Table 7.

Performance comparison between baseline models and CREAM variants under the high-error condition. The text correction and text–speech matching modules can be implemented with different settings, while the text rescoring module is implemented using BERT

MethodHigh-error scenario
AISHELL-1TEDLIUM-2Libri-otherLibri-clean
Baselines
Top-18.3312.0711.994.64
Oracle4.748.558.142.43
BART7.4311.7711.264.50
FastCorrect6.72
BERT6.0510.379.913.58
CREAM
Text correction moduleTextspeech matching module
BARTASR4.919.489.123.28
BARTMATE5.0510.419.433.93
BARTMIMICA5.2511.3310.744.83
FastCorrectASR5.01
FastCorrectMATE4.90
FastCorrectMIMICA5.02
BARTASR+MATE+MIMICA4.639.298.803.16
FastCorrectASR+MATE+MIMICA4.73

or Create an Account

Close subscription notice
Close access options