Language-oriented user interfaces for voice activated services

• US 20030046086A1
• Filed: 08/12/2002
• Published: 03/06/2003
• Est. Priority Date: 12/07/1999
• Status: Abandoned Application

• Current Assignee: Comverse Incorporated (Mavenir Group Holdings Ltd.)
• Original Assignee: Comverse Network Systems, Inc. (Verint Systems Incorporated)
• Inventors: Yuschik, Matthew John
• Application Number:
  US10/216,190
• Publication Number:
  US 20030046086A1
• Time in Patent Office: Days
• US Class Current: 704/275
• CPC Class Codes: G10L 15/065 Adaptation
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding

Abstract

A comprehensive system is provided for designing a voice activated user interface (VA UI) having a semantic and syntactic structure adapted to the culture and conventions of spoken language for the intended users. The system decouples the content dimension of speech (semantics) and the manner-of-speaking dimension (syntax) in a systematic way. By decoupling these dimensions, the VA UI can be optimized with respect to each dimension independently and jointly. The approach is general across languages and encompasses universal variables of language and culture. Also provided are voice activated user interfaces with semantic and syntactic structures so adapted, as well as a prompting grammar and error handling methods adapted to such user interfaces.

Claims

1. A method for selecting a vocabulary set for a voice activated user interface, the method comprising:

collecting responses to task-oriented questions eliciting commonly used names for tasks and task-related items; and

selecting a plurality of responses from the collected responses based on frequency of occurrence in the collected responses.

2. A method as recited in claim 1, further comprising:

dividing the selected responses into sub-vocabularies;

testing each sub-vocabulary for pairwise acoustic similarity among the responses thereof; and

replacing a specified response of a sub-vocabulary with an alternative one of the collected when the specified response is acoustically similar to another response in the sub-vocabulary.

3. A method as recited in claim 1, wherein collecting the responses to the task-oriented questions includes posing the task-oriented questions in a spoken conversational language of a target community to speakers of the spoken conversational language from the target community.

4. A method as recited in claim 1, further comprising:

dividing the selected responses into sub-vocabularies corresponding to command menus for a target service application of the user interface; and

testing a pair of the selected responses in a sub-vocabulary for pairwise acoustic similarity.

10. A method as recited in claim 1, further comprising modifying a first selected response for consistency with a second selected response when the first and second selected responses are grammatically distinct and refer to similar service functions for the user interface.

5. A computer system providing a service through a voice activated user interface, said computer system comprising:

a storage having a vocabulary of command words stored therein, each command word being selected from responses to questions posed to members of a test group; and

a processor interpreting a spoken response based on the stored command words.

6. A computer system as recited in claim 5, wherein the vocabulary includes plural sub-vocabularies each consisting of command words selected from the responses to be pairwise acoustically dissimilar.

7. A computer system as recited in claim 5, wherein each command word is selected from the responses based on frequency of occurrence in the responses.

8. A computer system as recited in claim 5, wherein said processor executes a command corresponding to the spoken response when the spoken response matches a corresponding command word of the vocabulary.

9. A computer system as recited in claim 5, wherein the service is a voice controlled voice mail service.

11. A method as recited in claim 11, wherein the first and second selected responses refer to service functions in different sub-menus of the user interface.

12. A computer readable medium encoded with a program for a voice activated user interface, said program comprising instructions for:

interpreting a spoken response based on a vocabulary of stored command words each selected from responses to questions posed to members of a test group; and

performing a command function based on the interpreted response.