Query construction from conversational inputs
Systems, methods, and other embodiments associated with defining an adaptive model for conversational interaction are described herein. In one embodiment, a computer-implemented method includes semantically processing a conversational input to separate the input into a plurality of structural entities. The method also includes constructing a query of a data provider based on one or more of the structural entities and a query paradigm. The method also includes transmitting the query to the data provider for execution. The method further includes presenting the query results through a user interface.
1. A computer-implemented method, comprising:
semantically processing a conversational input to separate the conversational input into a plurality of structural entities, wherein the semantic processing includes identifying as the structural entities at least an intent, a data object, and a field value;
constructing a query for execution by a data provider by inserting the data object and field value into a query paradigm that fits the intent, a type of the data object, and the field value;
transmitting the query to the data provider for execution; and
presenting results from the execution of the query through a user interface.
2. The computer-implemented method of claim 1 , wherein the query paradigm comprises a structure for constructing queries that is configured to accept the one or more structural entities, wherein constructing the query for execution by the data provider further comprises selecting the query paradigm from among a plurality of query paradigms based on the structural entities.
3. The computer-implemented method of claim 1 , wherein the query paradigm comprises a structure configured to accept the one or more structural entities, wherein constructing the query for execution by the data provider further comprises inserting the one or more structural entities into the query paradigm to produce the query of the data provider.
4. The computer-implemented method of claim 1 ,
wherein semantically processing the conversational input further comprises:
identifying one or more of the structural entities to be the intent,
identifying one or more of the structural entities to be the data object that is of a given type of object,
identifying a plurality of the structural entities to be field values for objects of the type of object,
identifying one or more of the structural entities to be a logical operator; and
wherein constructing the query for execution by the data provider further comprises:
determining that the query paradigm fits the intent, the type of object, the field values, and the logical operator,
selecting the query paradigm to be used to construct the query in response to the determination, and
constructing the query by inserting the object and the field values into the query paradigm.
5. The computer-implemented method of claim 1 , wherein constructing the query for execution by the data provider further comprises resolving an ambiguous value for the intent, an object type, the field value, or a logical operator from a current contextual state.
6. The computer-implemented method of claim 1 , wherein constructing the query for execution by the data provider further comprises selecting to be the query paradigm one query paradigm that is most likely to satisfy the conversational input from a plurality of query paradigms that fit the structural entities.
7. The computer-implemented method of claim 1 , wherein semantically processing the conversational input further comprises recognizing the plurality of structural entities in the conversational input using a machine learning model that has been trained to recognize structural entities in a training set of conversational requests.
8. The computer-implemented method of claim 1 , wherein the data provider provides data interface services for data stores.
9. One or more non-transitory computer-readable media storing computer-executable instructions that, when the instructions are executed by at least a processor of one or more computers, cause the computers to:
semantically process a conversational input received through a conversational user interface to produce a plurality of structural entities, wherein the semantic processing includes identifying as the structural entities at least an intent, a data object, and a field value;
construct a query for execution by a data provider by inserting the data object and field value into a query paradigm that fits the intent, a type of the data object, and the field value;
send the query to the data provider for execution; and
present results from the execution of the query through the conversational user interface.
10. The one or more non-transitory computer-readable media of claim 9 , wherein the query paradigm comprises a structure for constructing queries that is configured to accept the one or more structural entities, wherein the instructions to construct the query for execution by the data provider further cause the computers to select the query paradigm from among a plurality of query paradigms based on the structural entities.
11. The one or more non-transitory computer-readable media of claim 9 , wherein the query paradigm comprises a structure for constructing queries that is configured to accept the one or more structural entities, wherein the instructions to construct the query for execution by the data provider further cause the computers to generate the query of the data provider by inserting the one or more structural entities into the query paradigm.
12. The one or more non-transitory computer-readable media of claim 9 ,
wherein the instructions to semantically process the conversational input further cause the computers to:
identify the intent from the structural entities,
identify an object type from the structural entities, and
identify one or more of the structural entities to be field values for the data object, wherein the data object is of the object type; and
wherein the instructions to construct the query for execution by the data provider further cause the computers to:
determine that the query paradigm fits the intent, the object type, and the field values,
select the query paradigm to be used to construct the query in response to the determination, and
construct the query by inserting at least the field values into the query paradigm.
13. The one or more non-transitory computer-readable media of claim 9 , wherein the instructions to construct the query for execution by the data provider further cause the computers to resolve an ambiguous value for an intent, an object type, a field value, or a logical operator from a current contextual state.
14. The one or more non-transitory computer-readable media of claim 9 , wherein the instructions to construct the query for execution by the data provider further cause the computers to select to be the query paradigm one query paradigm that is most likely to satisfy the conversational input from a plurality of query paradigms that fit the structural entities.
15. The one or more non-transitory computer-readable media of claim 9 , wherein the instructions further cause the computers to:
train a natural language processing model to recognize structural entities in a training set of conversational requests; and
wherein the instructions to semantically process the conversational input further cause the computers to recognize the plurality of structural entities in the conversational input using the trained natural language processing model.
16. A computing system comprising:
a processor;
a memory operably connected to the processor; and
one or more non-transitory computer-readable media storing computer-executable instructions that, when the instructions are executed by at least the processor of the computer system, cause the computer system to:
extract a plurality of structural entities from a conversational input to a conversational user interface, wherein the semantic processing includes identifying as the structural entities at least an intent, a data object, and a field value;
construct a query for execution by a data provider at runtime by inserting the data object and field value into a query paradigm that fits the intent, a type of the data object and the field value;
transmit the query to the data provider for execution; and
present results from the execution of the query through the conversational user interface.
17. The computing system of claim 16 , wherein the query paradigm comprises a structure for constructing queries that is configured to accept the one or more structural entities, wherein the instructions to construct the query for execution by the data provider further cause the computing system to select the query paradigm from among a plurality of query paradigms based on the structural entities.
18. The computing system of claim 16 , wherein the query paradigm comprises a structure for constructing queries that is configured to accept the one or more structural entities, wherein the instructions to construct the query for execution by the data provider further cause the computing system to generate the query of the data provider by inserting the one or more structural entities into the query paradigm.
19. The computing system of claim 16 , wherein the instructions further cause the computers to:
train a natural language processing model to recognize structural entities in a training set of conversational requests and assign types to the structural entities, wherein the types include intent, data object, field value, and logical operator; and
wherein the instructions to extract a plurality of structural entities from the conversational input further cause the computers execute the trained natural language processing model to:
recognize the plurality of structural entities in the conversational input, and
associate the structural entities with the types.
20. The computing system of claim 16 , wherein the processor, the memory, and the computer-readable media are included in a mobile device, and wherein the data provider includes a data source that is remote from the mobile device.