Voice portal hosting system and method

• US 8,457,970 B2
• Filed: 04/13/2001
• Issued: 06/04/2013
• Est. Priority Date: 05/16/2000
• Status: Active Grant

• Current Assignee: InterDigital Patent Holdings, Inc. (InterDigital, Inc.)
• Original Assignee: Swisscom AG (Government of Switzerland)
• Inventors: Van Kommer, Robert
• Primary Examiner(s): Vo, Huyen X.
• Application Number:
  US09/835,237
• Publication Number:
  US 20010055370A1
• Time in Patent Office: 4,435 Days
• Field of Search: 704/270, 704/271, 704272-276, 704/270.1, 704/231, 704/255, 704/244, 704/243, 704/245, 704/250, 704 1- 10, 704/251, 704/257, 705/27, 379/88.01, 379/88.17
• US Class Current: 704/270
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00
• CPC Class Codes: H04M 3/493 Interactive information services, e.g. directory enquiries ; Arrangements therefor, e.g. interactive voice response [IVR] systems or voice portals

Abstract

Voice portal hosting system (2) with which a plurality of users (1) in a first voice telecommunication network (10) can establish a connection with a voice equipment (11),

• • comprising a memory (26) in which a plurality of interactive voice response applications have been independently uploaded through a second telecommunication network (3) by a plurality of independent value-added service providers (4),
  • wherein at least a plurality of said interactive voice response applications uses a common speech recognition module (24) run on said system.

Claims

1. A voice portal hosting system, allowing a plurality of users to establish a connection with said system using voice equipment for interacting with one or more of a plurality of service providers for ordering a product and/or service, said system comprising:

means for independently uploading a plurality of interactive voice response applications from said service providers, to said system, via a communication channel in advance of ordering products or services using the system, each of said voice response applications for providing interactive voice response functionality for a corresponding one of said service providers when executed by said hosting system, wherein said interactive voice response applications include an executable component for execution by said hosting system, said executable component comprising at least one of an executable file, a Java Bean, a Corba-component, a compiled software module, and a pre-compiled software module;

means for storing said plurality of interactive voice response applications;

a common speech recognition module;

means for storing a plurality of user-specific speech models and a plurality of language models adapted to specific users for use by the common speech recognition module, wherein the plurality of language models is updated for each specific user and service access;

a user identification module for identifying a user calling said system via another communication channel;

means for retrieving the user-specific speech model of the identified user from said plurality of models, wherein the identified user interacts with one or more of said interactive voice response applications, and wherein said one or more interactive voice response applications utilize said retrieved user-specific speech model via said common speech recognition module for recognizing speech of the identified user, and further wherein said common speech models are adaptable during dialogue between said users and any of said interactive voice response applications during ordering of products and/or services from the corresponding providers such that said adapted speech models are thereafter utilized by others of said voice response applications for ordering products and/or services of the corresponding other providers.

2. A voice portal hosting system, intended to be connected to a first voice telecommunication network in order for a plurality of users in said network to establish a connection with said system using voice equipment for ordering products and/or services from one of a plurality of providers, said system comprising:

a memory in which a plurality of interactive voice response applications providing interactive voice response functionality is stored, each of said applications including an executable component for execution by said hosting system;

a common speech recognition module;

means for storing a plurality of user-specific speech models and language models adapted to specific users for use by the common speech recognition module, wherein at least one language model is updated for each specific user and service access;

a user identification module for identifying a known user or a new user;

means for retrieving the user-specific speech model of the known user from said plurality of models;

means for updating said user-specific speech models to the new user without using any training phase;

and

uploading means for independently uploading said plurality of interactive voice response applications, to said system, by a plurality of independent value-added service providers, wherein

the identified user interacts with one or more of said interactive voice response applications, and wherein

said one or more interactive voice response applications utilize said retrieved user-specific speech model via said common speech recognition module for recognizing speech of the known user ordering a product and/or service from one of said providers, and for further adapting said user-specific speech model during said ordering such that said adapted user-specific speech model can thereafter be utilized for ordering a product and/or service from another of said providers, and wherein

speaker independent models are used for a new user prior to updating said user-specific speech models to make the new user into a known user.

3. The system of claim 2, wherein each of said interactive voice response applications includes an executable component for execution by said hosting system, said executable component comprising at least one of an executable file, a Java Bean, a Corba-component, a compiled software module, and a pre-compiled software module.

4. The system of claim 2, wherein each of said interactive voice response applications includes an executable component for execution by said hosting system, said executable component comprising at least one of an executable file, a Java Bean, a Corba-component, a compiled software module, and a pre-compiled software module.

5. A method for allowing each of a plurality of value-added service providers to set up an interactive voice response application including an executable component for execution by a voice portal hosting system commonly used by said plurality of value-added service providers, said voice response application for being used by a plurality of users for ordering products and/or services from said providers, said method comprising the steps of:

storing a plurality of user-specific speech models and language models adapted to known users for use by a common speech recognition module, wherein at least one language model is updated for each specific user and service access;

identifying a user calling said system as a known user or a new user;

retrieving the user-specific speech model of the known user from said plurality of models or else retrieving a speaker independent model for the new user and generating a user-specific speech model for the new user without using any training phase;

independently uploading, to said system, said interactive voice response applications which provide interactive voice response functionality;

the identified user interacting with one of said interactive voice response applications; and

said one of said interactive voice response applications using said retrieved user-specific speech model or said retrieved speaker independent speech model via said common speech recognition module for executing on said hosting system for recognizing speech of the known user or the new user, respectively for ordering a product or service from one of said providers, such that a user-specific speech model for the new user is created or the retrieved user-specific speech model of the known user is further adapted during said ordering such that said user-specific speech model is thereafter made available for the new or known user to interact with another of said interactive voice applications for ordering a service or product from another of said providers.

6. A method for allowing each of a plurality of independent value-added service providers to set up a corresponding interactive voice response application including an executable component for execution by a voice portal hosting system commonly used by said plurality of value-added service providers and which can be used by a plurality of users, said method comprising the steps of:

storing said interactive voice response applications from said providers on said system for providing interactive voice response functionality,

storing a plurality of user-specific speech models and language models, adapted to known users for use by a common speech recognition module, wherein at least one language model is updated for each specific user and service access,

identifying a user calling said system as a known user or new user,

retrieving the user-specific speech model of the known user from said plurality of models or retrieving a speaker independent model for a new user and adapting a user specific speech model for the new user,

and

executing one of said voice response applications associated with one of said providers in response to the user calling said system, said executing including interacting with the user via said common speech module using said retrieved user-specific speech model for recognizing the speech of the known user or using said retrieved speaker independent model for the new user, wherein

said common speech models are adapted during each dialogue between said users and any of said interactive voice response applications without using any training phase and wherein

the user-specific speech model for the user is further adapted during said interacting such that said user-specific speech model is thereafter made available for interacting with any other of said voice response applications associated with another of said providers.

7. The method of claim 6, wherein said interactive voice response applications include an executable component for execution by said hosting system, said executable component comprising at least one of an executable file, a Java Bean, a Corba-component, a compiled software module, and a pre-compiled software module.

8. A system for implementing the method of claim 6.

9. A method for allowing each of a plurality of independent value-added service providers to set up a corresponding interactive voice response application including an executable component for execution by a voice portal hosting system commonly used by said plurality of value-added service providers and which can be used by a plurality of users, said method comprising the steps of:

independently uploading, through a second telecommunication network, said interactive voice response applications to said system for providing interactive voice response functionality, wherein said interactive voice response applications include an executable component for execution by said hosting system,

storing a plurality of user-specific speech models and language models, adapted to known users for use by a common speech recognition module, wherein at least one language model is updated for each specific user and service access,

identifying a user calling said system as a known user or new user,

retrieving the user-specific speech model of the known user from said plurality of models or retrieving a speaker independent model for a new user and adapting a user specific speech model for the new user,

executing one or more of said voice response applications in response to the user calling said system, said executing including interacting with the user via said common speech module using said retrieved user specific speech model for recognizing the speech of the known user or using said retrieved speaker independent model for the new user, wherein

said common speech models are incrementally adapted during each dialogue between said users and any of said interactive voice response applications using recording speech samples and without using any training phase such that said adapted models are thereafter made available for use by all of said interactive voice response applications of the providers, and wherein

said common speech recognition module comprises a common user profile database including user preferences.

10. A voice portal hosting system, intended to be connected to a first voice telecommunication network in order for a plurality of users in said network to establish a connection with said system using voice equipment, said system comprising:

a memory in which a plurality of interactive voice response applications providing interactive voice response functionality is stored, each of said applications including an executable component for execution by said hosting system;

a common speech recognition module;

means for storing a plurality of user-specific speech and language models adapted to specific users for use by the common speech recognition module, wherein at least one language model is updated for each specific user and service access;

a user identification module for identifying a user;

means for retrieving the user-specific speech and language model of the identified user from said plurality of models;

and

uploading means for independently uploading said plurality of interactive voice response applications in advance, to said system, by a plurality of independent value added service providers, wherein

the identified user interacts with one of said interactive voice response applications for ordering a product or service from the provider corresponding to the one of said interactive voice response applications, and wherein

said one of said interactive voice response applications utilizes said retrieved user-specific speech and language model via said common speech recognition module for recognizing speech of the identified user during said ordering, and wherein said retrieved user specific speech model is further adapted to the specific user during said ordering from said corresponding service provider such that said further adapted retrieved user-specific speech model is thereafter utilized by any other of said interactive voice response applications for future ordering of products and/or services from the other corresponding service providers.

11. A method for allowing each of a plurality of value-added service providers to set up an interactive voice response application including an executable component for execution by a voice portal hosting system commonly used by said plurality of value-added service providers for selling products and/or services, said voice response application for being used by a plurality of users to order said products and services, said method comprising the steps of:

storing a plurality of user-specific speech and language models adapted to specific users for use by a common speech recognition module, wherein at least one language model is updated for each specific user and service access;

identifying a user calling said system;

retrieving the user-specific speech and language model of the identified user from said plurality of models;

independently uploading, to said system, said interactive voice response applications which provide interactive voice response functionality;

the identified user interacting with one of said interactive voice response applications; and

said one of said interactive voice response applications using said retrieved user-specific speech and language model via said common speech recognition module for executing on said hosting system for recognizing speech of the identified user for ordering products and/or services from the provider corresponding to said one of said interactive voice response applications, and wherein, said retrieved user specific speech model is further adapted to the identified user during said ordering such that said further adapted retrieved user-specific speech model is thereafter utilized for ordering of products and/or services from any others of said service providers using their corresponding interactive voice response application(s).

12. A voice portal hosting system, intended to be connected to a first voice telecommunication network in order for a plurality of users in said network to establish a connection with said system using voice equipment, said system comprising:

a memory in which a plurality of interactive voice response applications providing interactive voice response functionality is stored, each of said applications including an executable component for execution by said hosting system;

a common speech recognition module;

means for storing a plurality of user-specific speech models and language models, adapted to specific users for use by the common speech recognition module, wherein at least one language model is updated for each specific user and service access;

a user identification module for identifying a user;

means for retrieving the user-specific speech model of the identified user from said plurality of models;

and

uploading means for independently uploading said plurality of interactive voice response applications, to said system, by a plurality of independent value-added service providers, wherein

the identified user interacts with one of said interactive voice response applications of a correspond one of said providers, and wherein

said one of said interactive voice response applications utilize said retrieved user-specific speech model via said common speech recognition module for recognizing speech of the identified user, wherein said retrieved user-specific speech model is further adapted during said interacting and is thereafter made available for use by others of said interactive voice response applications of others of said providers, and wherein

said common speech recognition module, said user-specific speech models, and said plurality of interactive voice response applications are all hosted in a single host.

13. A method for allowing each of a plurality of value-added service providers to set up an interactive voice response application including an executable component for execution by a voice portal hosting system commonly used by said plurality of value-added service providers for selling products and/or services, said voice response application for being used by a plurality of users to order said products and services, said method comprising the steps of:

storing a plurality of user-specific speech and language models adapted to specific users for use by a common speech recognition module, wherein at least one language model is updated for each specific user and service access;

identifying a user calling said system;

retrieving the user-specific speech and language model of the identified user from said plurality of models;

independently uploading, to said system, said interactive voice response applications which provide interactive voice response functionality;

the identified user interacting with one of said interactive voice response applications of a corresponding one of said providers; and

said one or more of said interactive voice response applications using said retrieved user-specific speech and language model via said common speech recognition module for executing on said hosting system for recognizing speech of the identified user, wherein said interactive voice response applications include an executable component for execution by said hosting system, and wherein

said retrieved user-specific speech and language model is adapted during said interacting.

14. The method of claim 13, wherein said adapted retrieved user-specific speech and language model is made available for use by all others of said interactive voice response applications of the other providers.