Generation, ranking, and delivery of actions for entities in a video delivery system

• US 10,212,464 B2
• Filed: 01/05/2017
• Issued: 02/19/2019
• Est. Priority Date: 04/15/2016
• Status: Active Grant

• Current Assignee: Hulu LLC (The Walt Disney Company)
• Original Assignee: Hulu LLC (The Walt Disney Company)
• Inventors: Kaya, Lutfi Ilke, Tang, Bangsheng, Yang, Tong, Kehler, Christopher Russell, Zhang, Chi
• Primary Examiner(s): Kumar, Pankaj
• Assistant Examiner(s): Newlin, Timothy R
• Application Number:
  US15/399,702
• Publication Number:
  US 20170302979A1
• Time in Patent Office: 775 Days
• CPC Class Codes: G06F 16/70 of video data
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: H04N 21/234 Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
• CPC Class Codes: H04N 21/252 Processing of multiple end-users' preferences to derive collaborative data
• CPC Class Codes: H04N 21/4532 involving end-user characteristics, e.g. viewer profile, preferences monitoring of user activities for profile generation for accessing a video database G06F16/739; user profiles in network data switching protocols H04L67/306; processing of user preferences or user profiles in wireless networks H04W8/18
• CPC Class Codes: H04N 21/4667 Processing of monitored end-user data, e.g. trend analysis based on the log file of viewer selections
• CPC Class Codes: H04N 21/482 End-user interface for program selection broadcast systems using EPGs H04H60/72
• CPC Class Codes: H04N 21/8133 specifically related to the content, e.g. biography of the actors in a movie, detailed information about an article seen in a video program

Abstract

In one embodiment, a method sends videos to a user that is using a video delivery service. The method receives user behavior that includes actions taken by the user on the video delivery service. The method inputs the user behavior into a first predictor to generate a set of actions for a set of entities. Also, the method inputs the set of actions for the set of entities, a real-time context, and the user behavior into a second predictor to generate probabilities for the set of actions for the set of entities. A probability for an action indicates the probability the user would select that action for an entity when compared against other actions in the set of actions for the set of entities. An action feed is selected based on the ranking and dynamically output to a client while the user is using the video delivery service.

Claims

1. A method comprising:

sending, by a computing device, videos to a user that is using a video delivery service on a client;

receiving, by the computing device, user behavior for the user, wherein the user behavior includes actions taken by the user on the video delivery service;

receiving, by the computing device, a real-time context based on the user using the video delivery service;

inputting, by the computing device, the user behavior into a first predictor to generate a set of actions for a set of entities found on the video delivery service, wherein inputting the user behavior into the first predictor to generate the set of actions for the set of entities comprises;

determining, by the computing device, a first candidate set of entities the user has engaged within a first time period using the user behavior;

determining, by the computing device, if a probability the user would re-engage with each of the first candidate set of entities meets a first threshold; and

including, by the computing device, entities in the first candidate set of entities that have the probability the user would re-engage with each of the first candidate set of entities that meets the threshold in the set of entities;

inputting, by the computing device, the set of actions for the set of entities, the real-time context, and the user behavior into a second predictor to generate probabilities for the set of actions for the set of entities, wherein the probability for an action indicates the probability the user would select that action for an entity when compared against other actions in the set of actions for the set of entities;

selecting, by the computing device, an action feed based on the ranking, the action feed including at least a portion of the set of actions for the set of entities; and

dynamically outputting, by the computing device, the action feed to the client while the user is using the video delivery service, wherein an action on an entity in the action feed is performed when selected by the user.

2. The method of claim 1, further comprising:

training the first predictor using the user behavior, wherein the first predictor is trained to output the set of entities based on the probability the user would re-engage with the set of entities.

3. The method of claim 1, further comprising:

determining a second candidate set of entities the user has not engaged within a second time period using the user behavior, the second candidate set of entities determined to be related to entities the user engaged;

determining if the probability the user would engage with the second candidate set of entities meets a second threshold; and

including entities in the second candidate set of entities that have the probability the user would engage with the second candidate set of entities that meets the second threshold in the set of entities.

4. The method of claim 1, further comprising:

training the second predictor using the user behavior, wherein the second predictor is trained to rank actions for entities based on the probability the user would select an action for an entity.

5. The method of claim 4, wherein training the second predictor comprises:

training the second predictor using actions the user selected and actions for entities the user did not select in the user behavior.

6. The method of claim 1, wherein the first predictor generates an entity relationship graph that defines possible entities and relationships between the entities.

7. The method of claim 1, wherein the second predictor uses a function f(u, e, a, c) to generate the probabilities, where u includes the user behavior for the user, e is an entity, a is an action, and c is the real-time context associated with the user.

8. The method of claim 1, wherein selecting the action feed comprises:

filtering the set of actions for the set of entities based on filter criteria to select the at least the portion of the set of actions for the set of entities.

9. The method of claim 1, further comprising:

receiving a selection of an action in the action feed; and

using a link for the action to perform the action on the video delivery service.

10. A method comprising:

sending, by a computing device, videos to a user that is using a video delivery service on a client;

receiving, by the computing device, user behavior for the user, wherein the user behavior includes actions taken by the user on the video delivery service;

receiving, by the computing device, a real-time context based on the user using the video delivery service;

inputting, by the computing device, the user behavior into a first predictor to generate a set of actions for a set of entities found on the video delivery service, wherein the first predictor;

generates an entity relationship graph that defines possible entities and relationships between the entities;

determines which of the entities in the entity relationship graph are available for selection by the user on the video delivery service, and

selects the set of actions for the set of entities from the entities in the entity relationship graph that are available for selection by the usen;

inputting, by the computing device, the set of actions for the set of entities, the real-time context, and the user behavior into a second predictor to generate probabilities for the set of actions for the set of entities, wherein the probability for an action indicates the probability the user would select that action for an entity when compared against other actions in the set of actions for the set of entities;

selecting, by the computing device, an action feed based on the ranking, the action feed including at least a portion of the set of actions for the set of entities; and

dynamically outputting, by the computing device, the action feed to the client while the user is using the video delivery service, wherein an action on an entity in the action feed is performed when selected by the user.

11. A method comprising;

sending, by a computing device, videos to a user that is using a video delivery service on a client;

receiving, by the computing device, user behavior for the user, wherein the user behavior includes actions taken by the user on the video delivery service;

receiving, by the computing device, a real-time context based on the user using the video delivery service;

inputting, by the computing device, the user behavior into a first predictor to generate a set of actions for a set of entities found on the video delivery service;

inputting, by the computing device, the set of actions for the set of entities, the real-time context, and the user behavior into a second predictor to generate probabilities for the set of actions for the set of entities, wherein the probability for an action indicates the probability the user would select that action for an entity when compared against other actions in the set of actions for the set of entities, wherein the second predictor comprises a first ranker configured for performing;

receiving, by the computing device, first actions for manually added entities and second actions for engaged entities, the manually added entities being explicitly engaged by the user in the user behavior and the engaged entities being implicitly engaged by the user in the user behavior; and

ranking, by the computing device, the first actions for manually added entities and the second actions for engaged entities into a set of third actions;

selecting, by the computing device, an action feed based on the ranking, the action feed including at least a portion of the set of actions for the set of entities; and

dynamically outputting, by the computing device, the action feed to the client while the user is using the video delivery service, wherein an action on an entity in the action feed is performed when selected by the user.

12. The method of claim 11, wherein the second predictor comprises a second ranker configured for performing:

receiving fourth actions for discovered entities other than the engaged entities and manually added entities, the discovered entities being determined based a relevance between the engaged entities and the different entities; and

ranking the fourth actions for the discovered entities.

13. The method of claim 12, wherein the second predictor comprises a third ranker configured for performing:

receiving fifth actions for a set of campaigns, the set of campaigns promoting campaign entities; and

ranking the fifth actions for the campaign entities in the set of campaigns.

14. The method of claim 13, wherein the third ranker is configured for performing:

ranking the fifth actions based on a campaign urgency; and

ranking the fifth actions based on entity availability on the video delivery service for each campaign and the campaign urgency.

15. The method of claim 14, wherein the second predictor comprises a fourth ranker configured for performing:

receiving the ranked set of third actions and the ranked fourth actions; and

ranking the ranked set of third actions and the ranked fourth actions together into a set of sixth actions.

16. The method of claim 15, wherein the second predictor comprises a fifth ranker configured for performing:

receiving the set of sixth actions and the ranked fifth actions; and

ranking the set of sixth actions and the ranked fifth actions together into a set of seventh actions, wherein the set of seventh actions is used to select the action feed.

17. A non-transitory computer-readable storage medium containing instructions, that when executed, control a computer system to be configured for:

sending videos to a user that is using a video delivery service on a client;

receiving user behavior for the user, wherein the user behavior includes actions taken by the user on the video delivery service;

receiving a real-time context based on the user using the video delivery service;

inputting the user behavior into a first predictor to generate a set of actions for a set of entities found on the video delivery service, wherein inputting the user behavior into the first predictor to generate the set of actions for the set of entities comprises;

determining a first candidate set of entities the user has engaged within a first time period using the user behavior;

determining if a probability the user would re-engage with each of the first candidate set of entities meet a first threshold; and

including entities in the first candidate set of entities that have the probability the user would re-engage with each of the first candidate set of entities that meets the threshold in the set of entities;

inputting the set of actions for the set of entities, the real-time context, and the user behavior into a second predictor to generate probabilities for the set of actions for the set of entities, wherein the probability for an action indicates the probability the user would select that action for an entity when compared against other actions in the set of actions for the set of entities;

selecting an action feed based on the ranking, the action feed including at least a portion of the set of actions for the set of entities; and

dynamically outputting the action feed to the client while the user is using the video delivery service, wherein an action on an entity in the action feed is performed when selected by the user.

18. The non-transitory computer-readable storage medium of claim 17, wherein:

the first predictor determines which of the entities in the entity relationship graph are available for selection by the user on the video delivery service, and

the first predictor selects the set of actions for the set of entities from the entities in the entity relationship graph that are available for selection by the user.

19. A method comprising:

sending, by a computing device, videos to a user that is using a video delivery service on a client;

receiving, by the computing device, user behavior for the user, wherein the user behavior includes actions taken by the user on the video delivery service;

ranking first actions for manually added entities and second actions for engaged entities into a set of third actions, the manually added entities being explicitly engaged by the user in the user behavior and the engaged entities being implicitly engaged by the user in the user behavior;

ranking fourth actions for discovered entities other than the engaged entities and manually added entities, the discovered entities being determined based a relevance between the engaged entities and the different entities;

ranking fifth actions for a set of campaigns, the set of campaigns promoting campaign entities;

ranking the ranked set of third actions and the ranked fourth actions into a set of sixth actions;

selecting, by the computing device, an action feed based on the ranking of the set of sixth actions and the ranked fifth actions, the action feed including at least a portion of the set of actions for the set of entities; and

dynamically outputting, by the computing device, the action feed to the client while the user is using the video delivery service, wherein an action on an entity in the action feed is performed when selected by the user.

20. The method of claim 19, wherein at least one of the rankings uses a function f(u, e, a, c) to generate probabilities the user would select an action for an entity, where u includes the user behavior for the user, e is the entity, a is the action, and c is a real-time context associated with the user using the video delivery service.