Paper: Are Some Speech Recognition Errors Easier to Detect than Others?

ACL ID N07-2040
Title Are Some Speech Recognition Errors Easier to Detect than Others?
Venue Human Language Technologies
Session Short Paper
Year 2007
Authors

This study investigates whether some speech recognition (SR) errors are eas- ier to detect and what patterns can be identified from those errors. Specifically, SR errors were examined from both non- linguistic and linguistic perspectives. The analyses of non-linguistic properties re- vealed that high error ratios and consecu- tive errors lowered the ease of error detec- tion. The analyses of linguistic properties showed that ease of error detection was as- sociated with changing parts-of-speech of reference words in SR errors. Addition- ally, syntactic relations themselves and the change of syntactic relations had impact on the ease of error detection.