Encoder-Decoder Models Can Benefit from Pre-trained Masked Language Models in Grammatical Error Correction.
Masahiro KanekoMasato MitaShun KiyonoJun SuzukiKentaro InuiPublished in: ACL (2020)
Keyphrases
- error correction
- language model
- error control
- probabilistic model
- noisy channel
- statistical language models
- language modeling
- turbo codes
- language modelling
- n gram
- smoothing methods
- error detection
- reed solomon
- pre trained
- ldpc codes
- speech recognition
- relevance model
- low complexity
- channel coding
- neural network
- retrieval model
- active learning
- feature selection