Autonomous learning of entity values in artificial intelligence conversational systems

1. A computer system comprising:

a memory storing one or more associations between one or more entities and one or more fields of response data;

a network interface; and

at least one processor coupled to the memory and the network interface and configured to:

receive a skill service request to process an intent, the skill service request specifying at least one entity having an entity value;

generate a fulfillment request to fulfill the intent, the fulfillment request including the entity value;

transmit, via the network interface, the fulfillment request to a fulfillment service that is configured to execute a programmatic action to fulfill the intent using the entity value;

receive, from the fulfillment service via the network interface, fulfillment response data specifying one or more response values of the one or more fields, the fulfillment response data further including an indicator of whether the fulfillment service successfully executed the programmatic action to fulfill the intent;

identify the one or more entities corresponding to the one or more fields within the fulfillment response data using the one or more associations between the one or more entities and the one or more fields of response data stored in the memory;

store, within the memory, the one or more response values of the one or more fields as one or more new entity values of the one or more entities; and

retrain a natural language processor using the one or more new entity values of the one or more entities for processing a further intent.

2. The computer system of claim 1 , the at least one processor being further configured to:

receive, from a thesaurus service via the network interface, a response specifying thesaurus data associated with the one or more response values of the one or more fields, the thesaurus data comprising one or more of a synonym, a verb, and a plural form of the one or more values of the one or more fields; and

store, within the memory, the thesaurus data.

3. The computer system of claim 2 , the at least one processor being further configured to determine, for at least one entity of the one or more entities, whether the system is configured to import thesaurus data regarding the entity.

4. The computer system of claim 1 , the at least one processor being further configured to store the one or more response values of the one or more fields within dictionary data accessible by a spellchecker.

5. The computer system of claim 1 , the at least one processor being further configured to:

receive an utterance from a user in human language expressing the intent;

identify, via execution of the natural language processor, the utterance as expressing the intent;

generate a response to the intent; and

render the response to the user in the human language.

6. The computer system of claim 1 , wherein the intent is a first intent, the at least one processor being further configured to:

receive a request to process a second intent;

generate a fulfillment request to fulfill the second intent, the fulfillment request to fulfill the second intent including at least one parameter value equal to at least one value of the one or more values of the one or more fields;

transmit, via the network interface, the fulfillment request to fulfil the second intent to the fulfillment service; and

receive, via the network interface, a fulfillment response to the second intent, the fulfillment response to the second intent including fulfillment response data derived from the parameter value.

7. The computer system of claim 1 , the at least one processor being further configured to determine, for at least one entity of the one or more entities, whether the system is configured to auto-discover values of the entity.

8. A method of autonomously learning new entity values executed by a conversational system comprising a memory storing one or more associations between one or more entities and one or more fields of response data, the method comprising:

receiving a skill service request to process an intent, the skill service request specifying at least one entity having an entity value;

generating a fulfillment request to fulfill the intent, the fulfillment request including the entity value;

transmitting, via a network interface, the fulfillment request to a fulfillment service that is configured to execute a programmatic action to fulfill the intent using the entity value;

receiving, from the fulfillment service via the network interface, fulfillment response data specifying one or more response values of the one or more fields, the fulfillment response data further including an indicator of whether the fulfillment service successfully executed the programmatic action to fulfill the intent;

identifying the one or more entities corresponding to the one or more fields within the fulfillment response data using the one or more associations between the one or more entities and the one or more fields of response data stored in the memory;

storing, within the memory, the one or more response values of the one or more fields as one or more new entity values of the one or more entities; and

retraining a natural language processor using the one or more new entity values of the one or more entities for processing a further intent.

9. The method of claim 8 , further comprising:

receiving, from a thesaurus service via the network interface, a response specifying thesaurus data associated with the one or more response values of the one or more fields, the thesaurus data comprising one or more of a synonym, a verb, and a plural form of the one or more values of the one or more fields; and

storing, within the memory, the thesaurus data.

10. The method of claim 9 , further comprising determining, for at least one entity of the one or more entities, whether the conversational system is configured to import thesaurus data regarding the entity.

11. The method of claim 8 , further comprising storing the one or more response values of the one or more fields within dictionary data accessible by a spellchecker.

12. The method of claim 8 , further comprising:

receiving an utterance from a user in human language expressing the intent;

identifying, via execution of the natural language processor, the utterance as expressing the intent;

generating a response to the intent; and

rendering the response to the user in the human language.

13. The method of claim 8 , wherein the intent is a first intent, the method further comprising:

receiving a request to process a second intent;

generating a fulfillment request to fulfill the second intent, the fulfillment request to fulfill the second intent including at least one parameter value equal to at least one value of the one or more values of the one or more fields;

transmitting, via the network interface, the fulfillment request to fulfil the second intent to the fulfillment service; and

receiving, via the network interface, a fulfillment response to the second intent, the fulfillment response to the second intent including fulfillment response data derived from the parameter value.

14. The method of claim 8 , further comprising determining, for at least one entity of the one or more entities, whether the conversational system is configured to auto-discover values of the entity.

15. A non-transitory computer readable medium storing executable sequences of instructions to implement an autonomous learning process within a conversational system comprising a memory storing one or more associations between one or more entities and one or more fields of response data, the sequences of instructions comprising instructions to:

receive a skill service request to process an intent, the skill service request specifying at least one entity having an entity value;

generate a fulfillment request to fulfill the intent, the fulfillment request including the entity value;

transmit, via the network interface, the fulfillment request to a fulfillment service that is configured to execute a programmatic action to fulfill the intent using the entity value;

receive, from the fulfillment service via the network interface, fulfillment response data specifying one or more response values of the one or more fields, the fulfillment response data further including an indicator of whether the fulfillment service successfully executed the programmatic action to fulfill the intent;

identify the one or more entities corresponding to the one or more fields within the fulfillment response data using the one or more associations between the one or more entities and the one or more fields of response data stored in the memory;

store, within the memory, the one or more response values of the one or more fields as one or more new entity values of the one or more entities; and

retrain a natural language processor using the one or more new entity values of the one or more entities for processing a further intent.

16. The non-transitory computer readable medium of claim 15 , the sequences of instructions further comprising instructions to:

receive, from a thesaurus service via the network interface, a response specifying thesaurus data associated with the one or more response values of the one or more fields, the thesaurus data comprising one or more of a synonym, a verb, and a plural form of the one or more values of the one or more fields; and

store, within the memory, the thesaurus data.

17. The non-transitory computer readable medium of claim 16 , the sequences of instructions further comprising instructions to determine, for at least one entity of the one or more entities, whether the conversational system is configured to import thesaurus data regarding the entity.

18. The non-transitory computer readable medium of claim 15 , the sequences of instructions further comprising instructions to store the one or more response values of the one or more fields within dictionary data accessible by a spellchecker.

19. The non-transitory computer readable medium of claim 15 , wherein the intent is a first intent, the sequences of instructions further comprising instructions to:

receive a request to process a second intent;

generate a fulfillment request to fulfill the second intent including at least one parameter value equal to at least one value of the one or more values of the one or more fields;

transmit, via the network interface, the fulfillment request to fulfil the second intent to the fulfillment service; and

receive, via the network interface, a fulfillment response to the second intent including fulfillment response data derived from the parameter value.

20. The non-transitory computer readable medium of claim 15 , the sequences of instructions further comprising instructions to determine, for at least one entity of the one or more entities, whether the conversational system is configured to auto-discover values of the entity.