Knowledge Distillation via Module Replacing for Automatic Speech Recognition with Recurrent Neural Network Transducer.
Kaiqi ZhaoHieu NguyenAnimesh JainNathan SusanjAthanasios MouchtarisLokesh GuptaMing ZhaoPublished in: INTERSPEECH (2022)
Keyphrases
- recurrent neural networks
- automatic speech recognition
- speech recognition
- neural network
- feed forward
- knowledge base
- complex valued
- speech signal
- hidden markov models
- reservoir computing
- back propagation
- artificial neural networks
- speech retrieval
- noisy environments
- neural model
- recurrent networks
- conversational speech
- spoken words
- echo state networks
- hidden layer
- natural images
- machine learning