ICASSP
2011
IEEE
13 years 4 months ago
2011
IEEE
Long-span language models that capture syntax and semantics are seldom used in the first pass of large vocabulary continuous speech recognition systems due to the prohibitive sea...