CN116028721A - 5G message pushing system - Google Patents
5G message pushing system Download PDFInfo
- Publication number
- CN116028721A CN116028721A CN202310329665.6A CN202310329665A CN116028721A CN 116028721 A CN116028721 A CN 116028721A CN 202310329665 A CN202310329665 A CN 202310329665A CN 116028721 A CN116028721 A CN 116028721A
- Authority
- CN
- China
- Prior art keywords
- module
- data
- message
- user
- recall
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000004927 fusion Effects 0.000 claims description 21
- 238000001914 filtration Methods 0.000 claims description 19
- 230000006399 behavior Effects 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 12
- 238000000034 method Methods 0.000 claims description 11
- 238000012163 sequencing technique Methods 0.000 claims description 9
- 238000013486 operation strategy Methods 0.000 claims description 7
- 238000007499 fusion processing Methods 0.000 claims description 6
- 230000001960 triggered effect Effects 0.000 claims description 4
- 238000010801 machine learning Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000003993 interaction Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 230000036651 mood Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Information Transfer Between Computers (AREA)
Abstract
The invention discloses a 5G message pushing system, and belongs to the technical field of 5G message pushing systems. A5G message pushing system comprises a 5G message pushing system, wherein the 5G message pushing system comprises an enterprise data module, a big data platform, a pushing module and a user, the output end of the enterprise data module is connected with the input end of the big data platform, the output end of the big data platform is connected with the input end of the pushing module, the big data platform can enter a coarse ranking module in a ranking module after being fused by a candidate set, and at the moment, the fine ranking module needs to infer and screen data in the candidate set of the coarse ranking module according to the click rate of the user on the message, the multi-target reading time length, the sharing and other recommendation indexes related to the user, so that the accurate pushing of the message to the user is ensured, and information overload can not occur, so that the user is difficult to accurately find valuable information for the user in the pushed messages.
Description
Technical Field
The invention relates to the technical field of 5G message pushing systems, in particular to a 5G message pushing system.
Background
The 5G message, namely a rich media communication (RCS) message, enables us to realize various rich applications of APP on a short message interface, namely, the 5G message is multimedia and can be used for interactive service, and the 5G message service is divided into two information interactions, one is between individual users, and the other is between enterprises and individual users; for the individual user, the 5G message breaks through the length limitation of the traditional short message on each piece of information, the content aspect breaks through the text limitation, the effective fusion of the information such as text, picture audio, video, position and the like is realized, meanwhile, for the enterprise, the 5G message provides an information interaction interface with the individual user, and the enterprise can output personalized service and consultation to the user in a rich media mode such as text, voice, tab and the like; at present, 5G is combined with rich media information, real-time interaction, one-stop service and other services for providing information, more and more information contents and services are transmitted and pushed to users, so that serious information overload is easy to generate, and the users can hardly accurately find valuable information in the pushed information.
Disclosure of Invention
1. Technical problem to be solved
The invention aims to provide a 5G message pushing system, which aims to solve the problems in the background art:
the 5G message, namely a rich media communication (RCS) message, enables us to realize various rich applications of APP on a short message interface, namely, the 5G message is multimedia and can be used for interactive service, and the 5G message service is divided into two information interactions, one is between individual users, and the other is between enterprises and individual users; for the individual user, the 5G message breaks through the length limitation of the traditional short message on each piece of information, the content aspect breaks through the text limitation, the effective fusion of the information such as text, picture audio, video, position and the like is realized, meanwhile, for the enterprise, the 5G message provides an information interaction interface with the individual user, and the enterprise can output personalized service and consultation to the user in a rich media mode such as text, voice, tab and the like; at present, 5G is combined with rich media information, real-time interaction, one-stop service and other services for providing information, more and more information contents and services are transmitted and pushed to users, so that serious information overload is easy to generate, and the users can hardly accurately find valuable information in the pushed information.
2. Technical proposal
The 5G message pushing system comprises an enterprise data module, a big data platform, a pushing module and a user, wherein the output end of the enterprise data module is connected with the input end of the big data platform, the output end of the big data platform is connected with the input end of the pushing module, the output end of the pushing module is connected with the input end of the user, and the output end of the user is connected with the input end of the big data platform;
the big data platform comprises a message content source module and a message content library module, wherein the output end of the enterprise data module is connected with the input end of the message content source module, the output end of the message content source module is connected with the input end of the message content library module, and the output end of the message content library module is connected with the input end of the pushing module;
the pushing module comprises a data module, a recall module, a fusion filtering module and a sequencing module, wherein the output end of the data module is connected with the input end of the recall module, the output end of the recall module is connected with the input end of the fusion filtering module, the output end of the fusion filtering module is connected with the input end of the sequencing module, and the output end of the sequencing module is connected with the input end of a user.
Preferably, the message content source module takes text, video, graphics context, audio and the like authored by an enterprise content producer on a platform as a content source, wherein data in the content source have large differences, the data in the content source are subjected to normalization processing, and the platform is subjected to unified processing of the content, including a label system, a theme, categories and the like and fed back to an application party.
Preferably, the message content library module performs content understanding on a message content source and constructs a message content library, the message content library module processes titles and data of a large amount of text, audio, video and other contents in the message content source, and the message content library module classifies and sorts the contents to obtain a content library, wherein the content understanding mainly comprises text understanding, multimedia understanding, content tendency and target estimation.
Preferably, the data module cleans and processes the data in the message content library module by using various data processing tools and falls into different systems, and the data in the message content library module is used in the system modules such as a feedback algorithm model, a user portrait and the like.
Preferably, the recall module adopts a multi-path recall architecture, the recall module refers to data in different systems of the data module in several major categories such as model categories, attributes, manual intervention, heuristics, operation strategies and the like, and the recall module triggers the recall strategy and generates an initial candidate set through reference conditions such as user attributes, user portraits, historical behaviors, real-time behaviors, geographic positions and the like.
Preferably, the fusion filtering module fuses the candidate sets generated by different recall strategies in the recall module, and the fusion filtering module filters the content of the candidate set which is screened out by the recall module and does not meet the condition.
Preferably, the sorting module reorders the screened candidate sets by using a machine learning model, the sorted candidate sets are divided into a coarse sorting module and a fine sorting module, the candidate sets in the coarse sorting module are generally ten thousands of pieces of data, the candidate sets in the fine sorting module are generally thousands of pieces of data, the fine sorting module needs to infer according to the click rate of a user on candidate messages, multi-target reading time, sharing and other recommendation indexes related to the user, and the fine sorting module screens the data in the coarse sorting module candidate sets.
A5G message pushing method comprises the following steps:
s1: firstly, a text, a video, an image-text, an audio and the like created by a content producer in a platform in an enterprise data module are used as a content source and stored in a message content source module in a big data platform, and meanwhile, the click rate of a message, multi-target reading time length, sharing and other data content of recommendation indexes related to the user are stored in the message content source module in the big data platform.
S2: and the message content source module in the big data platform normalizes the data stored in the content source of the big data platform and feeds the normalized data back to the application party, and then the message content library module classifies and sorts the data in the message content source to obtain a content library.
S3: according to the click rate of the user on the message, the multi-target reading time length, the sharing and other data of recommendation indexes which are closely related to the user are taken as references, the data module in the pushing module cleans and processes the data of the message content library module in the big data platform and falls into different systems, and then the recall module is utilized to refer to the data in different systems of the data module by referring to the model class, the attribute, the manual intervention, the heuristic class, the operation strategy and other large classes, so that the recall module triggers the recall strategy and generates an initial recommendation candidate set through the reference conditions of the attribute, the user portraits, the historical behaviors, the real-time behaviors, the geographic positions and the like of the user.
S4: when the recall strategy is triggered in the recall module and an initial recommended candidate set is generated, the fusion filtering module fuses the candidate sets generated by different recall strategies in the recall module, and screens and filters out the unconditional content candidate set in the candidate set fusion process.
S5: when the candidate sets are fused, the candidate sets enter a coarse ranking module in the ranking module, and at the moment, the fine ranking module needs to infer and screen data in the coarse ranking module candidate sets according to the click rate of the user on the message, the multi-target reading time length, the sharing and other recommendation indexes which are closely related to the user, so that the accurate pushing of the message to the user is ensured.
3. Advantageous effects
Compared with the prior art, the invention has the advantages that:
1) According to the invention, the rough ranking module in the ranking module can be entered after the candidate sets are fused, and at the moment, the fine ranking module needs to infer and screen data in the rough ranking module candidate sets according to the click rate of the messages, multi-target reading time, sharing and other recommendation indexes which are closely related to the users by referring to the users, so that the messages are pushed to the users accurately, and therefore valuable information for the users is difficult to find in the pushed messages accurately due to information overload.
2) In the invention, the candidate sets generated by different recall strategies in the recall module are fused through the fusion filtering module, and the unconditional content candidate sets are screened and filtered in the fusion process of the candidate sets, so that the condition of pushing the junk message to the user can be avoided, and the condition that the mood is influenced because the junk message is received by the user can be avoided.
3) In the invention, the message is pushed to the user by using the pushing system, so that the APP applet does not need to be downloaded, thereby obtaining the message required by the user, saving the memory of the mobile phone of the user on one hand, reducing the possibility of personal information of the user on the other hand, and simultaneously improving the convenience of the user on a certain basis.
Drawings
FIG. 1 is a schematic diagram of the overall system design of the present invention;
fig. 2 is a schematic flow chart of a push module system according to the present invention.
Detailed Description
Examples: referring to fig. 1-2, a 5G message pushing system includes an enterprise data module, a big data platform, a pushing module and a user, wherein an output end of the enterprise data module is connected with an input end of the big data platform, an output end of the big data platform is connected with an input end of the pushing module, an output end of the pushing module is connected with an input end of the user, and an output end of the user is connected with an input end of the big data platform;
the big data platform comprises a message content source module and a message content library module, wherein the output end of the enterprise data module is connected with the input end of the message content source module, the output end of the message content source module is connected with the input end of the message content library module, and the output end of the message content library module is connected with the input end of the pushing module;
the pushing module comprises a data module, a recall module, a fusion filtering module and a sequencing module, wherein the output end of the data module is connected with the input end of the recall module, the output end of the recall module is connected with the input end of the fusion filtering module, the output end of the fusion filtering module is connected with the input end of the sequencing module, and the output end of the sequencing module is connected with the input end of a user.
Specifically, the rough ranking module in the ranking module can be entered after the candidate sets are fused, at this time, the fine ranking module needs to infer and screen data in the rough ranking module candidate sets according to the click rate of the messages, multi-target reading time, sharing and other recommendation indexes which are closely related to the users by referring to the users, so that accurate pushing of the messages to the users is ensured, and therefore information overload cannot occur, the users are difficult to accurately find information valuable to the users in the pushed messages, and the candidate sets are screened data sets.
The message content source module takes texts, videos, graphics context, audios and the like created by an enterprise content producer on a platform as content sources, wherein the data in the content sources have large differences, the data in the content sources are subjected to normalization processing, and the platform is subjected to unified processing of the content, comprises a label system, a theme, categories and the like and feeds the unified processing back to an application party.
The message content library module is used for carrying out content understanding on a message content source and constructing a message content library, the message content library module is used for processing titles and data of a large amount of texts, audios, videos and the like in the message content source, the message content library module is used for classifying and sorting the content to obtain a content library, the content understanding mainly comprises text understanding, multimedia understanding and content tendency, and release target estimation, wherein the content tendency refers to age tendency, regional tendency and the like, and the release target estimation refers to understanding of a recommendation push target of the content through historical data behaviors of a user.
The data module cleans and processes the data in the message content library module by utilizing various data processing tools and falls into different systems, and the data in the message content library module is used in the system modules such as a feedback algorithm model, a user portrait and the like.
Specifically, the candidate sets generated by different recall strategies in the recall module are fused through the fusion filtering module, and unconditional content candidate sets are screened and filtered in the candidate set fusion process, so that the situation that the junk message is pushed to the user can be avoided, and the situation that the mood is influenced because the junk message is received by the user can be avoided.
The recall module adopts a multi-path recall architecture, the recall module refers to data in different systems of the data module in a plurality of categories such as model categories, attributes, manual intervention, heuristics, operation strategies and the like, and the recall module triggers the recall strategy and generates an initial candidate set through reference conditions such as user attributes, user portraits, historical behaviors, real-time behaviors, geographic positions and the like.
And the fusion filtering module fuses the candidate sets generated by different recall strategies in the recall module, and filters the content of the candidate set which is screened out by the recall module and does not meet the condition.
The sorting module reorders the screened candidate sets by using a machine learning model, the sorted candidate sets are divided into a coarse sorting module and a fine sorting module, the candidate sets in the coarse sorting module are generally ten thousands of pieces of data, the candidate sets in the fine sorting module are generally thousands of pieces of data, the fine sorting module needs to infer the click rate of candidate messages, multi-target reading time, sharing and other recommendation indexes relevant to the users according to the users, and the fine sorting module screens the data in the coarse sorting module candidate sets.
Specifically, the pushing module is used for pushing the message for the user, so that the APP applet does not need to be downloaded, and the message required by the user is obtained, so that on one hand, the mobile phone memory of the user can be saved, on the other hand, the possibility of personal information of the user is reduced, and meanwhile, the convenience of the user is improved on a certain basis.
A5G message pushing method comprises the following steps:
s1, firstly, texts, videos, graphics context, audios and the like created by a content producer in an enterprise data module on a platform are used as content sources and stored in a message content source module in a big data platform, and meanwhile, the click rate of a message, multi-target reading time length, sharing and other data contents of recommendation indexes which are closely related to the user are stored in the large data platform.
S2: and the message content source module in the big data platform normalizes the data stored in the content source of the big data platform and feeds the normalized data back to the application party, and then the message content library module classifies and sorts the data in the message content source to obtain a content library.
S3: according to the click rate of the user on the message, the multi-target reading time length, the sharing and other data of recommendation indexes which are closely related to the user are taken as references, the data module in the pushing module cleans and processes the data of the message content library module in the big data platform and falls into different systems, and then the recall module is utilized to refer to the data in different systems of the data module by referring to the model class, the attribute, the manual intervention, the heuristic class, the operation strategy and other large classes, so that the recall module triggers the recall strategy and generates an initial recommendation candidate set through the reference conditions of the attribute, the user portraits, the historical behaviors, the real-time behaviors, the geographic positions and the like of the user.
S4: when the recall strategy is triggered in the recall module and an initial recommended candidate set is generated, the fusion filtering module fuses the candidate sets generated by different recall strategies in the recall module, and screens and filters out the unconditional content candidate set in the candidate set fusion process.
S5: when the candidate sets are fused, the candidate sets enter a coarse ranking module in the ranking module, and at the moment, the fine ranking module needs to infer and screen data in the coarse ranking module candidate sets according to the click rate of the user on the message, the multi-target reading time length, the sharing and other recommendation indexes which are closely related to the user, so that the accurate pushing of the message to the user is ensured.
Working principle: firstly, a text, a video, an image text, an audio and the like created by a content producer in an enterprise data module on a platform are used as a content source and stored in a message content source module in a large data platform, meanwhile, the click rate of a message, multi-target reading time length, sharing and other data content of recommendation indexes related to the user are stored in the message content source module in the large data platform, the message content source module in the large data platform normalizes the data stored in the content source of the large data platform and feeds the normalized data back to an application party, then the data in the message content source is classified and arranged through a message content library module to obtain a content library, and the click rate of the message, the multi-target reading time length, the sharing and other data of the recommendation indexes related to the user are used as references according to the user, the data module in the pushing module cleans and processes the data of the message content library module in the big data platform and falls into different systems, then the recall module is utilized to refer to the data in the different systems of the data module by a plurality of major categories such as model category, attribute, manual intervention, heuristic category, operation strategy and the like, the recall module triggers the recall strategy and generates an initial recommendation candidate set through the reference conditions such as user attribute, user portrait, historical behavior, real-time behavior, geographic position and the like, when the recall strategy is triggered in the recall module and generates the initial recommendation candidate set, the fusion filtering module fuses the candidate sets generated by the different recall strategies in the recall module, screens and filters the non-conforming content candidate sets in the candidate set fusion process, the non-conforming content candidate sets enter the coarse ranking module in the ranking module after the candidate sets are fused, and the fine ranking module needs to refer to the click rate of the user on the message, the multi-target reading time length, sharing and other recommendation indexes closely related to the user are presumed, and data in the coarse-ranking module candidate set are screened, so that accurate pushing of the message to the user is ensured.
The foregoing has shown and described the basic principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the above-described embodiments, and that the above-described embodiments and descriptions are only preferred embodiments of the present invention, and are not intended to limit the invention, and that various changes and modifications may be made therein without departing from the spirit and scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (8)
1. A 5G message pushing system, characterized by: the 5G message pushing system comprises an enterprise data module, a big data platform, a pushing module and a user, wherein the output end of the enterprise data module is connected with the input end of the big data platform, the output end of the big data platform is connected with the input end of the pushing module, the output end of the pushing module is connected with the input end of the user, and the output end of the user is connected with the input end of the big data platform;
the big data platform comprises a message content source module and a message content library module, wherein the output end of the enterprise data module is connected with the input end of the message content source module, the output end of the message content source module is connected with the input end of the message content library module, and the output end of the message content library module is connected with the input end of the pushing module;
the pushing module comprises a data module, a recall module, a fusion filtering module and a sequencing module, wherein the output end of the data module is connected with the input end of the recall module, the output end of the recall module is connected with the input end of the fusion filtering module, the output end of the fusion filtering module is connected with the input end of the sequencing module, and the output end of the sequencing module is connected with the input end of a user.
2. A 5G message pushing system according to claim 1, wherein: the message content source module takes texts, videos, graphics context and audios authored by enterprise content producers on a platform as content sources, wherein the data in the content sources have larger difference, the data in the content sources are subjected to normalization processing, and the platform performs unified processing of the content, including a label system, a theme and categories and feeds the content back to an application party.
3. A 5G message pushing system according to claim 1, wherein: the message content library module is used for carrying out content understanding on a message content source and constructing a message content library, the message content library module is used for processing titles and data of a large number of text, audio and video contents in the message content source, and the message content library module is used for classifying and sorting the contents to obtain a content library.
4. A 5G message pushing system according to claim 1, wherein: the data module cleans and processes the data in the message content library module by utilizing various data processing tools and falls into different systems, and the data in the message content library module is used in a feedback algorithm model and a user portrait system module.
5. A 5G message pushing system according to claim 1, wherein: the recall module adopts a multi-path recall architecture, the recall module refers to data in different systems of the data module in a plurality of categories of model, attribute, manual intervention, heuristic and operation strategy, and the recall module triggers the recall strategy and generates an initial candidate set through user attribute, user portrait, historical behavior, real-time behavior and geographic position reference conditions.
6. A 5G message pushing system according to claim 1, wherein: and the fusion filtering module fuses the candidate sets generated by different recall strategies in the recall module, and filters the content of the candidate set which is screened out by the recall module and does not meet the condition.
7. A 5G message pushing system according to claim 1, wherein: the sorting module reorders the screened candidate sets by using a machine learning model, the sorted candidate sets are divided into a coarse sorting module and a fine sorting module, the candidate sets in the coarse sorting module are generally ten thousands of pieces of data, the candidate sets in the fine sorting module are generally thousands of pieces of data, the fine sorting module needs to infer the click rate of candidate messages, multi-target reading time, sharing and other recommendation indexes relevant to the users according to the users, and the fine sorting module screens the data in the coarse sorting module candidate sets.
8. A 5G message pushing method, according to any of claims 1-7, wherein:
s1: firstly, taking texts, videos, graphics context and audios created by a content producer in a platform in an enterprise data module as content sources and storing the texts, videos, graphics context and audios into a message content source module in a big data platform, and simultaneously storing the click rate of a message, multi-target reading time, sharing and other data contents of recommendation indexes related to the user, and storing the data contents into the message content source module in the big data platform;
s2: the message content source module in the big data platform normalizes the data stored in the content source of the big data platform and feeds the normalized data back to the application party, and then the message content library module classifies and sorts the data in the message content source to obtain a content library;
s3: according to the click rate of a user on a message, multi-target reading time length, sharing and other data of recommendation indexes which are closely related to the user as references, a data module in a pushing module cleans and processes the data of a message content library module in a big data platform and falls into different systems, and then a recall module is utilized to refer to data in different systems of the data module by referring to model classes, attributes, manual intervention, heuristic classes and operation strategies, so that the recall module triggers a recall strategy and generates an initial recommendation candidate set through the attribute, user portraits, historical behaviors, real-time behaviors and geographic position reference conditions of the user;
s4: when a recall strategy is triggered in the recall module and an initial recommended candidate set is generated, the fusion filtering module fuses the candidate sets generated by different recall strategies in the recall module, and filters out the unconditional content candidate sets in the candidate set fusion process;
s5: when the candidate sets are fused, the candidate sets enter a coarse ranking module in the ranking module, and at the moment, the fine ranking module needs to infer and screen data in the coarse ranking module candidate sets according to the click rate of the user on the message, the multi-target reading time length, the sharing and other recommendation indexes which are closely related to the user, so that the accurate pushing of the message to the user is ensured.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310329665.6A CN116028721A (en) | 2023-03-30 | 2023-03-30 | 5G message pushing system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310329665.6A CN116028721A (en) | 2023-03-30 | 2023-03-30 | 5G message pushing system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116028721A true CN116028721A (en) | 2023-04-28 |
Family
ID=86074494
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310329665.6A Pending CN116028721A (en) | 2023-03-30 | 2023-03-30 | 5G message pushing system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116028721A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112131411A (en) * | 2020-09-21 | 2020-12-25 | 腾讯科技(深圳)有限公司 | Multimedia resource recommendation method and device, electronic equipment and storage medium |
CN113095888A (en) * | 2021-03-12 | 2021-07-09 | 上海意略明数字科技股份有限公司 | Message pushing method and device, storage medium and computer equipment |
CN114168790A (en) * | 2021-09-27 | 2022-03-11 | 中南大学 | Personalized video recommendation method and system based on automatic feature combination |
CN114297434A (en) * | 2021-12-30 | 2022-04-08 | 镇江多游网络科技有限公司 | Short video information stream intelligent recommendation method based on GPU cluster |
CN114329176A (en) * | 2021-11-08 | 2022-04-12 | 腾讯科技(武汉)有限公司 | Information recommendation method and device, computer equipment, storage medium and program product |
-
2023
- 2023-03-30 CN CN202310329665.6A patent/CN116028721A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112131411A (en) * | 2020-09-21 | 2020-12-25 | 腾讯科技(深圳)有限公司 | Multimedia resource recommendation method and device, electronic equipment and storage medium |
CN113095888A (en) * | 2021-03-12 | 2021-07-09 | 上海意略明数字科技股份有限公司 | Message pushing method and device, storage medium and computer equipment |
CN114168790A (en) * | 2021-09-27 | 2022-03-11 | 中南大学 | Personalized video recommendation method and system based on automatic feature combination |
CN114329176A (en) * | 2021-11-08 | 2022-04-12 | 腾讯科技(武汉)有限公司 | Information recommendation method and device, computer equipment, storage medium and program product |
CN114297434A (en) * | 2021-12-30 | 2022-04-08 | 镇江多游网络科技有限公司 | Short video information stream intelligent recommendation method based on GPU cluster |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US9792356B2 (en) | System and method for supporting natural language queries and requests against a user's personal data cloud | |
CN1297930C (en) | Apparatus and method for authoring multimedia document | |
US8528018B2 (en) | System and method for evaluating visual worthiness of video data in a network environment | |
CN102473189B (en) | Providing in real time to the link of portion of media object during social networks upgrades | |
US20050246324A1 (en) | System and associated device, method, and computer program product for performing metadata-based searches | |
US10496604B2 (en) | System and method for providing and/or obtaining electronic documents | |
US20130091071A1 (en) | Product catalog dynamically tailored to user-selected media content | |
US20200028816A1 (en) | Automated classification and time-based relevancy prioritization of electronic mail items | |
CN101578601A (en) | Initial impression analysis tool for an online dating services | |
CN102984050A (en) | Method, client and system for searching voices in instant messaging | |
WO2023016536A1 (en) | Interaction method, apparatus and device, and storage medium | |
CN113574555A (en) | Intelligent summarization based on context analysis of auto-learning and user input | |
US11343114B2 (en) | Group management in a messaging service | |
US20220394068A1 (en) | Method for a video content distribution and messaging platform providing personalized video feeds | |
US20230394395A1 (en) | Conversation-based messaging method and system | |
US11290409B2 (en) | User device messaging application for interacting with a messaging service | |
US11645466B2 (en) | Categorizing conversations for a messaging service | |
CN116028721A (en) | 5G message pushing system | |
EP1698986A2 (en) | Creation and composition of sets items | |
CN107870938B (en) | Integrated system for picture interactive processing application interface | |
JP7478898B2 (en) | Messaging Services | |
US11349800B2 (en) | Integration of an email, service and a messaging service | |
JP2022524470A (en) | Systems and methods for improved search and categorization of media content items based on media content item destinations | |
CN111159435A (en) | Multimedia resource processing method, system, terminal and computer readable storage medium | |
CN107870713B (en) | Picture and text integrated picture processing method with compatibility |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20230428 |