Table 8 Results obtained with ULT+WRE approach compared with the baseline. Command Error Rate depending on the weight reestimation rules (MLE and FD) with ULT. 11 136 tests performed on VODIS corpus (noisy).
From: Compact Acoustic Models for Embedded Speech Recognition
ULT+WRE | Baseline | ||
|---|---|---|---|
MLE | FD | ||
Very compact model | 5.25% | 5.11% | 5.80% |
Compact model | 4.01% | 4.27% | 4.80% |