CTC-internal scores saturate for hypothesis selection on LibriSpeech, but RoBERTa-based MBR decoding cuts WER by 0.535 pp while MWER training fails near convergence.
CR-CTC: Consistency regularization on CTC for improved speech recognition
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
The Anatomy of the CTC Oracle Gap: Acoustic Exhaustion and Linguistic Recovery
CTC-internal scores saturate for hypothesis selection on LibriSpeech, but RoBERTa-based MBR decoding cuts WER by 0.535 pp while MWER training fails near convergence.