Knowledge Transfer from Pre-trained Language Models to Cif-based Speech Recognizers via Hierarchical Distillation.
Minglun HanFeilong ChenJing ShiShuang XuBo XuPublished in: CoRR (2023)
Keyphrases
- language model
- knowledge transfer
- pre trained
- speech recognition
- language modeling
- knowledge sharing
- n gram
- transfer learning
- document retrieval
- training data
- retrieval model
- probabilistic model
- query expansion
- information retrieval
- training examples
- test collection
- automatic speech recognition
- relevance model
- neural network
- multi modal
- word segmentation
- out of vocabulary