| 摘要 |
PURPOSE:To obtain the natural language dialog device which can vocalize overlapping with the vocalization of a speaker (overlap vocalization). CONSTITUTION:This device is equipped with a vocalization hypothesis generation part 2 which outputs vocalization hypothesis information showing at least one hypothesis of the contents of natural language input on the basis of language analytic information of a language analysis part 1 that gradually analyzes the natural language input from a user and outputs language analytic information showing the analytic result of syntax information, rhythm information, or meaning information, a language generation part 4 which converts the answer contents information of an answer contents determination part 3, outputting the answer contents information showing answer contents on the basis of the vocalization hypothesis information into natural language output, and an idle talk management part 5 which judges whether instantaneous vocalization to the natural language input of the user or vocalization overlapping with the natural language input of the user with time is performed on the basis of the operation state of at least one of the language analysis part 1, vocalization hypothesis generation part 2, answer contents determination part 3, and language generation part 4 and displays the natural language output of the language generation part 4 to the user when judging that the vocalization is performed. |
