Voice interactive service system and method for providing different speech-based services
View Patent ↗A voice interactive service system provides different speech-based services to a plurality of users. Using a communication terminal, the services are accessed via a telecommunication network through service-specific connectivity ports. The system comprises processing cores which have different configurations of speech processing resources for performing different services. For performing a requested service, a connection module establishes a connection between the respective connectivity port and a processing core having a configuration of speech processing resources suitable for performing the requested service. Because of the service-specific resourcing of cores, there is no need for requesting and allocating processing resources from external resource servers. Moreover, the port-dedicated resourcing of the cores ensures that a successful access to a connectivity port leads to a successful provision of the requested service.
1. A voice interactive service system providing different speech-based services to a plurality of users accessing the system in each case with a communication terminal via a telecommunication network through one of a plurality of connectivity ports, the system comprising a plurality of processing cores, wherein
the connectivity ports are service-specific connectivity ports, wherein:
each of the connectivity ports is associated with at least one of the different speech-based services, and
associating of the connectivity ports with the speech-based services is performed before the service-specific connectivity ports receive user service requests from the plurality of users via the telecommunication network;
the processing cores are hardware based processing units having different configurations of speech processing resources which are stored in memory associated with the respective processing core, the configuration of one service-specific processing core comprising the speech processing resources for performing one of the different speech-based services associated with at least one of the service-specific connectivity ports; and
for performing a specific one of the different speech-based services requested by one of the users through a respective one of the service-specific connectivity ports, the system comprises means for connecting the respective one of the service-specific connectivity ports to a selected one of the service-specific processing cores having a configuration of speech processing resources for performing the specific one of the speech-based services.
2. The system of claim 1 , wherein the configurations of the processing cores include in each case at least one of a service-specific set and a port-specific set of speech processing resources stored permanently in memory associated with the respective processing core.
3. The system of claim 1 , wherein the speech processing resources comprise at least one of speech recognition resources, language-specific speech recognition resources, speech synthesis resources, language-specific speech synthesis resources, and dialogue control resources.
4. The system of claim 1 , wherein the system comprises a data store having stored therein user-specific speech processing parameters; and a personalization module configured to load into memory associated with the selected one of the processing cores the processing parameters associated with the one of the users requesting the speech-based service, for a personalized configuration of the speech processing resources.
5. The system of claim 4 , wherein the user-specific speech processing parameters comprise at least one of a voice model, a vocabulary, terminal characteristics, dialogue experience, service history, and synthetic voice preferences.
6. The system of claim 1 , wherein the processing cores are in each case a core of a multi-core processor; and the processing cores located on the same multi-core processor have the same configuration of speech processing resources.
7. The system of claim 1 , wherein the means for connecting the connectivity ports to a selected one of the processing cores comprise at least one of fixed connections linking in each case one service-specific connectivity port to one service-specific processing core, fixed connections linking in each case one connectivity port to one port-specific processing core, and a matrix switch interconnecting service-specific connectivity ports to service-specific processing cores.
8. A method of providing different voice interactive speech-based services to a plurality of users accessing the services in each case with a communication terminal via a telecommunication network through one of a plurality of connectivity ports, comprising:
establishing the connectivity ports as service-specific connectivity ports, wherein:
each of the connectivity ports is associated with at least one of the different speech-based services, and
associating of the connectivity ports with the speech-based services is performed before the service-specific connectivity ports receive user service requests from the plurality of users via the telecommunication network;
associating a plurality of processing cores, which are hardware based processing units, with different configurations of speech processing resources, which are stored in memory associated with the respective processing core, the configuration of one service-specific processing core comprising the speech processing resources for performing one of the different speech-based services associated with at least one of the service-specific connectivity ports; and
performing a specific one of the different speech-based services, requested by one of the users through a respective one of the service-specific connectivity ports, by connecting the respective one of the service-specific connectivity ports to a selected one of the service-specific processing cores having a configuration of speech processing resources for performing the specific one of the speech-based services.
9. The method of claim 8 , wherein the method further comprises storing permanently in memory associated in each case with one of the processing cores a configuration of the respective processing core including at least one of a service-specific set and a port-specific set of speech processing resources associated with the respective processing core.
10. The method of claim 8 , wherein the speech processing resources comprise at least one of speech recognition resources, language-specific speech recognition resources, speech synthesis resources, language-specific speech synthesis resources, and dialogue control resources.
11. The method of claim 8 , wherein the method further comprises storing in a data store user-specific speech processing parameters; and loading into memory associated with the selected one of the processing cores the processing parameters associated with the one of the users requesting the speech-based service, for a personalized configuration of the speech processing resources.
12. The method of claim 11 , wherein the user-specific speech processing parameters comprise at least one of a voice model, vocabulary, terminal characteristics, dialogue experience, service history, and synthetic voice preferences.
13. The method of claim 8 , wherein the processing cores are implemented on a multi-core processor; and the same configuration of speech processing resources is associated with the processing cores located on the same multi-core processor.
14. The method of claim 8 , wherein connecting the respective one of the connectivity ports to a selected one of the processing cores includes at least one of linking in each case one service-specific connectivity port to one service-specific processing core through a fixed connection, linking in each case one connectivity port to one port-specific processing core through a fixed connection, and interconnecting service-specific connectivity ports to service-specific processing cores through a matrix switch.