| 摘要 |
A speech analysis system associates speakers with respective segments (fig. 4) of an audio conversation and ranks the speaker as an expert based on the number of times their segments are associated with a term or topic via keyword tags (408) and whether this exceeds a threshold. The level of expertise may be based on an expertise vector (fig. 6) such as communication length or how recently the topic has been referenced. Automatic Speech Recognition (ASR 202), Spoken Language Understanding (SLU 204) and Dialog Management (DM 206) modules convert the speech to text, provide tags or indexes for the speech and interpret the texts meaning, and then Spoken Language Generation (SLG 208) and synthesising (210) modules respond with audible speech. Speaker identity tags may be provided via a Lightweight Directory Access Protocol (LDAP) directory. |
