Automatic generation of narratives from data using communication goals and narrative analytics

Patent metadata

Field	Value
Publication number	US-9977773-B1
Application number	US-201414570834-A
Country	US
Kind code	B1
Filing date	Dec 15, 2014
Priority date	Jan 7, 2011
Publication date	May 22, 2018
Grant date	May 22, 2018

Abstract

The exemplary embodiments described herein are related to techniques for automatically generating narratives about data based on communication goal data structures that are associated with configurable content blocks. The use of such communication goal data structures facilitates modes of operation whereby narratives can be generated in real-time and/or interactive manners.

First claim

What is claimed is:

1. A method comprising:
explicitly representing a plurality of communication goals via a plurality of communication goal data structures that are stored in a memory;
storing a plurality of content block data structures in the memory, wherein the content block data structures comprise models that define how narratives are to be generated from data;
establishing associational relationships between the communication goal data structures and the content block data structures; and
a processor automatically generating a plurality of narratives about data in an interactive mode based on a plurality of inputs by (1) receiving an input relating to a request for a narrative story about a set of data, (2) determining a communication goal for the requested narrative based on the received input, (3) accessing a communication goal data structure in the memory based on the determined communication goal, (4) determining the content block data structure associated with the accessed communication goal data structure based on the established associational relationships, (5) accessing the determined content block data structure in the memory, (6) instantiating the accessed content block data structure, the instantiated content block specifying a model for a plurality of data components that need to be analyzed to generate a narrative and a model for a plurality of computational components for analyzing the data components to generate a narrative, (7) automatically generating a narrative about the data based on the instantiated content block data structure as applied to the data, and (8) repeating the receiving, communication goal determining, communication goal data structure accessing, content block data structure determining, content block data structure accessing, instantiating, and automatically generating steps such that as new input is received the processor generates a plurality of different narratives about the data that satisfy a plurality of communication goals determined from the inputs.

2. The method of claim 1 wherein the repeating step comprises the processor generating the narratives iteratively as new input is received.

3. The method of claim 2 further comprising:
storing a plurality of the communication goal data structures in the memory as a communication goal library; and
wherein the communication goal data structure accessing step comprises selecting a communication goal data structure from the communication goal library that corresponds to the determined communication goal.

4. The method of claim 3 wherein each communication goal data structure is associated with a communication goal and comprises data representative of a plurality of parameters for its associated communication goal and data representative of its associated content block data structure.

5. The method of claim 4 wherein each communication goal data structure is further associated with another communication goal data structure and comprises data indicative the associated another communication goal data structure.

6. The method of claim 3 further comprising:
the processor determining a value for a parameter of the accessed communication goal data structure based on the received input; and
wherein the instantiating step comprises the processor instantiating the accessed content block data structure using the determined parameter value.

7. The method of claim 6 wherein the determined parameter value is a specification of a feature to be discussed by the narrative.

8. The method of claim 6 wherein the determined parameter value is specification of a subject for the narrative.

9. The method of claim 6 wherein the input receiving step comprises the processor receiving the input from a user, the user input comprising a query from the user, the method further comprising:
the processor determining a parameter in response to the query; and
wherein the parameter value determining step comprises using the parameter value specification from the query as the determined parameter value.

10. The method of claim 6 wherein the input receiving step comprises the processor receiving the input from a user, the user input comprising a text string from the user, the method further comprising:
the processor determining a parameter in response to the text string; and
wherein the parameter value determining step comprises using the parameter value specification from the text string as the determined parameter value.

11. The method of claim 6 wherein the input receiving step comprises the processor receiving the input from a user, the user input comprising a selection from the user of an option presented in a user interface (UI), the method further comprising:
the processor determining a parameter in response to the UI selection; and
wherein the parameter value determining step comprises using the parameter value specification from the selection of the object in the GUI as the determined parameter value.

12. The method of claim 6 wherein the iteratively generating step comprises the processor generating the narratives as new input is received such that the processor uses a parameter value specification from an instantiated content block data structure for a prior input when instantiating another content block data structure for a new input.

13. The method of claim 6 further comprising:
the processor analyzing the received input to determine if a parameter value needed for narrative generation remains unspecified; and
in response to a determination that a parameter value needed for narrative generation remains unspecified, the processor generating a prompt for communication to a user that is configured to solicit a specification of the remaining parameter value from the user;
the processor receiving a specification of the remaining parameter value from the user in response to the prompt; and
wherein the instantiating step comprises the processor instantiating the accessed content block data structure using the received remaining parameter value.

14. The method of claim 3 further comprising:
storing a plurality of the content block data structures in the memory as a content block library; and
wherein the content block data structure determining step comprises selecting the content block data structure in the content block library that is associated with the accessed communication goal data structure.

15. The method of claim 14 wherein a plurality of the content block data structures are parameterized via a subject parameter such that the parameterized content block data structures pertain to a plurality of different subjects, wherein the accessed content block data structure comprises a parameterized content block data structure, and wherein the instantiating step comprises the processor instantiating the accessed parameterized content block data structure based on a specification of a value for the subject parameter.

16. The method of claim 14 wherein a plurality of the content block data structures are parameterized via a feature parameter such that the parameterized content block data structures pertain to a plurality of different features, wherein the accessed content block data structure comprises a parameterized content block data structure, and wherein the instantiating step comprises the processor instantiating the accessed parameterized content block data structure based on a specification of a value for the feature parameter.

17. The method of claim 2 wherein the accessed communication goal data structure has an association with another communication goal data structure, and

Assignees

• Narrative Science Inc

Inventors

• Birnbaum Lawrence
• Hammond Kristian J
• Nichols Nathan Drew
• Paley Andrew R

Classifications

• G06Q10/0639

  Performance analysis of employees; Performance analysis of enterprise or organisation operations · CPC title

• G06Q30/0201

  Market modelling; Market analysis; Collecting market data · CPC title

• G06F40/253

  Grammatical analysis; Style critique · CPC title

• G06N20/00

  Machine learning · CPC title

• G06Q10/06398

  Performance of employee with respect to a job function · CPC title