Spoken language understanding

The spoken language understanding (SLU) is aimed at interpreting user's intentions from speech utterances. Although the process is closely related to natural language understanding (NLU), it focuses only on specific application domains. To address this limitation, statistical models have been formulated as extensions to hidden Markov models (HMMs) as segment models, where a multiple block with word dependency is generated from each underlying Markov state corresponding to individual semantic slot defined from application domain. As a result, a natural bridge between automatic speech recognition (ASR) and SLU is achieved.