Multimedia content publishing method and apparatus, and electronic device and storage medium

Patent metadata

Field	Value
Publication number	US-12182195-B2
Application number	US-202118029074-A
Country	US
Kind code	B2
Filing date	Sep 8, 2021
Priority date	Sep 30, 2020
Publication date	Dec 31, 2024
Grant date	Dec 31, 2024

Abstract

Provided are a multimedia content publishing method and apparatus, and an electronic device and a storage medium. The method includes: determining multimedia content to be published; acquiring at least one candidate time-sensitivity tag that matches the multimedia content; determining at least one target time-sensitivity tag selected from the at least one candidate time-sensitivity tag; and generating multimedia content publishing information that includes the target time-sensitivity tag. In the disclosure, candidate time-sensitivity tags are provided for a publisher of multimedia content, such that the publisher selects a target time-sensitivity tag therefrom; and the added target time-sensitivity tag can be used for reference when related search results are provided for a user who initiates a search request, thereby improving the accuracy of the search results.

First claim

What is claimed is:

1. A method of publishing multimedia content to an online platform, comprising:
determining multimedia content to be published, wherein the multimedia content to be published comprises multimedia content uploaded by a client computing device;
acquiring at least one candidate time-sensitivity tag that matches the multimedia content, wherein the at least one time candidate time-sensitivity tag is dynamically updated based on user real-time search data;
determining at least one target time-sensitivity tag selected from the at least one candidate time-sensitivity tag; and
generating multimedia content publishing information that comprises the target time-sensitivity tag.

2. The method according to claim 1 , wherein after generating multimedia content publishing information that comprises the target time-sensitivity tag, the method further comprises:
publishing, in response to a multimedia content publishing request, the generated multimedia content publishing information that comprises the target time-sensitivity tag.

3. The method according to claim 1 , wherein the acquiring at least one candidate time-sensitivity tag that matches the multimedia content comprises:
acquiring, in response to a time-sensitivity tag acquisition operation, at least one candidate time-sensitivity tag that matches the multimedia content; or,
acquiring at least one candidate time-sensitivity tag that matches the multimedia content after determining that the multimedia content is time-sensitive content based on content attribute information and/or author attribute information associated with the multimedia content.

4. The method according to claim 1 , wherein the determining multimedia content to be published comprises:
acquiring multimedia content uploaded by a target user and title information added for the multimedia content, and taking the multimedia content uploaded by the target user and the title information as the multimedia content to be published.

5. A multimedia content publishing method of publishing multimedia content to an online platform, comprising:
acquiring multimedia content to be published;
generating and dynamically updating a candidate time-sensitivity tag set based on user real-time search data;
selecting at least one candidate time-sensitivity tag that matches the multimedia content from the candidate time-sensitivity tag set, and returning the selected at least one candidate time-sensitivity tag to a first client computing device;
receiving multimedia content publishing information that comprises at least one target time-sensitivity tag, the at least one target time-sensitivity tag belonging to the selected at least one candidate time-sensitivity tag; and
publishing the multimedia content based on the multimedia content publishing information.

6. The method according to claim 5 , further comprising:
storing target time-sensitivity tags associated with the multimedia content;
finding, if a search request initiated by a second client is received, a target time-sensitivity tag matching the search request from the stored target time-sensitivity tags associated with the multimedia content; and
pushing multimedia content associated with the found target time-sensitivity tag to the second client.

7. The method according to claim 5 , wherein the selecting at least one candidate time-sensitivity tag that matches the multimedia content from a candidate time-sensitivity tag set comprises:
determining a correlation between the multimedia content and each time- sensitivity tag in the candidate time-sensitivity tag set; and
selecting at least one candidate time-sensitivity tag from the candidate time- sensitivity tag set based on the correlation.

8. The method according to claim 7 , wherein the determining a correlation between the multimedia content and each time-sensitivity tag in the candidate time-sensitivity tag set comprises:
extracting a multimedia feature vector from the multimedia content, and extracting a text feature vector from each time-sensitivity tag in the candidate time-sensitivity tag set;
determining a vector correlation between the multimedia feature vector and each text feature vector; and
determining a correlation between the multimedia content and each time-sensitivity tag in the candidate time-sensitivity tag set based on each determined vector correlation.

9. The method according to claim 8 , wherein the multimedia content to be published comprises multimedia content uploaded by the first client and title information added for the multimedia content, and
wherein the extracting a multimedia feature vector from the multimedia content comprises:
extracting a content feature vector from the multimedia content uploaded by the first client, and extracting a text feature vector from the title information added for the multimedia content; and
determining the extracted content feature vector and the text feature vector as the multimedia feature vector.

10. The method according to claim 8 , wherein the determining a vector correlation between the multimedia feature vector and each text feature vector comprises:
determining the vector correlation between the multimedia feature vector and each text feature vector by using a trained correlation model.

11. The method according to claim 10 , wherein the correlation model is trained based on the following steps:
acquiring multiple historical search terms and a multimedia content search result associated with each historical search term;
for each multimedia content search result, taking a historical search term associated with the multimedia content search result as a positive class time-sensitivity tag of the multimedia content search result, and taking a historical search term associated with other multimedia content search results other than the multimedia content search result as a negative class time-sensitivity tag of the multimedia content search result; and
taking each multimedia content search result, the positive class time-sensitivity tag of the each multimedia content search result, and the negative class time-sensitivity tag of the each multimedia content search result as one group of training sample data, and training the correlation model to be trained based on a plurality of groups of training sample data to obtain the trained correlation model.

12. The method according to claim 11 , wherein the negative class time-sensitivity tag of each multimedia content search result is determined based on the following steps:
adding the same identifier for the same historical search term of each group of training sample data; and
for each multimedia content search result, determining other multimedia content search result whose corresponding identifier is different from the identifier of the historical search term associated with the multimedia content search result, and taking the historical search term associated with the determined other multimedia content search results as the negative class time-sensitivity tag of the multimedia content search result.

13. The method according to claim 5 , further comprising:
taking time-sensitivity tags with the semantic similarity being greater than a predetermined threshold in the candidate time-sensitivity tag set as one time-sensitivity tag group;
for each time-sensitivity tag group, calculating a word overlap degree between each time-sensitivity tag in the time-sensitivity tag group and other time-sensitivity tags in the time-sensitivity tag group except for the time-sensitivity tag, and updating the time-sensitivity tag group based on a plurality of word overlap degrees obtained by the calculation to obtain

Assignees

• Beijing Bytedance Network Tech Co Ltd

Inventors

• Su Kai
• Wang Changhu

Classifications

• G06F16/44

  Browsing; Visualisation therefor · CPC title

• G06F16/45

  Clustering; Classification · CPC title

• G06F16/7867

  using information manually generated, e.g. tags, keywords, comments, title and artist information, manually generated time, location and usage information, user ratings · CPC title

• G06F16/489 Primary

  using time information · CPC title

• G06F16/7844 Primary

  using original textual content or text extracted from visual content or transcript of audio data · CPC title