TECHNIQUES FOR PROVIDING A NATURAL LANGUAGE NARRATIVE

• US 20130174026A1
• Filed: 12/28/2011
• Published: 07/04/2013
• Est. Priority Date: 12/28/2011
• Status: Active Grant

• Current Assignee: CBS Interactive Inc. (ViacomCBS Incorporated)
• Original Assignee: CBS Interactive Inc. (ViacomCBS Incorporated)
• Inventors: Locke, Matthew
• Granted Patent: US 10,540,430 B2
• Time in Patent Office: Days
• US Class Current: 715/254
• CPC Class Codes: G06F 40/10 Text processing natural language analysis G06F40/20; semantic analysis G06F40/30; processing or translation of natural language G06F40/40
• CPC Class Codes: G06F 40/166 Editing, e.g. inserting or deleting
• CPC Class Codes: G06F 40/253 Grammatical analysis; Style critique
• CPC Class Codes: G06Q 10/10 Office automation; Time management

Abstract

Various embodiments for providing a natural language narrative are described. Embodiments may include a summary engine that collects data about an event, and selects a template to narrate a story about the event. A template may include static text, slots for data from the event, and slots for dynamic word choice. Templates may be selected according to a story type and story “flavor.” Templates may also include conditions that determine a choice of the use of template sub-sections. Various aspects of the templates may be selected randomly, to create the impression that the story is human-generated. The stories may be published to a variety of locations and formats. Other embodiments are described and claimed.

Claims

1. A computer-implemented method comprising:

receiving data about an event;

selecting a narrative template for the event comprising empty slots;

filling a data slot in the narrative template with a data item;

filling a word slot in the narrative template with a word selected from a word list for the word slot; and

publishing the narrative template with the slots filled.

2. The method of claim 1, wherein the data comprises at least one of:

a score;

a participant name;

a date of the event;

a time of the event;

a characteristic of a participant;

a position of a participant;

a team name;

a characteristic of an action taken during the event;

statistical data about a participant'"'"'s performance during the event;

statistical data about a team'"'"'s performance during the event;

play-by-play data;

historical data about a team;

a result of a participant action;

an election result; and

historical data about a participant.

3. The method of claim 1, wherein the template comprises a condition, the method further comprising:

determining a state of the condition according to the data; and

selecting a first template sub-section to include when the condition is true and a second template sub-section to include when the condition is false.

4. The method of claim 3, wherein a template sub-section comprises at least one of:

a word, a sentence, and a paragraph.

5. The method of claim 1, further comprising:

randomly selecting a word from the word list to fill the empty word slot.

6. The method of claim 5, wherein the word list is bound to a data item in the event data and a word in the word list comprises a margin, the method further comprising:

selecting a word from the word list when the data item to which the word list is bound is within a margin of the selected word; and

randomly selecting a word from a different word list when the data item is not within the margin.

7. The method of claim 1, wherein the event is one of:

a high school sports event;

a little league sports event;

a recreational league sports event;

a college sports event;

a professional sports event;

an amateur sports event;

a team sports event;

an individual sports event;

a contest;

a reality television show;

a concert;

a live event;

a game;

a tournament;

a political event; and

a fund raising event.

8. An apparatus comprising:

a logic device;

a data store to store narrative templates and communicatively coupled to the logic device;

a summary engine executing on the logic device to automatically;

select a narrative template for an event comprising empty slots;

fill a first slot in the narrative template with a data item from data received about the event; and

fill a second slot in the narrative template with a word selected according to the data received about the event; and

a publishing engine executing on the logic device to;

publish the narrative template with the slots filled.

9. The apparatus of claim 8, the publishing engine further to publish to at least one of:

a website;

a real simple syndication feed;

an e-mail message;

a news ticker;

a social network site;

a short messaging service message;

a multimedia messaging service message;

a text-to-speech telephone message;

an audio stream; and

a radio broadcast.

10. The apparatus of claim 8, the narrative template further comprising:

a story type, comprising a plurality of story flavors, wherein a story flavor comprises a plurality of paragraph types, wherein a paragraph type comprises a plurality of paragraph definitions, wherein a paragraph definition comprises at least one of;

a static sentence, and a dynamic sentence comprising at least one of a word slot and a data slot.

11. The apparatus of claim 10, the summary engine further to:

select a story type from a plurality of story types according to criteria;

select a story flavor randomly from the story flavors of the selected story type;

select a paragraph definition randomly for each paragraph type of the selected story flavor; and

construct a master template from the selected paragraph definitions.

12. The apparatus of claim 11, wherein the master template includes a condition, the summary engine further to:

determine a state of the condition according to the data; and

select a first template sub-section to include when the condition is true and a second template sub-section to include when the condition is false.

13. The apparatus of claim 12, wherein a template sub-section comprises at least one of:

a word, a sentence, and a paragraph definition.

14. The apparatus of claim 12, wherein a story type comprises a word list, and a paragraph definition comprises a word slot associated with the word list, the summary engine further to randomly select a word from the word list to fill the word slot.

15. A machine-readable storage medium comprising instructions that when executed cause a computing system to:

receiving data about an event;

select a story type;

construct a master template according to the story type for the event comprising empty slots;

fill the empty slots of the master template with data from the event to generate a narrative summary; and

publish the narrative summary.

16. The storage medium of claim 15, wherein a story type comprises a plurality of story flavors, wherein a story flavor comprises a plurality of paragraph types, wherein a paragraph type comprises a plurality of paragraph definitions, wherein a paragraph definition comprises at least one of:

a static sentence, and a dynamic sentence comprising at least one of;

a word slot and a data slot.

17. The storage medium of claim 16, further comprising instructions that when executed cause the computing system to:

select a story flavor randomly from the story flavors of the selected story type;

select a paragraph definition randomly for each paragraph type of the selected story flavor; and

construct the master template from the selected paragraph definitions.

18. The storage medium of claim 16, further comprising instructions that when executed cause the computing system to:

fill a data slot in the master template with a data item from the received data; and

fill a word slot in the master template with a word randomly selected from a word list associated with the word slot.

19. The storage medium of claim 15, further comprising instructions that when executed cause the computing system to format the narrative summary to publish to at least one of:

a website;

a real simple syndication feed;

an e-mail message;

a news ticker;

a social network site;

a short messaging service message;

a multimedia messaging service message;

a text-to-speech telephone message;

an audio stream; and

a radio broadcast.

20. The storage medium of claim 16, wherein the master template includes a condition, the storage medium further comprising instructions that when executed cause the computing system to:

determine a state of the condition according to the data; and

select a first template sub-section to include when the condition is true and a second template sub-section to include when the condition is false.