Systems and methods to predict hashtags for content items

Patent metadata

Field	Value
Publication number	US-10698945-B2
Application number	US-201514829537-A
Country	US
Kind code	B2
Filing date	Aug 18, 2015
Priority date	Aug 18, 2015
Publication date	Jun 30, 2020
Grant date	Jun 30, 2020

Abstract

Systems, methods, and non-transitory computer readable media configured to acquire data associated with a content item, the data associated with the content item including contextual information. The data associated with the content item can be provided to a model trained by machine learning. A set of hashtags associated with the content item can be determined based on the model.

First claim

What is claimed is:

1. A computer-implemented method comprising:
acquiring, by a computing system, training data associated with content items, wherein the content items are associated with hashtags already created by users;
training, by the computing system, a machine learning model based on the training data and the hashtags associated with the content items;
upon new content items associated with new hashtags becoming available, refining, by the computing system, the machine learning model based on new data and the new hashtags associated with the new content items; and
in response to detection of keystrokes typed by a user who intends to create a hashtag for a content item,
determining, by the computing system, a set of hashtags associated with the content item based on the machine learning model, and
presenting, by the computing system, one or more hashtags of the set of hashtags for selection by the user through an user interface, wherein the one or more hashtags are consistent with the intent as indicated by the keystrokes of the user.

2. The computer-implemented method of claim 1 , wherein the training data associated with the content items further includes at least one of text information, user information, an image, or video associated with the content item.

3. The computer-implemented method of claim 1 , wherein the determining the set of hashtags further comprises:
determining a confidence value for each hashtag in the set of hashtags; and
sorting the set of hashtags based on confidence values.

4. The computer-implemented method of claim 3 , wherein the determining the set of hashtags further comprises:
selecting a subset of the set of hashtags based on a predetermined threshold.

5. The computer-implemented method of claim 4 , further comprising:
providing the subset of the set of hashtags to a client computing device.

6. The computer-implemented method of claim 5 , wherein one or more hashtags from the subset of the set of hashtags are presented for selection in a user interface displayed by the client computing device based on keystrokes typed by a user.

7. The computer-implemented method of claim 1 , wherein the machine learning includes use of a neural network.

8. The computer-implemented method of claim 1 , wherein the training data associated with the content items includes contextual information, wherein the contextual information includes at least one of time of day, day of week, week of year, or location associated with creation of the content item.

9. A system comprising:
at least one processor; and
a memory storing instructions that, when executed by the at least one processor, cause the system to perform:
acquiring training data associated with content items, wherein the content items are associated with hashtags already created by users;
training a machine learning model based on the training data and the hashtags associated with the content items;
upon new content items associated with new hashtags becoming available, refining the machine learning model based on new data and the new hashtags associated with the new content items; and
in response to detection of keystrokes typed by a user who intends to create a hashtag for a content item,
determining a set of hashtags associated with the content item based on the machine learning model, and
presenting one or more hashtags of the set of hashtags for selection by the user through an user interface, wherein the one or more hashtags are consistent with the intent as indicated by the keystrokes of the user.

10. The system of claim 9 , wherein the training data associated with the content items further includes at least one of text information, user information, an image, or video associated with the content item.

11. The system of claim 9 , wherein the determining the set of hashtags further comprises:
determining a confidence value for each hashtag in the set of hashtags; and
sorting the set of hashtags based on confidence values.

12. The system of claim 11 , wherein the determining the set of hashtags further comprises:
selecting a subset of the set of hashtags based on a predetermined threshold.

13. The system of claim 12 , further comprising:
providing the subset of the set of hashtags to a client computing device.

14. The sysetm of claim 9 , wherein the training data associated with the content items includes contextual information, wherein the contextual information includes at least one of time of day, day of week, week of year, or location associated with creation of the content item.

15. A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method comprising:
acquiring training data associated with content items, wherein the content items are associated with hashtags already created by users;
training a machine learning model based on the training data and the hashtags associated with the content items;
upon new content items associated with new hashtags becoming available, refining the machine learning model based on new data and the new hashtags associated with the new content items; and
in response to detection of keystrokes typed by a user who intends to create a hashtag for a content item,
determining a set of hashtags associated with the content item based on the machine learning model, and
presenting one or more hashtags of the set of hashtags for selection by the user through an user interface, wherein the one or more hashtags are consistent with the intent as indicated by the keystrokes of the user.

16. The non-transitory computer-readable storage medium of claim 15 , wherein the training data associated with the content items further includes at least one of text information, user information, an image, or video associated with the content item.

17. The non-transitory computer-readable storage medium of claim 15 , wherein the determining the set of hashtags further comprises:
determining a confidence value for each hashtag in the set of hashtags; and
sorting the set of hashtags based on confidence values.

18. The non-transitory computer-readable storage medium of claim 17 , wherein the determining the set of hashtags further comprises:
selecting a subset of the set of hashtags based on a predetermined threshold.

19. The non-transitory computer-readable storage medium of claim 18 , further comprising:
providing the subset of the set of hashtags to a client computing device.

20. The non-transitory computer-readable storage medium of claim 14 , wherein the training data associated with the content items includes contextual information, wherein the contextual information includes at least one of time of day, day of week, week of year, or location associated with creation of the content item.

Assignees

• Facebook Inc

Inventors

• State Bogdan
• Herda{Umlaut Over (G)}Delen Amaç
• Boucher Maxime
• Weinsberg Ehud

Classifications

• G06N20/00

  Machine learning · CPC title

• G06F16/438

  Presentation of query results · CPC title

• G06F16/381

  using identifiers, e.g. barcodes, RFIDs (for URLs G06F16/9554) · CPC title

• G06F16/48 Primary

  Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually · CPC title