Attention-based alignment of non-native and native phone sequences creates compact mispronunciation lexicons that improve English ASR for Korean speakers by around 13% relative WER.
Acoustic data-driven pronun- ciation lexicon generation for logographic languages,
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
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Data-Driven Mispronunciation Pattern Discovery for Robust Speech Recognition
Attention-based alignment of non-native and native phone sequences creates compact mispronunciation lexicons that improve English ASR for Korean speakers by around 13% relative WER.