Applied artificial intelligence technology for narrative generation based on a conditional outcome framework
Artificial intelligence (AI) technology can be used in combination with composable communication goal statements to facilitate a user's ability to quickly structure story outlines in a manner usable by an NLG narrative generation system without any need for the user to directly author computer code. Narrative analytics that are linked to communication goal statements can employ a conditional outcome framework that allows the content and structure of resulting narratives to intelligently adapt as a function of the nature of the data under consideration. This AI technology permits NLG systems to determine the appropriate content for inclusion in a narrative story about a data set in a manner that will satisfy a desired communication goal.
1. An artificial intelligence apparatus for generating a narrative from structured data according to a narrative generation process, the structured data comprising a plurality of data values associated with a plurality of data parameters, the apparatus comprising:
a processor configured to:
parameterize a model based on (1) user input identifying a communication goal for analyzing a specified group of entities and (2) a plurality of the data parameters;
execute the parameterized model with respect to the structured data, wherein as part of the execution the processor is further configured to (1) determine a group characteristic for the specified entity group based on the structured data and (2) summarize a subset of the structured data selected automatically based on the determined group characteristic to determine information to include in the narrative; and
generate a narrative about the structured data based on the execution operation, wherein the narrative expresses the determined at least one idea in natural language and satisfies the communication goal.
2. The artificial intelligence apparatus recited in claim 1 , the processor being further configured to:
display a graphical user interface for allowing a user to specify the communication goal, the communication goal being one of a plurality of communication goals that may be specified via the graphical user interface, the user input being provided via the graphical user interface.
3. The artificial intelligence apparatus of claim 1 wherein the processor is further configured to select the model to be parameterized from among a plurality of models based on the communication goal for analyzing the specified group of entities with respect to a specified attribute.
4. The artificial intelligence apparatus of claim 1 wherein the communication goal for analyzing the specified group of entities with respect to a specified attribute is identified based on data comprising a communication goal statement.
5. The artificial intelligence apparatus of claim 1 wherein as part of the execution operation the processor is further configured to:
select a plurality of idea data structures within the model based on the determined group characteristic, wherein each selected idea data structure corresponds to a different idea and has an associated condition; and
evaluate the conditions of the selected idea data structures based on the structured data to determine which of the selected idea data structures is applicable to the structured data and thus serves as the determined at least one idea.
6. The artificial intelligence apparatus of claim 1 wherein the group characteristic comprises a determined group distribution characteristic.
7. The artificial intelligence apparatus of claim 6 wherein the model comprises a plurality of conditional outcome data structures corresponding to different characterizations for a distribution of members of the specified entity group with respect to values of the specified attribute.
8. The apparatus of claim 7 wherein the determined group distribution characteristic comprises a typical distribution.
9. The apparatus of claim 7 wherein the determined group distribution characteristic comprises a clump at top distribution.
10. The apparatus of claim 7 wherein the determined group distribution characteristic comprises a flat distribution.
11. The apparatus of claim 7 wherein the determined group distribution characteristic comprises a long-tail distribution.
12. The apparatus of claim 7 wherein the determined group characteristic comprises a determined group size characteristic.
13. The apparatus of claim 12 wherein the model comprises a plurality of conditional outcome data structures corresponding to different characterizations for a size of the specified entity group.
14. The apparatus of claim 1 wherein the processor is further configured to:
parameterize the model based on an ontology, wherein the ontology comprises a plurality of ontological objects that define a knowledge base about the structured data for the narrative generation process, wherein the ontological objects include a plurality of entities, wherein the specified group of entities comprises entities within the ontology; and
execute the parameterized model with respect to the structured data and the ontology.
15. The apparatus of claim 14 wherein the ontological objects further comprise a plurality of entity types, and wherein the specified entity group comprises a plurality of entities within the ontology that share a common entity type.
16. The apparatus of claim 15 wherein the ontology associates the entity types with a plurality of expressions for the entity types, and wherein the narrative describes the entities within the specified entity group using an expression associated by the ontology with the entity type corresponding to the specified entity group.
17. The apparatus of claim 14 wherein the ontological objects further comprise a plurality of attributes associated with the entities.
18. A method for generating a narrative from structured data according to a narrative generation process, the structured data comprising a plurality of data values associated with a plurality of data parameters, the method comprising:
parameterizing a model based on (1) user input identifying a communication goal for analyzing a specified group of entities and (2) a plurality of the data parameters;
executing the parameterized model with respect to the structured data via a processor, wherein as part of the execution the processor is further configured to (1) determine a group characteristic for the specified entity group based on the structured data and (2) summarize a subset of the structured data selected automatically based on the determined group characteristic to determine information to include in the narrative; and
generating a narrative about the structured data based on the execution operation, wherein the narrative expresses the determined at least one idea in natural language and satisfies the communication goal.
19. The method recited in claim 18 , the method further comprising:
displaying a graphical user interface for allowing a user to specify the communication goal, the communication goal being one of a plurality of communication goals that may be specified via the graphical user interface, the user input being provided via the graphical user interface.
20. One or more non-transitory computer readable media having instructions stored thereon for performing a method for generating a narrative from structured data according to a narrative generation process, the structured data comprising a plurality of data values associated with a plurality of data parameters, the method comprising:
parameterizing a model based on (1) user input identifying a communication goal for analyzing a specified group of entities and (2) a plurality of the data parameters;
executing the parameterized model with respect to the structured data via a processor, wherein as part of the execution the processor is further configured to (1) determine a group characteristic for the specified entity group based on the structured data and (2) summarize a subset of the structured data selected automatically based on the determined group characteristic to determine information to include in the narrative; and
generating a narrative about the structured data based on the execution operation, wherein the narrative expresses the determined at least one idea in natural language and satisfies the communication goal.