Synthetic phrase-level mixed audio, built by splicing monolingual speech, improves code-switch ASR for three Southeast Asian language pairs, but the evaluation sets are largely synthetic.
M ONO Data We utilize monolingual data from four languages: EN, ZH, BM, and TA, incorporating both local and non-local accents
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Can we train ASR systems on Code-switch without real code-switch data? Case study for Singapore's languages
Synthetic phrase-level mixed audio, built by splicing monolingual speech, improves code-switch ASR for three Southeast Asian language pairs, but the evaluation sets are largely synthetic.