System and method for aligning ASR model weights with NLU concepts

1. A computer-implemented method comprising:

identifying, by a computing device, at least one language model component of a plurality of language model components in at least one application associated with automatic speech recognition (ASR) and natural language understanding (NLU) usage;

receiving, from a user via a user interface, a contribution bias for the at least one language model component, the contribution bias comprising a weight entered by the user and assigned to the at least one language model component; and

aligning the ASR and NLU between the plurality of language model components based upon, at least in part, the contribution bias, wherein aligning the ASR and the NLU includes aligning together vocabulary, coverage of possible sentence structure, and coverage of concepts and variants of the coverage concepts.

2. The computer-implemented method of claim 1 wherein the at least one application includes a pre-trained language model with the at least one language model component of the plurality of language model components.

3. The computer-implemented method of claim 1 wherein the plurality of language model components includes a general language model component.

4. The computer-implemented method of claim 1 wherein the plurality of language model components includes a domain specific language model component.

5. The computer-implemented method of claim 1 wherein the at least one language model component of the plurality of language model components includes a common domain language model component.

6. The computer-implemented method of claim 1 further comprising configuring the at least one application with the plurality of language model components at run-time through an explicit directive.

7. The computer-implemented method of claim 1 further comprising configuring the at least one application with the plurality of language model components at run-time through an identifying tuple.

8. A computer program product residing on a non-transitory computer readable storage medium having a plurality of instructions stored thereon which, when executed across one or more processors, causes at least a portion of the one or more processors to perform operations comprising:

identifying at least one language model component of a plurality of language model components in at least one application associated with automatic speech recognition (ASR) and natural language understanding (NLU) usage;

receiving, from a user via a user interface, a contribution bias for the at least one language model component, the contribution bias comprising a weight entered by the user and assigned to the at least one language model component; and

aligning the ASR and NLU between the plurality of language model components based upon, at least in part, the contribution bias, wherein aligning the ASR and the NLU includes aligning together vocabulary, coverage of possible sentence structure, and coverage of concepts and variants of the coverage concepts.

9. The computer program product of claim 8 wherein the at least one application includes a pre-trained language model with the at least one language model component of the plurality of language model components.

10. The computer program product of claim 8 wherein the plurality of language model components includes a general language model component.

11. The computer program product of claim 8 wherein the plurality of language model components includes a domain specific language model component.

12. The computer program product of claim 8 wherein the at least one language model component of the plurality of language model components includes a common domain language model component.

13. The computer program product of claim 8 wherein the operations further comprise configuring the at least one application with the plurality of language model components at run-time through an explicit directive.

14. The computer program product of claim 8 wherein the operations further comprise configuring the at least one application with the plurality of language model components at run-time through an identifying tuple.

15. A computing system including one or more processors and one or more memories configured to perform operations comprising:

identifying at least one language model component of a plurality of language model components in at least one application associated with automatic speech recognition (ASR) and natural language understanding (NLU) usage;

receiving, from a user via a user interface, a contribution bias for the at least one language model component, the contribution bias comprising a weight entered by the user and assigned to the at least one language model component; and

aligning the ASR and NLU between the plurality of language model components based upon, at least in part, the contribution bias, wherein aligning the ASR and the NLU includes aligning together vocabulary, coverage of possible sentence structure, and coverage of concepts and variants of the coverage concepts.

16. The computing system of claim 15 wherein the at least one application includes a pre-trained language model with the at least one language model component of the plurality of language model components.

17. The computing system of claim 15 wherein the plurality of language model components includes a general language model component.

18. The computing system of claim 15 wherein the plurality of language model components includes a domain specific language model component.

19. The computing system of claim 15 wherein the at least one language model component of the plurality of language model components includes a common domain language model component.

20. The computing system of claim 15 wherein the operations further comprise configuring the at least one application with the plurality of language model components at run-time through one of an explicit directive and an identifying tuple.