Integrated speech recognition and semantic classification

摘要: A novel system integrates speech recognition and semantic classification, so that acoustic scores in a recognizer accepts spoken utterances may be taken into account when training both language models classification models. For example, joint association score defined is indicative of correspondence class word sequence for an signal. The incorporate parameters such as weighting signal-to-class modeling the signal, model scores, scores. revised to raise target with relative competitor class. designed errors data are minimized.