Configurable and portable system for generating narratives

Patent metadata

Field	Value
Publication number	US-11790164-B2
Application number	US-202016876556-A
Country	US
Kind code	B2
Filing date	May 18, 2020
Priority date	Jan 7, 2011
Publication date	Oct 17, 2023
Grant date	Oct 17, 2023

Abstract

The subject invention functions to use available data and information to automatically create narrative stories that describes domain events, circumstances and/or entities in a comprehensible and compelling and audience customized, manner. Computer executable instructions provide for generating a narrative story using standard and uniform structures and data for receiving domain related data and a story specification, parsing the story specification to provide constituent components, transforming the constituent components into executable code, instantiating content blocks having at least one feature for the domain according to the story specification and rendering the narrative story using the constituent components specified by the content blocks.

First claim

What is claimed is:

1. A method for automatically generating a narrative story about a data set using a narrative generation platform that is horizontally scalable to a plurality of different story types or content verticals, the data set comprising structured data in a plurality of fields and/or categories that describe one or more events, situations, entities, and/or combinations thereof, the method comprising:
adapting the narrative generation platform to a specific story type or a specific content vertical by defining a story specification in response to input from a user, wherein the narrative generation platform includes a modular angle, wherein the modular angle corresponds to a characterization of data, wherein the modular angle is re-usable with respect to a plurality of different story types or content verticals, wherein the adapting step comprises:
parameterizing the modular angle, wherein the parameterizing operates to configure the modular angle for the specific story type or the specific content vertical; and
including a specification of the parameterized modular angle in the story specification, wherein the story specification organizes a narrative structure for the narrative story, the story specification comprising (1) specifications of a plurality of data models and (2) the specification of the parameterized modular angle, wherein the specifications of the data models and the parameterized modular angle are not expressed as program code directly executable by a processor and are not expressed in a general purpose programming language;

processing the story specification to generate executable code that defines an analytical framework for (1) analyzing the structured data based on the parameterized modular angle and (2) finding a summarization of the data set that fits the data set, wherein the summarization includes an analytical interpretation of the analyzed structured data;
executing the executable code to find the summarization that fits the data set by (1) analyzing the structured data based on the parameterized modular angle and (2) determining whether the characterization corresponding to the parameterized modular angle validly characterizes the structured data; and
in response to a determination that the characterization corresponding to the parameterized modular angle validly characterizes the structured data, generating the narrative story in accordance with the narrative structure from the story specification, and wherein the narrative story provides the analytical interpretation of the analyzed structured data by expressing the determined characterization in natural language;
wherein the processing, executing, and generating steps are performed by a processor within the narrative generation platform; and
wherein the adapting step is performed without requiring the user to directly author or edit program code.

2. The method of claim 1 wherein the modular angle is associated with a plurality of applicability condition parameters for testing to determine whether the characterization validly characterizes data, wherein the parameterizing step links the applicability condition parameters to the specific story type or content vertical and one or more elements or derived features of the structured data.

3. The method of claim 2 wherein the executing step comprises testing the linked applicability condition parameters against the structured data or data derived therefrom to determine whether the characterization validly characterizes the structured data.

4. The method of claim 3 wherein the story specification further comprises a specification of a plurality of derived features that are not expressed as program code directly executable by a processor and are not expressed in a general purpose programming language; and
wherein the analytical framework is configured to compute the derived features; and
wherein at least one of the applicability condition parameters is linked to at least one of the derived features.

5. The method of claim 1 wherein the story specification further comprises a plurality of content blocks, wherein each content block defines how a corresponding section of the narrative story is to be generated, wherein each content block includes a specification of a modular angle with a corresponding characterization for consideration with respect to the corresponding section.

6. The method of claim 1 wherein the parameterizing step comprises parameterizing a plurality of modular angles that are re-usable with respect to a plurality of different story types or content verticals, wherein each modular angle corresponds to a different characterization of data, wherein the parameterizing operates to configure the modular angles for the specific story type or the specific content vertical, and wherein the story specification comprises a plurality of specifications of the parameterized modular angles;
wherein the executing step comprises executing the executable code with respect to the structured data to determine whether any of the characterizations corresponding to the parameterized modular angles validly characterize the structured data; and
wherein the generating step comprises, in response to a determination that a plurality of the characterizations corresponding to the parameterized modular angles validly characterize the structured data, generating the narrative story in accordance with the narrative structure from the story specification, and wherein the narrative story expresses, in natural language, a plurality of the characterizations determined to validly characterize the structured data.

7. The method of claim 6 further comprising:
determining an importance value applicable to each of the parameterized modular angles whose corresponding characterization was determined to validly characterize the structured data; and
selecting which of the characterizations determined to validly characterize the structured data are to be expressed in the narrative story based on the determined importance values applicable to the parameterized modular angles corresponding thereto.

8. The method of claim 1 further comprising:
providing a user-interactive tool for authoring and/or editing the story specification in response to input from the user without the necessity of the user authoring, editing, or reviewing computer executable code in a general purpose programming language; and
wherein the parameterizing and including steps are performed in response to user inputs received through the provided user-interactive tool.

9. The method of claim 1 further comprising:
performing the parameterizing and including steps for a plurality of different story specifications to build a library of story specifications for a plurality of different specific story types or specific content verticals; and
wherein the method further comprises (1) selecting a story specification from the library in response to user input and (2) performing the processing, executing, and generating steps with respect to the selected story specification.

10. A computer program product for automatically generating a narrative story about a data set using a narrative generation platform that is horizontally scalable to a plurality of different story types or content verticals, the data set comprising structured data in a plurality of fields and/or categories that describe one or more events, situations, entities, and/or combinations thereof, the computer program product comprising:
a plurality of processor-executable instructions that are resident on a non-transitory computer-readable storage medium, wherein the instructions are configured for execution by a processor within the narrative generation platform to cause the

Assignees

• Narrative Science Inc

Inventors

• Nichols Nathan Drew
• Birnbaum Lawrence A
• Hammond Kristian J

Classifications

• G06F40/20 Primary

  Natural language analysis (semantic analysis of natural language G06F40/30) · CPC title

• G06F40/00

  Handling natural language data (speech analysis or synthesis, speech recognition G10L) · CPC title

• G06F40/35 Primary

  Discourse or dialogue representation · CPC title

• G06F40/205

  Parsing · CPC title