural networks.
Jie LiHeng ZhangXinyuan CaiBo XuPublished in: INTERSPEECH (2015)
Keyphrases
- recurrent neural networks
- speech recognition
- end to end
- long short term memory
- speech synthesis
- speaker independent
- neural network
- feed forward
- hidden markov models
- automatic speech recognition
- speech signal
- speech recognizer
- language model
- noisy environments
- speech recognition technology
- pattern recognition
- echo state networks
- artificial neural networks
- word segmentation
- congestion control
- speaker identification
- image quality
- speaker dependent
- speech recognition systems
- expert systems
- mobile devices
- english text
- data mining
- artificial intelligence
- feature extraction
- maximum likelihood