PERL is a Pinyin-fused masked rephrasing model with a length predictor that reduces Chinese ASR character error rate on N-best hypotheses.
The con- struction process is summarized as follows: Data Processing:We start from ECSpell data [17], a text do- main error correction dataset
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
PERL: Pinyin Enhanced Rephrasing Language Model for Chinese ASR N-best Error Correction
PERL is a Pinyin-fused masked rephrasing model with a length predictor that reduces Chinese ASR character error rate on N-best hypotheses.