On-line spotting of ASR errors by means of prosody

Given the state of the art of current speech technology, errors are unavoidable in present spoken dialogue systems. Therefore, one of the main concerns in dialogue design is how to decide whether or not the system has understood the user correctly. We identify methods to distinguish between correctly and incorrectly recognized utterances, scored by hand for semantic concept accuracy, using acoustic/prosodic characteristics. The analysis was performed on data collected during independent experiments done with an interactive voice response system that provides travel information over the phone