WO2024174985A1

WO2024174985A1 - Advertisement copy generation method, network device and computer storage medium

Info

Publication number: WO2024174985A1
Application number: PCT/CN2024/077705
Authority: WO
Inventors: 姚小虎; 王豪; 丁洋凯; 朱金霞
Original assignee: 华为技术有限公司
Priority date: 2023-02-25
Filing date: 2024-02-20
Publication date: 2024-08-29
Also published as: CN118552248A

Abstract

Provided in the present application are an advertisement copy generation method, a network device and a computer storage medium. The method is applied to a network device, and comprises: acquiring first advertisement information of a first object, the first advertisement information comprising a first copy; acquiring a first advertisement request sent by a first device, and acquiring a first keyword and a second keyword from the first advertisement request; determining the first advertisement information according to the first keyword; determining a second copy according to the first keyword, the second keyword and the first copy; and sending to the first device first response information, the first response information comprising the second copy, and the first response information being used for instructing the first device to display the second copy. The present application can generate advertisement copies better satisfying the requirements and preferences of users, and recommend same to the users, thereby improving advertisement launching effects.

Description

Advertisement copy generation method, network device and computer storage medium

This application claims the priority of the Chinese patent application filed with the State Intellectual Property Office of China on February 25, 2023, with application number 202310208054.6, and the priority of the Chinese patent application with the invention name "Advertising copy generation method, network device and computer storage medium", all contents of which are incorporated by reference in this application.

Technical Field

The present application relates to the field of computer technology, and in particular to an advertising copy generation method, a network device and a computer storage medium.

Background Art

Currently, advertisers can upload the advertising copy of their products to the advertising platform so that the advertising platform can place advertisements for the products. When a mobile phone or other terminal sends an advertisement request to the advertising platform, the advertising platform can return the advertising copy uploaded by the advertiser, and the terminal can display the advertising copy to the user. However, these advertising copies are generated in advance and may not meet the current needs and preferences of users, resulting in poor advertising results.

Summary of the invention

The present application discloses an advertising copy generation method, a network device and a computer storage medium, which can generate online advertising copy that better meets the needs and preferences of users and recommend it to users, thereby improving the effect of advertising delivery.

In a first aspect, the present application provides an advertising copy generation method, which is applied to a network device, and the method includes: obtaining first advertising information of a first object, the first advertising information including a first copy; obtaining a first advertising request sent by a first device, and obtaining a first keyword and a second keyword from the first advertising request; determining the first advertising information based on the first keyword; determining a second copy based on the first keyword, the second keyword and the first copy; sending a first response message to the first device, the first response message including the second copy, and the first response message being used to instruct the first device to display the second copy.

In some examples, the first object is a physical product or a non-physical service.

In some examples, the first copy includes at least one of the following: an advertisement name, theme, content summary, description, and advertisement copy input by an advertiser of the first object, and text obtained by crawling a link input by the advertiser of the first object by the network device.

In some examples, the first keyword and the second keyword are acquired by the first device according to a received search request.

In some examples, the first keyword is the name of the object category to which the first object belongs, and the second keyword is an adjective for the first keyword. For example, the first object is "Brand A mobile phone", the first keyword is "mobile phone", and the second keyword is "high performance".

In the above method, the network device can generate the second copy online based on the keyword searched online by the user and the first copy in the advertisement information, and return the second copy to the first device for display, instead of directly returning the first copy generated in advance to the first device for display. Compared with the first copy, the second copy generated online can well meet the current needs of online users (reflected by keywords), so it can improve the efficiency of advertising delivery and also improve the user experience.

In one embodiment, the above method also includes: obtaining a second advertising request sent by the first device, obtaining the first keyword and the third keyword from the second advertising request; determining the first advertising information based on the first keyword; determining the third copy based on the first keyword, the third keyword and the first copy; sending a second response message to the first device, the second response message including the third copy, and the second response message being used to instruct the first device to display the third copy.

In the above method, the keywords searched by the user of the first device at different times/scenarios may be different, and the network device will generate advertising copy based on the keywords searched by the user online. Therefore, when the network device receives different keywords, the copy generated for the same object may be different. This can meet the different needs of the same user in different search scenarios and effectively improve the efficiency of advertising delivery.

In one implementation, the first advertisement information includes at least one of the following: a first picture, a first video, and a first link of the first object, and the first response information also includes at least one of the following: the first picture, the first video, and the first link.

In the above method, the first advertising information of the first object includes not only the first copy but also pictures, videos, links and other information. The network device can send the second copy generated online and non-text content such as pictures, videos, links in the first advertising information to the first device instead of generating non-text content online, thereby reducing processing time and further improving the efficiency of advertising delivery.

In one implementation, the second text includes a first sentence, the first text includes the first sentence, and the first sentence is associated with the second keyword.

In the above method, the network device can directly extract the first sentence associated with the second keyword in the first copy to form the second copy, which can not only make the obtained second copy meet the current needs of the user, but also reduce the time for generating copy online, and further improve the efficiency of advertising delivery.

In one implementation, the second copy includes a second sentence, the first copy does not include the second sentence, and the second sentence is associated with the second keyword.

In the above method, the network device may not be limited to the content in the first copy, and may generate a second sentence associated with the second keyword to form a second copy, so that the obtained second copy can well meet the current needs of the user and improve the user experience.

In one embodiment, the first copy includes M sentences, and determining the second copy based on the first keyword, the second keyword and the first copy includes: selecting N sentences ranked in the top N positions in terms of relevance to the first keyword and the second keyword from the M sentences; determining the second copy based on the N sentences, where N is a positive integer less than M, and M is a positive integer greater than 1.

In one embodiment, the first advertising request includes an identifier of a first user; the method further includes: obtaining characteristic information and/or historical behavior data of the first user based on the identifier of the first user; determining the second copy based on the first keyword, the second keyword and the first copy includes: determining the second copy based on the first keyword, the second keyword, the first copy, and characteristic information and/or historical behavior data of the first user.

In the above method, the network device can not only determine the second copy based on the keywords searched online by the first user, but also determine the second copy based on the characteristic information and/or historical behavior data of the first user. The second copy obtained in this way can not only meet the current needs of the first user, but also meet the first user's preference for the copy type/style (which can be reflected through characteristic information and historical behavior data), so that the first user will be more interested in the second copy, thereby further improving the efficiency of advertising delivery.

In one implementation, before obtaining the first advertisement request sent by the first device, the method further includes: obtaining feature information and/or historical behavior data related to the identifier of the first user.

In some examples, obtaining characteristic information and/or historical behavior data related to the identifier of the first user includes: receiving and storing characteristic information and/or historical behavior data related to the identifier of the first user sent by the first device.

In the above method, the network device can receive and store the characteristic information and/or historical behavior data of the first user in advance. After subsequently obtaining the first advertising request sent by the first device, the characteristic information and/or historical behavior data of the first user can be directly obtained from the memory instead of obtaining it from the first device, thereby reducing data interaction when generating copy online, reducing processing time, and further improving advertising delivery efficiency.

In one implementation, the historical behavior data of the first user includes a fourth copy of a second object clicked by the first user, the fourth copy includes a third sentence, and the second copy includes the third sentence.

In the above method, the network device can directly extract the third sentence in the fourth copy clicked by the first user to form the second copy, which can not only make the obtained second copy meet the user's preference for copy style, but also reduce the time for online copy generation, thereby further improving the efficiency of advertising delivery.

In one embodiment, the characteristic information of the first user includes at least one of the following: age, gender, location, type of electronic device, and user portrait; the historical behavior data of the first user includes at least one of the following: clicking on advertisements, browsing advertisements, closing advertisements, purchasing corresponding objects through links in advertisements, downloading applications through links in advertisements, and duration of browsing advertisements.

In the above method, the historical behavior data and feature information that can be used to generate the second copy online are rich and varied, and are all data that can reflect the user's preference for the type/style of advertising copy. Therefore, the second copy generated based on these data can well meet the needs of users and has a high advertising delivery efficiency.

In one embodiment, the first advertising information includes first indication information, and the first indication information indicates that the first function is in an on state or a off state; determining the first advertising information based on the first keyword includes: judging whether the first indication information indicates that the first function is in an on state; when the first indication information indicates that the first function is in an on state, determining the first advertising information based on the first keyword; the method also includes: when the first indication information indicates that the first function is in an off state, sending a third response information to the first device, the third response information including the first text, and the third response information being used to instruct the first device to display the first text.

In some examples, the advertiser of the first object may set the first function to be turned on or off through the advertisement configuration interface provided by the network device. The first function is, for example, called an online personalized copywriting function.

In the above method, the advertiser can choose whether to turn on the first function. When the first function is turned on, the network device will generate the second copy online based on the keywords searched by the user online and return it to the first device. When the first function is turned off, the network device will directly return the first copy in the first advertising information to the first device. The advertiser can flexibly control the advertising copy delivered to enhance the advertiser's experience.

In a second aspect, the present application provides another method for generating advertising copy, which is applied to a network device, and the method includes: obtaining first advertising information of a first object, the first advertising information including a first copy; obtaining a first advertising request sent by a first device, the first advertising request including an identifier of a first user; obtaining characteristic information and/or historical behavior data of the first user based on the identifier of the first user; determining a second copy based on the first copy, the characteristic information and/or historical behavior data of the first user; and sending a first response message to the first device, the first response message including the second copy, and the first response message being used to instruct the first device to display the second copy.

Among them, the examples of the above-mentioned first object and the above-mentioned first copy can refer to the examples of the first object and the first copy in the first aspect, and will not be repeated here.

In the above method, the network device can generate the second copy online based on the online user's feature information and/or historical behavior data, and the first copy in the advertisement information, and return the second copy to the first device for display, rather than directly returning the first copy generated in advance to the first device for display. Compared with the first copy, the second copy generated online can well reflect the online user's preference for the copy type/style and meet the user's needs, thereby improving the efficiency of advertising delivery and the user experience.

In one embodiment, the above method also includes: obtaining a second advertising request sent by a second device, the second advertising request including an identifier of a second user; obtaining characteristic information and/or historical behavior data of the second user based on the identifier of the second user; determining a third copy based on the first copy and the characteristic information and/or historical behavior data of the second user; and sending a second response message to the second device, the second response message including the third copy, and the third response message being used to instruct the second device to display the third copy.

In the above method, after the network device receives advertising requests from different devices, it can generate advertising copy online based on the characteristic information and/or historical behavior data of the corresponding user. Therefore, for advertising requests from different devices/different users, the copy generated by the network device for the same object can be different, which can meet the different needs of different users and effectively improve the efficiency of advertising delivery.

In one implementation, the historical behavior data of the first user includes a fourth copy of a second object clicked by the first user, the fourth copy includes a first sentence, and the second copy includes the first sentence.

In the above method, the network device can directly extract the first sentence in the fourth copy clicked by the first user to form the second copy, which can not only make the obtained second copy meet the user's preference for copy style, but also reduce the time for online copy generation, thereby further improving the efficiency of advertising delivery.

In one embodiment, the first advertising information includes first indication information, and the first indication information indicates that the first function is in an on state or a off state; the obtaining of the characteristic information and/or historical behavior data of the first user based on the identifier of the first user includes: judging whether the first indication information indicates that the first function is in an on state; when the first indication information indicates that the first function is in an on state, obtaining the characteristic information and/or historical behavior data of the first user based on the identifier of the first user; the method also includes: when the first indication information indicates that the first function is in an off state, sending a third response information to the first device, the third response information including the first text, and the third response information being used to instruct the first device to display the first text.

In some examples, the advertiser of the first object can set the first function through the advertisement configuration interface provided by the network device. Turn on or off, the first function is called, for example, an online personalized copywriting function.

In the above method, the advertiser can choose whether to turn on the first function. When the first function is turned on, the network device will generate the second copy based on the user's characteristic information and/or historical behavior data line and return it to the first device. When the first function is turned off, the network device will directly return the first copy in the first advertising information to the first device. The advertiser can flexibly control the advertising copy delivered to enhance the advertiser's experience.

In a third aspect, the present application provides another method for generating advertising copy, which is applied to an electronic device, the method comprising: obtaining a first search request, and obtaining a first keyword and a second keyword based on the first search request; sending a first advertisement request to a network device, the first advertisement request including the first keyword and the second keyword; receiving a first response information sent by the network device, the first response information including a first copy; displaying the first copy; wherein the first copy is determined by the network device based on the first keyword, the second keyword and the second copy, and the second copy is the copy in the first advertisement information of the first object determined by the network device based on the first keyword.

In one embodiment, the method also includes: obtaining a second search request, and obtaining the first keyword and the third keyword based on the second search request; sending a second advertising request to the network device, the second advertising request including the first keyword and the third keyword; receiving a second response message sent by the network device, the second response message including a third copy; displaying the third copy; wherein the third copy is determined by the network device based on the first keyword, the third keyword and the second copy.

In one embodiment, the first advertising information includes at least one of the following: a first picture, a first video, and a first link of the first object, and the first response information includes at least one of the following: the first picture, the first video, and the first link; and displaying the first text includes: displaying the first text, and displaying at least one of the first picture, the first video, and the first link.

In one implementation, the first copy includes a first sentence, the second copy includes the first sentence, and the first sentence is associated with the second keyword.

In one implementation, the first text includes a second sentence, the second text does not include the second sentence, and the second sentence is associated with the second keyword.

Among them, the beneficial effects of the third aspect and the implementation methods of the third aspect are similar to the beneficial effects of the first aspect and the implementation methods of the first aspect, and will not be repeated here.

In a fourth aspect, the present application provides another method for generating advertising copy, which is applied to an electronic device, the method comprising: sending a first advertising request to a network device, the first advertising request including an identifier of a first user, the identifier of the first user being used by the network device to obtain characteristic information and/or historical behavior data of the first user; receiving a first response message sent by the network device, the first response message including a first copy; displaying the first copy; wherein the first copy is determined by the network device based on the characteristic information and/or historical behavior data of the first user and a second copy, and the second copy is the copy in the first advertising information of the first object determined by the network device.

In one embodiment, the first advertising information includes at least one of the following: a first picture, a first video, and a first link of the first object, and the first response information includes at least one of the following: the first picture, the first video, and the first link; and the displaying of the first text includes: displaying the first text, and displaying at least one of the first picture, the first video, and the first link.

In one embodiment, before sending the first advertisement request to the network device, the method further includes: sending feature information and/or historical behavior data related to the identifier of the first user to the network device.

In one implementation, the historical behavior data of the first user includes a fourth text of a second object clicked by the first user, and the second text includes a first sentence in the fourth text.

In one embodiment, the characteristic information of the first user includes at least one of the following: age, gender, location, type of electronic device used, and user portrait; the historical behavior data of the first user includes at least one of the following: clicking on advertisements, browsing advertisements, closing advertisements, purchasing corresponding objects through links in advertisements, downloading applications through links in advertisements, duration of browsing advertisements, and types of advertisements operated or browsed.

Among them, the beneficial effects of the fourth aspect and the implementation methods of the fourth aspect are similar to the beneficial effects of the second aspect and the implementation methods of the second aspect, and will not be repeated here.

In a fifth aspect, the present application provides a network device, including a transceiver, a processor and a memory; the above-mentioned memory is used to store computer program code, the above-mentioned computer program code includes computer instructions, and the above-mentioned processor calls the above-mentioned computer instructions to enable the above-mentioned network device to execute the advertising copy generation method provided by the first aspect, the second aspect and any one of the implementation methods of the first aspect and the second aspect of the present application.

In the sixth aspect, the present application provides an electronic device, including a transceiver, a processor and a memory; the above-mentioned memory is used to store computer program code, the above-mentioned computer program code includes computer instructions, and the above-mentioned processor calls the above-mentioned computer instructions to enable the above-mentioned electronic device to execute the advertising copy generation method provided by the third aspect, the fourth aspect and any one of the implementation methods of the third aspect and the fourth aspect of the present application.

In the seventh aspect, the present application provides a computer storage medium, which stores a computer program, and the computer program includes program instructions. When the program instructions are executed by a processor, they are used to execute the advertising copy generation method provided by the first to fourth aspects of the present application, and any one of the implementation methods of the first to fourth aspects.

In an eighth aspect, the present application provides a computer program product. When the computer program product runs on an electronic device, the electronic device executes the advertising copy generation method provided by any one of the first to fourth aspects of the present application and any implementation method of the first to fourth aspects.

In a ninth aspect, the present application provides an electronic device, the electronic device comprising a method or device for executing any aspect or implementation of the present application. The electronic device is, for example, a chip.

It should be understood that the description of technical features, technical solutions, beneficial effects or similar language in this application does not imply that all features and advantages can be realized in any single embodiment. On the contrary, it is understood that the description of features or beneficial effects means that specific technical features, technical solutions or beneficial effects are included in at least one embodiment. Therefore, the description of technical features, technical solutions or beneficial effects in this application does not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions and beneficial effects described in this application can also be combined in any appropriate manner. Those skilled in the art will understand that this application can be implemented without one or more specific technical features, technical solutions or beneficial effects of a specific embodiment. In other embodiments, additional technical features and beneficial effects can also be identified in specific embodiments that do not embody all embodiments.

BRIEF DESCRIPTION OF THE DRAWINGS

The following is an introduction to the drawings used in this application.

FIG1A is a schematic diagram of the architecture of an advertising copy generation system provided by the present application;

FIG1B is a schematic diagram of the architecture of another advertising copy generation system provided by the present application;

FIG2A is a schematic diagram of the hardware structure of an electronic device provided by the present application;

FIG2B is a schematic diagram of a software architecture of an electronic device provided by the present application;

FIG2C is a schematic diagram of the hardware structure of a network device provided by the present application;

FIG3 is a schematic diagram of a user interface provided by the present application;

FIG4 is a schematic diagram of another user interface provided by the present application;

FIG5 is a flow chart of a method for generating advertisement copy provided by the present application;

FIG6 is a schematic diagram of another user interface provided by the present application;

FIG7 is a schematic diagram of another user interface provided by the present application;

FIG8 is a flow chart of another method for generating advertisement copy provided by the present application;

FIG9 is a flow chart of a model training process provided by the present application;

FIG10 is a schematic diagram of another user interface provided by the present application;

FIG11 is a flow chart of another method for generating advertisement copy provided by the present application;

FIG12A is a schematic diagram of an advertisement configuration interface provided by the present application;

FIG12B is a schematic diagram of another advertisement configuration interface provided by the present application;

FIG13 is a flow chart of an advertisement delivery method provided in the present application.

DETAILED DESCRIPTION

The technical solution in the embodiment of the present application will be described below in conjunction with the accompanying drawings. In the description of the embodiment of the present application, unless otherwise specified, "/" means or, for example, A/B can mean A or B; "and/or" in the text is only a description of the association relationship of associated objects, indicating that there can be three relationships, for example, A and/or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiment of the present application, "multiple" means two or more than two.

In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

The current advertising copywriting generation technology is to generate copywriting offline. For example, the advertising platform generates copywriting from the purchase words for the advertisement provided by the advertiser. The text content related to the keywords is extracted from the information such as the title (such as car), title (such as X brand car, pick up the car immediately, get the insurance price of all models), description (such as X brand car, breakthrough technology, enlighten the future, enjoy a wonderful travel life) and landing page, and the copy of the current advertisement is generated in advance/offline based on these text contents, and then the copy is delivered to the corresponding platform. When any user is recommended this advertisement later, the mobile phone or other terminal used by the user will display the above offline generated copy. However, these copies may not meet the needs and preferences of users. For example, in the search advertising scenario, users will enter the search keywords, and when the terminal displays the search results related to the keywords, it often recommends some related product advertisements. At this time, the advertising copy generated offline based on the above advertising copy generation technology can be displayed. However, when the copy is generated offline, the real-time needs of online users cannot be perceived in advance. For example, the user needs represented by the keywords searched by users cannot be perceived, so the current interests of users cannot be reflected in the advertising copy. For example, when users browse different ads, they often choose to click on some ads that they are interested in. However, when generating ads offline, it is impossible to determine the specific users and their historical behavior data in advance, so it is impossible to generate ads based on the user's preferences for different ad types represented by historical behavior data. This will lead to poor advertising results, such as a low click-through rate (CTR), where CTR is the actual number of clicks on an ad divided by the number of ad impressions.

The present application provides a method for generating advertising copy, which can be applied to advertising delivery/display scenarios. The network equipment of the advertising platform can use online information to generate online copy. For example, in a search advertising scenario, the advertising copy can be generated online and in real time based on the keywords searched by the first user and the historical behavior data of the first user, and the advertising copy is sent to the electronic device for display. On the one hand, the advertising copy is generated based on the keywords searched by the user online, so it can match the user's current interests. On the other hand, the advertising copy is generated based on the user's historical behavior data, so it can match the user's preference for the type of copy. The advertising copy generated by this application is more in line with the needs and preferences of users, effectively improving the efficiency and effectiveness of advertising delivery, and also improving the user experience.

Any advertisement in this application is used to promote the corresponding object. The object in this application can be a physical commodity/product or a non-physical service, etc. This application does not limit this.

This application does not limit the software for displaying advertisements on electronic devices, such as but not limited to applications (APPs), mini-programs, atomic services, or web pages. For the convenience of explanation, the following description will take applications as an example.

Please refer to FIG. 1A , which exemplarily shows a schematic diagram of the architecture of an advertisement copy generation system 10 .

As shown in FIG. 1A , the advertisement copy generation system 10 may include an electronic device 100 , an electronic device 200 , a network device 300 , an electronic device 400 , and an electronic device 500 . Among them:

The electronic device 100/electronic device 200 can establish a connection and communicate with the network device 300 via wired and/or wireless means. Wired means include, but are not limited to, high definition multimedia interface (HDMI), universal serial bus (USB), network cable (such as twisted pair, coaxial cable or optical fiber), etc., and wireless means include, but are not limited to, wireless local area networks (WLAN) (such as wireless fidelity (Wi-Fi)), satellite, cellular communication, etc.

The network device 300 can be used to maintain the advertising content of multiple objects, such as but not limited to the advertising form, the text, links, pictures, videos and sounds that appear in the advertisement. Multiple advertisers can upload the advertising content to be delivered to the network device 300. For example, the advertiser 1 of advertisement 1 can input the configuration content of delivery of advertisement 1 in the electronic device 400. For specific examples, please refer to Figures 12A and 12B below. The electronic device 400 can send the configuration content of delivery of advertisement 1 to the network device 300, and the network device 300 can implement the delivery of advertisement 1 based on the configuration content. The advertiser 2 of advertisement 2 can also send the configuration content of delivery of advertisement 2 to the network device 300 through the electronic device 500, and the network device 300 can implement the delivery of advertisement 2 based on the configuration content.

The electronic device 100 may be installed with one or more APPs. When the user 1 uses the APP in the electronic device 100, the electronic device 100 may obtain advertisement copy (such as the copy of the advertisement 1 or advertisement 2 mentioned above) from the network device 300, and display the obtained advertisement copy in the user interface of the APP. The user 2 may also browse the advertisement copy through the APP in the electronic device 200.

In one implementation, in the first advertising delivery scenario, the electronic device 100 may record the user's operation behavior and/or browsing behavior on the advertisement and information related to the advertisement, and the recorded data may be referred to as historical behavior data. When the electronic device 100 subsequently obtains the advertisement copy, the advertisement request sent to the network device 300 may carry the above historical behavior data. After receiving the advertisement request, the network device 300 may generate the advertisement copy based on the keyword and return the advertisement copy to the electronic device 100 for display.

In another embodiment, in the first search advertisement scenario, after the electronic device 100 receives the keyword input by the user, it can not only obtain the search results related to the keyword, but also send an advertisement request to the network device 300 to obtain the advertisement copy, and the advertisement request can carry the above-mentioned keyword. After receiving the advertisement request, the network device 300 can generate the advertisement copy based on the keyword and return the advertisement copy to the electronic device 100. When the electronic device 100 displays the search results related to the keyword, it can also display the advertisement copy.

In another embodiment, in the second search advertisement scenario, after receiving the keyword input by the user, the electronic device 100 may send an advertisement request to the network device 300 to obtain the advertisement copy. The advertisement request may carry the above-mentioned keyword and the After receiving the advertisement request, the network device 300 may generate an advertisement text based on the keyword and the historical behavior data, and return the advertisement text to the electronic device 100. When the electronic device 100 displays search results related to the keyword, it may also display the advertisement text.

Not limited to the architecture shown in FIG1A , in another embodiment, the number of electronic devices used by users may be more or less, the number of advertisers may be more or less, and the number of electronic devices used by advertisers may be more or less. For example, the electronic device used by any advertiser may be network device 300. Network device 300 may provide an advertisement configuration interface for advertisers to upload advertisement configuration content. Examples of the advertisement configuration interface may be found in FIGS. 12A and 12B below. This application does not limit the form and quantity of devices in the system.

Please refer to Fig. 1B, which exemplarily shows a schematic diagram of the architecture of another advertising copy generation system 10. Fig. 1B takes the above second search advertising scenario as an example for explanation.

As shown in FIG1B , the advertisement copy generation system 10 may include an electronic device 100 (user is user 1), a network device 300 and an electronic device 400 (user is advertiser 1). The electronic device 100 may include an input module 101, an advertisement acquisition module 102 and an output module 103. The network device 300 may include an advertisement material acquisition service 301, an advertisement copy generation service 302, a short-term interest extraction service 303 and a long-short interest fusion service 304. Among them:

The electronic device 400 can be used to receive the input of the advertiser 1, and obtain the configuration content of the advertisement 1 based on the input, and then send the configuration content to the advertising material acquisition service 301 of the network device 300. After receiving the configuration content of the advertisement 1, the advertising material acquisition service 301 can obtain the information of the advertisement 1 based on the configuration content, wherein the information of the advertisement (which can be referred to as the advertisement information) includes, for example but not limited to: the object of the advertisement (such as goods or services), the advertisement name, theme, content summary, description (such as the function description of the product advertisement or the service content description of the service advertisement), advertisement copy and other texts entered by the advertiser, non-text content such as pictures, videos, links (also called landing pages) uploaded by the advertiser, and text, pictures, videos, links and other information obtained after crawling the advertisement landing page. The advertisement information obtained by the advertising material acquisition service 301 can be provided to the advertisement copy generation service 302. The above description can be understood as advertiser 1 turning on the advertisement 1 delivery service. After advertiser 1 turns on the advertisement 1 delivery service, network device 300 can send the content of advertisement 1 (including advertising copy, and optionally non-text content such as pictures, videos, links, etc.) to any electronic device for display.

Next, the process of the network device 300 sending the content of the advertisement 1 to the electronic device 100 for display is exemplified.

The input module 101 of the electronic device 100 can be used to receive input from the user 1. After receiving the keyword input by the user 1, the input module 101 can send the keyword to the advertisement acquisition module 102. The advertisement acquisition module 102 can be used to send an advertisement request to the network device 300 to obtain an advertisement copy. The advertisement request can carry the keyword input by the user 1 and the identity document (ID) of the user 1.

After receiving the advertisement request, the advertisement copy generation service 302 of the network device 300 can determine the advertisement (assuming it is advertisement 1) of the object (such as a product or service) to be currently delivered, and obtain the information of advertisement 1 from the advertisement material acquisition service 301. Then, the advertisement copy generation service 302 can send keywords and the information of advertisement 1 to the short-term interest extraction service 303, and can send the ID of user 1 and the information of advertisement 1 to the long-term and short-term interest fusion service 304.

After the short-term interest extraction service 303 of the network device 300 receives the keywords searched by user 1 and the information of advertisement 1, it can extract text descriptions/copy fragments (which may be called short-term interests) related to the keywords from the information of advertisement 1 (including text descriptions) in real time and online, and send the short-term interests to the long- and short-term interest fusion service 304 for the generation of advertising copy.

After receiving the ID of user 1, the long-short interest fusion service 304 of the network device 300 can obtain the static features (also referred to as feature information) and historical behavior data of user 1 according to the ID of user 1. The static features include, but are not limited to, age, gender, city, device type of electronic device 100 used, user portrait, etc. The static features and historical behavior data of user 1 can be sent by the electronic device 100 to the network device 300 before receiving the input of user 1. The long-short interest fusion service 304 can generate the copy of advertisement 1 in real time online based on the acquired static features of user 1, the acquired historical behavior data of user 1, the short-term interests sent by the short-term interest extraction service 303, and the information of advertisement 1 sent by the advertisement copy generation service 302, and send the copy of advertisement 1 to the advertisement copy generation service 302.

After receiving the copy of Advertisement 1, the advertisement copy generation service 302 of the network device 300 may send the copy of Advertisement 1 to the advertisement acquisition module 102 of the electronic device 100, and the advertisement acquisition module 102 then sends the copy of Advertisement 1 to the output module 103 for display.

Not limited to the architecture shown in FIG. 1B , in another embodiment, the network device 300 may also not include the long-short interest fusion service 304. For example, in the first search advertising scenario described in FIG. 1A , the short-term interest extraction service 303 may send the short-term interest to the advertising copy generation service 302. The advertising copy generation service 302 may generate the copy of advertisement 1 based on the short-term interest, and return the copy to the electronic device 100 for display. In another embodiment, the network device 300 may also not include the short-term interest extraction service 303. For example, in the first advertising delivery scenario described in FIG. 1A , user 1 may not input keywords. The long-short interest fusion service 304 may generate the copy of advertisement 1 based on the static characteristics of user 1, the historical behavior data of user 1, and the information of advertisement 1. The generation of the advertising copy does not involve the keywords input by the user and the short-term interests obtained based on the keywords. The device in the advertising copy generation system 10 may include more or fewer modules. For example, a module may include The present application does not limit this, and the present application does not limit this, but includes multiple sub-modules. For example, multiple modules can be coupled together.

The electronic device in the present application can be a mobile phone, a tablet computer, a handheld computer, a desktop computer, a laptop computer, an ultra-mobile personal computer (UMPC), a netbook, a cellular phone, a personal digital assistant (PDA), as well as smart home devices such as smart TVs and smart cameras, wearable devices such as smart bracelets, smart watches, and smart glasses, extended reality (XR) devices such as augmented reality (AR), virtual reality (VR), and mixed reality (MR), vehicle-mounted devices or smart city devices. The embodiments of the present application do not impose any special restrictions on the specific types of electronic devices.

Next, the electronic device in this application is introduced as an example.

FIG. 2A exemplarily shows a schematic diagram of the hardware structure of an electronic device 100 .

As shown in Figure 2A, the electronic device 100 may include a processor 110, an external memory interface 120, an internal memory 121, a universal serial bus (USB) interface 130, a charging management module 140, a power management module 141, a battery 142, an antenna 1, an antenna 2, a mobile communication module 150, a wireless communication module 160, an audio module 170, a speaker 170A, a receiver 170B, a microphone 170C, an earphone interface 170D, a sensor module 180, a button 190, a motor 191, an indicator 192, a camera 193, a display screen 194, and a subscriber identification module (SIM) card interface 195, etc. The sensor module 180 may include a pressure sensor 180A, a gyroscope sensor 180B, an air pressure sensor 180C, a magnetic sensor 180D, an acceleration sensor 180E, a distance sensor 180F, a proximity light sensor 180G, a fingerprint sensor 180H, a temperature sensor 180J, a touch sensor 180K, an ambient light sensor 180L, a bone conduction sensor 180M, etc.

It is to be understood that the structure illustrated in the embodiment of the present application does not constitute a specific limitation on the electronic device 100. In other embodiments of the present application, the electronic device 100 may include more or fewer components than shown in the figure, or combine some components, or split some components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

The processor 110 may include one or more processing units, for example, the processor 110 may include an application processor (AP), a modem processor, a graphics processor (GPU), an image signal processor (ISP), a controller, a video codec, a digital signal processor (DSP), a baseband processor, and/or a neural-network processing unit (NPU), etc. Different processing units may be independent devices or integrated in one or more processors.

The controller can generate operation control signals according to the instruction operation code and timing signal to complete the control of instruction fetching and execution.

The processor 110 may also be provided with a memory for storing instructions and data. In one embodiment, the memory in the processor 110 is a cache memory. The memory may store instructions or data that the processor 110 has just used or cyclically used. If the processor 110 needs to use the instruction or data again, it may be directly called from the memory. This avoids repeated access, reduces the waiting time of the processor 110, and thus improves the efficiency of the system.

In one embodiment, the processor 110 may include one or more interfaces. The interface may include an inter-integrated circuit (I2C) interface, an inter-integrated circuit sound (I2S) interface, a pulse code modulation (PCM) interface, a universal asynchronous receiver/transmitter (UART) interface, a mobile industry processor interface (MIPI), a general-purpose input/output (GPIO) interface, a subscriber identity module (SIM) interface, and/or a universal serial bus (USB) interface, etc.

The charging management module 140 is used to receive charging input from a charger. The charger may be a wireless charger or a wired charger. In some wired charging implementations, the charging management module 140 may receive charging input from a wired charger through the USB interface 130. In some wireless charging implementations, the charging management module 140 may receive wireless charging input through a wireless charging coil of the electronic device 100. While the charging management module 140 is charging the battery 142, it may also power the electronic device 100 through the power management module 141.

The power management module 141 is used to connect the battery 142, the charging management module 140 and the processor 110. The power management module 141 receives input from the battery 142 and/or the charging management module 140 to power the processor 110, the internal memory 121, the display screen 194, the camera 193, and the wireless communication module 160.

The wireless communication function of the electronic device 100 can be implemented through the antenna 1, the antenna 2, the mobile communication module 150, the wireless communication module 160, the modem processor and the baseband processor.

Antenna 1 and antenna 2 are used to transmit and receive electromagnetic wave signals. Each antenna in electronic device 100 can be used to cover a single or multiple communication frequency bands. Different antennas can also be reused to improve the utilization of antennas. For example, antenna 1 can be reused as a diversity antenna for a wireless local area network. In another embodiment, the antenna can be used in combination with a tuning switch.

The mobile communication module 150 can provide solutions for wireless communications including 2G/3G/4G/5G/6G applied to the electronic device 100. The mobile communication module 150 may include at least one filter, a switch, a power amplifier, a low noise amplifier (LNA), etc. The mobile communication module 150 can receive electromagnetic waves from the antenna 1, and filter, amplify, and process the received electromagnetic waves, and transmit them to the modulation and demodulation processor for demodulation. The mobile communication module 150 can also amplify the signal modulated by the modulation and demodulation processor, and convert it into electromagnetic waves for radiation through the antenna 1. In one embodiment, at least some of the functional modules of the mobile communication module 150 can be set in the processor 110. In one embodiment, at least some of the functional modules of the mobile communication module 150 can be set in the same device as at least some of the modules of the processor 110.

The modem processor may include a modulator and a demodulator. Among them, the modulator is used to modulate the low-frequency baseband signal to be sent into a medium-high frequency signal. The demodulator is used to demodulate the received electromagnetic wave signal into a low-frequency baseband signal. The demodulator then transmits the demodulated low-frequency baseband signal to the baseband processor for processing. After the low-frequency baseband signal is processed by the baseband processor, it is passed to the application processor. The application processor outputs a sound signal through an audio device (not limited to a speaker 170A, a receiver 170B, etc.), or displays an image or video through a display screen 194. In one embodiment, the modem processor may be an independent device. In another embodiment, the modem processor may be independent of the processor 110 and be set in the same device as the mobile communication module 150 or other functional modules.

The wireless communication module 160 can provide wireless communication solutions including wireless local area networks (WLAN) (such as wireless fidelity (Wi-Fi) network), bluetooth (BT), global navigation satellite system (GNSS), frequency modulation (FM), near field communication (NFC), infrared (IR) and the like applied to the electronic device 100. The wireless communication module 160 can be one or more devices integrating at least one communication processing module. The wireless communication module 160 receives electromagnetic waves via the antenna 2, modulates the frequency of the electromagnetic wave signal and performs filtering processing, and sends the processed signal to the processor 110. The wireless communication module 160 can also receive the signal to be sent from the processor 110, modulate the frequency of the signal, amplify the signal, and convert it into electromagnetic waves for radiation through the antenna 2.

In one embodiment, the antenna 1 of the electronic device 100 is coupled to the mobile communication module 150, and the antenna 2 is coupled to the wireless communication module 160, so that the electronic device 100 can communicate with the network and other devices through wireless communication technology. The wireless communication technology may include global system for mobile communications (GSM), general packet radio service (GPRS), code division multiple access (CDMA), wideband code division multiple access (WCDMA), time-division code division multiple access (TD-SCDMA), long term evolution (LTE), BT, GNSS, WLAN, NFC, FM, and/or IR technology. The GNSS may include a global positioning system (GPS), a global navigation satellite system (GLONASS), a Beidou navigation satellite system (BDS), a quasi-zenith satellite system (QZSS) and/or a satellite based augmentation system (SBAS).

The electronic device 100 implements the display function through a GPU, a display screen 194, and an application processor. The GPU is a microprocessor for image processing, which connects the display screen 194 and the application processor. The GPU is used to perform mathematical and geometric calculations for graphics rendering. The processor 110 may include one or more GPUs that execute program instructions to generate or change display information.

The display screen 194 is used to display images, videos, etc. The display screen 194 includes a display panel. The display panel can be a liquid crystal display (LCD), an organic light-emitting diode (OLED), an active-matrix organic light-emitting diode or an active-matrix organic light-emitting diode (AMOLED), a flexible light-emitting diode (FLED), Miniled, MicroLed, Micro-oLed, quantum dot light-emitting diodes (QLED), etc. In one embodiment, the electronic device 100 may include 1 or N display screens 194, where N is a positive integer greater than 1.

The electronic device 100 can realize the shooting function through ISP, camera 193, video codec, GPU, display screen 194 and application processor.

ISP is used to process the data fed back by camera 193. For example, when taking a photo, the shutter is opened, and the light is transmitted to the camera photosensitive element through the lens. The light signal is converted into an electrical signal, and the camera photosensitive element transmits the electrical signal to ISP for processing and converts it into an image visible to the naked eye. ISP can also perform algorithm optimization on the noise, brightness, etc. of the image. ISP can also optimize the exposure, color temperature and other parameters of the shooting scene. In one embodiment, ISP can be set in camera 193.

The camera 193 is used to capture still images or videos. The object generates an optical image through the lens and projects it onto the photosensitive element. The photosensitive element can be a charge coupled device (CCD) or a complementary metal-oxide-semiconductor (CMOS) phototransistor. The photosensitive element converts the optical signal into an electrical signal, and then transmits the electrical signal to the ISP for conversion into a digital image signal. The ISP outputs the digital image signal to the DSP for processing. The DSP converts the digital image signal into an image signal in a standard RGB, YUV or other format. In one embodiment, the electronic device 100 may include 1 or N cameras 193, where N is greater than 1. A positive integer.

Digital signal processors are used to process digital signals. Video codecs are used to compress or decompress digital videos.

NPU is a neural-network (NN) computing processor, through which the intelligent cognition of the electronic device 100 can be realized, such as image recognition, face recognition, voice recognition, text understanding, etc.

The external memory interface 120 can be used to connect an external memory card, such as a Micro SD card, to expand the storage capacity of the electronic device 100. The external memory card communicates with the processor 110 through the external memory interface 120 to implement a data storage function. For example, files such as music and videos can be stored in the external memory card.

The internal memory 121 can be used to store computer executable program codes, which include instructions. The internal memory 121 may include a program storage area and a data storage area. The program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc. The data storage area may store data created during the use of the electronic device 100 (such as audio data, a phone book, etc.), etc. In addition, the internal memory 121 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, a universal flash storage (UFS), etc. The processor 110 executes various functional applications and data processing of the electronic device 100 by running instructions stored in the internal memory 121 and/or instructions stored in a memory provided in the processor.

The electronic device 100 can implement audio functions such as music playing and recording through the audio module 170, the speaker 170A, the receiver 170B, the microphone 170C, the headphone jack 170D, and the application processor.

The audio module 170 is used to convert digital audio information into analog audio signal output, and is also used to convert analog audio input into digital audio signals.

The speaker 170A, also called a "speaker", is used to convert an audio electrical signal into a sound signal. The electronic device 100 can listen to music or listen to a hands-free call through the speaker 170A.

The receiver 170B, also called a "earpiece", is used to convert audio electrical signals into sound signals. When the electronic device 100 receives a call or voice message, the voice can be received by placing the receiver 170B close to the human ear.

Microphone 170C, also called "microphone" or "microphone", is used to convert sound signals into electrical signals. When making a call or sending a voice message, the user can speak by approaching the microphone 170C with his mouth to input the sound signal into the microphone 170C.

The earphone jack 170D is used to connect a wired earphone.

The pressure sensor 180A is used to sense pressure signals and can convert pressure signals into electrical signals. In one embodiment, the pressure sensor 180A can be set on the display screen 194. When a touch operation is applied to the display screen 194, the electronic device 100 detects the touch operation strength according to the pressure sensor 180A. The electronic device 100 can also calculate the touch position according to the detection signal of the pressure sensor 180A. In one embodiment, touch operations acting on the same touch position but with different touch operation strengths can correspond to different operation instructions. For example: when a touch operation with a touch operation strength less than the first pressure threshold acts on the short message application icon, an instruction to view the short message is executed. When a touch operation with a touch operation strength greater than or equal to the first pressure threshold acts on the short message application icon, an instruction to create a new short message is executed.

The gyroscope sensor 180B can be used to determine the motion posture of the electronic device 100. The air pressure sensor 180C is used to measure the air pressure. The magnetic sensor 180D includes a Hall sensor. The electronic device 100 can use the magnetic sensor 180D to detect the opening and closing of the flip leather case. The acceleration sensor 180E can detect the magnitude of the acceleration of the electronic device 100 in various directions (generally three axes). The distance sensor 180F is used to measure the distance. The proximity light sensor 180G may include, for example, a light emitting diode (LED) and a light detector, such as a photodiode. The ambient light sensor 180L is used to sense the brightness of the ambient light. The fingerprint sensor 180H is used to collect fingerprints. The temperature sensor 180J is used to detect the temperature.

The touch sensor 180K is also called a "touch control device". The touch sensor 180K can be set on the display screen 194. The touch sensor 180K and the display screen 194 form a touch screen, also called a "touch control screen". The touch sensor 180K is used to detect touch operations acting on or near it. The touch sensor can pass the detected touch operation to the application processor to determine the type of touch event. Visual output related to the touch operation can be provided through the display screen 194. In another embodiment, the touch sensor 180K can also be set on the surface of the electronic device 100, which is different from the position of the display screen 194.

The bone conduction sensor 180M can obtain vibration signals. The button 190 includes a power button, a volume button, etc. The button 190 can be a mechanical button. It can also be a touch button. The motor 191 can generate a vibration prompt. The indicator 192 can be an indicator light, which can be used to indicate the charging status, power change, and can also be used to indicate messages, missed calls, notifications, etc. The SIM card interface 195 is used to connect a SIM card.

The software system of the electronic device 100 may adopt a layered architecture, an event-driven architecture, a micro-core architecture, a micro-service architecture, or a cloud architecture. For example, the software system of the layered architecture may be an Android system, or a Harmony operating system (OS), or other software systems. The embodiment of the present application takes the Android system of the layered architecture as an example to illustrate the software structure of the electronic device 100.

FIG. 2B exemplarily shows a schematic diagram of a software architecture of an electronic device 100 .

The layered architecture divides the software into several layers, each with a clear role and division of labor. The layers communicate with each other through software interfaces. In one embodiment, the Android system is divided into four layers, from top to bottom: the application layer, the application framework layer, the Android runtime and system library, and the kernel layer.

The application layer can include a series of application packages.

As shown in FIG2B , the application package may include a search engine, a browser, a shopping application, a social media application, an application market, a video, music, etc. The application in the present application may also be replaced by other software such as a mini-program, an atomization service, etc. Any of the above-mentioned examples of application may be used to display the advertising copy generated online based on the advertising copy generation method of the present application.

The application framework layer provides application programming interface (API) and programming framework for the applications in the application layer. The application framework layer includes some predefined functions.

As shown in FIG. 2B , the application framework layer may include a window manager, a content provider, a view system, a telephony manager, a resource manager, a notification manager, and the like.

The window manager is used to manage window programs. The window manager can obtain the display screen size, determine whether there is a status bar, lock the screen, capture the screen, etc.

Content providers are used to store and retrieve data and make it accessible to applications. The data may include videos, images, audio, calls made and received, browsing history and bookmarks, phone books, etc.

The view system includes visual controls, such as controls for displaying text, controls for displaying images, etc. The view system can be used to build applications. A display interface can be composed of one or more views. For example, a display interface including a text notification icon can include a view for displaying text and a view for displaying images.

The phone manager is used to provide communication functions of the electronic device 100, such as management of call status (including connecting, hanging up, etc.).

The resource manager provides various resources for applications, such as localized strings, icons, images, layout files, video files, and so on.

The notification manager enables applications to display notification information in the status bar, which can be used to convey notification-type messages and can disappear automatically after a short stay without user interaction. For example, the notification manager is used to notify download completion, message reminders, etc. The notification manager can also be a notification that appears in the system top status bar in the form of a chart or scroll bar text, such as a notification of an application running in the background, or a notification that appears on the screen in the form of a dialog window. For example, a text message is prompted in the status bar, a prompt sound is emitted, the electronic device 100 vibrates, an indicator light flashes, etc.

Android Runtime includes core libraries and virtual machines. Android Runtime is responsible for the scheduling and management of the Android system.

The core library consists of two parts: one part is the function that needs to be called by the Java language, and the other part is the Android core library.

The application layer and the application framework layer run in a virtual machine. The virtual machine executes the Java files of the application layer and the application framework layer as binary files. The virtual machine is used to perform functions such as object life cycle management, stack management, thread management, security and exception management, and garbage collection.

The system library can include multiple functional modules, such as surface manager, media library, 3D graphics processing library (such as OpenGL ES), 2D graphics engine (such as SGL), etc.

The surface manager is used to manage the display subsystem and provide the fusion of 2D and 3D layers for multiple applications.

The media library supports playback and recording of a variety of commonly used audio and video formats, as well as static image files, etc. The media library can support a variety of audio and video encoding formats, such as: MPEG4, H.264, MP3, AAC, AMR, JPG, PNG, etc.

The 3D graphics processing library is used to implement 3D graphics drawing, image rendering, compositing, and layer processing.

A 2D graphics engine is a drawing engine for 2D drawings.

The kernel layer is the layer between hardware and software. The kernel layer contains at least display driver, camera driver, audio driver, and sensor driver.

The following is an illustrative description of the workflow of the software and hardware of the electronic device 100 in conjunction with capturing a photo scene.

When the touch sensor 180K receives a touch operation, the corresponding hardware interrupt is sent to the kernel layer. The kernel layer processes the touch operation into a raw input event (including touch coordinates, timestamp of the touch operation, and other information). The raw input event is stored in the kernel layer. The application framework layer obtains the raw input event from the kernel layer and identifies the control corresponding to the input event. For example, if the touch operation is a touch single-click operation and the control corresponding to the single-click operation is the control of the camera application icon, the camera application calls the interface of the application framework layer to start the camera application, and then starts the camera driver by calling the kernel layer to capture static images or videos through the camera 193.

Next, the network device 300 in the present application is introduced as an example.

FIG. 2C exemplarily shows a schematic diagram of the hardware structure of a network device 300 .

As shown in FIG2C , the network device 300 may include one or more processors 310, a communication interface 320, and a memory 330, wherein the processor 310, the communication interface 320, and the memory 330 may be connected via a bus or other means, and FIG2C takes the connection via a bus 340 as an example. In: The processor 310 may be composed of one or more general-purpose processors, such as a CPU. The processor 310 may be used to run program codes related to the advertising copy generation method. The communication interface 320 may be a wired interface (such as an Ethernet interface) or a wireless interface (such as a cellular network interface or a wireless local area network interface), which is used to communicate with other nodes, such as communicating with the electronic device 100 and the electronic device 200. The memory 330 may include a volatile memory (volatile memory), such as a random access memory (random access memory, RAM); the memory may also include a non-volatile memory (non-vlatile memory), such as a read-only memory (read-only memory, ROM), a flash memory (flash memory), a hard disk drive (hard disk drive, HDD) or a solid state disk (solid state disk, SSD). The memory 330 may also include a combination of the above-mentioned types of memory. The memory 330 may be used to store a set of program codes so that the processor 310 calls the program codes stored in the memory 330 to implement the method implemented on the network device 300 side in the embodiment of the present application. In the embodiment of the present application, the memory 330 may also be a storage array, and so on.

In one embodiment, the network device 300 may include at least one server, for example, the network device 300 is a server cluster composed of multiple servers. Any server may be a hardware server or a cloud server, for example, a web server, a backend server, an application server, a download server, etc. The hardware structure of any server may refer to the hardware structure of the network device 300 shown in FIG. 2C .

The network device 300 shown in FIG. 2C is an implementation example of an embodiment of the present application. In actual applications, the network device 300 may also include more or fewer components, and the present application does not limit this.

The following describes the application scenarios involved in the embodiments of the present application, user interface examples in the application scenarios, and an advertising copy generation method that can implement the application scenarios. The method can be applied to the advertising copy generation system 10 shown in FIG1A , and can also be applied to the advertising copy generation system 10 shown in FIG1B .

Scenario 1: When user 1 uses electronic device 100 to search for keyword 1 and keyword 2 respectively, electronic device 100 displays different advertisements for the same object. For specific examples, see FIGS. 3 and 4 below.

Figures 3 and 4 illustrate the advertising scenario of the product "Brand A sun hat" as an example. It is assumed that the advertising information of "Brand A sun hat" includes "Brand A sun hat, a super practical summer sun hat, made of high-quality fabrics, can effectively prevent the invasion of ultraviolet rays and give you a refreshing summer. It is a must-have sun hat for trendsetters, with novel styles and rich colors to show your handsomeness." Figure 3 shows the process of user 1 searching for keyword 1 (assuming it is "practical sun hat"), and Figure 4 shows the process of user 1 searching for keyword 2 (assuming it is "handsome sun hat").

As shown in FIG. 3 (A), the electronic device 100 may display a user interface 310 of a browser, and the user interface 310 may include a search box 311 and a search control 312. The search box 311 may be used to input search keywords, and the search control 312 may be used to trigger a search for web page results related to the keywords in the search box 311. The electronic device 100 may receive the keyword 1 "practical sun hat" input by the user in the search box 311, and then receive a user operation (such as a touch operation, such as a click operation) on the search control 312. In response to the user operation, the electronic device 100 may obtain and display web page results related to the keyword 1, and advertising content related to the keyword 1. For specific examples, see the user interface 320 shown in FIG. 3 (B).

As shown in (B) of FIG. 3 , the user interface 320 of the browser may include a search box 311, a display column of at least one advertisement related to keyword 1 (e.g., advertisement column 321), and at least one web page result related to keyword 1 (e.g., web page result 322 and web page result 323). Among them, the search box 311 may display the currently searched keyword 1 "practical sun hat". The advertisement column 321 may be used to display an advertisement for the product "Brand A sun hat", and may include an advertisement picture 321A, an advertisement copy 321B, an advertisement source 321C (i.e., "shopping platform 1"), and an advertisement indication 321D (i.e., "advertisement"). The advertisement copy 321B is "super practical summer sun hat, which can effectively prevent ultraviolet invasion and give you a refreshing summer". The web page result 322 may include a title 322A (including the characters "summer sun hat collection, versatile and practical..."), video information 322B in the web page, and a source URL 322C (i.e., "URL 111"). The web page result 323 may include a title 323A (including characters “[outdoor sun hat] price-brand-mall 111…”), a web page description 323B and a source URL 323C (ie, “URL 222”).

As shown in (A) of Figure 4, the electronic device 100 can display the user interface 310 shown in (A) of Figure 3. The electronic device 100 can receive the keyword 2 "handsome sun hat" input by the user in the search box 311, and then receive the user operation on the search control 312 (for example, a touch operation, such as a click operation). In response to the user operation, the electronic device 100 can obtain and display web page results related to keyword 2, as well as advertising content related to keyword 2. For specific examples, see the user interface 410 shown in (B) of Figure 4.

As shown in (B) of FIG. 4 , the user interface 410 of the browser may include a search box 311, a display column of at least one advertisement related to keyword 2 (e.g., advertisement column 411), and at least one web page result related to keyword 2 (e.g., web page result 412 and web page result 413). Among them, the search box 311 may display the currently searched keyword 2 “handsome sun hat”. The advertisement column 411 may be used to display an advertisement for the product “Brand A sun hat”, and may include an advertisement picture 411A, advertisement text 411B, advertisement source 411C (i.e., “shopping platform 1”), and advertisement indication 411D (i.e., “advertisement”). The advertisement text 411B is “a must-have sun hat for trendsetters, colorful and showing your handsomeness”. The web page result 412 may include a title 412A (including the characters “handsome sun hat - product - hot sale on the entire network…”), multiple pictures 412B in the web page, and a source network 411C. Web page result 413 may include title 413A (including characters "Baseball caps are out of fashion, sun hats are popular now ..."), web page description 413B and source URL 413C (ie, "URL 444").

Among them, the advertisement column 321 in the user interface 320 shown in (B) of FIG. 3 and the advertisement column 411 in the user interface 410 shown in (B) of FIG. 4 are both used to display the advertisement content of "Brand A sun hat". The advertisement picture 321A and the advertisement source 321C in the advertisement column 321 are respectively the same as the advertisement picture 411A and the advertisement source 411C in the advertisement column 411, but the advertisement copy 321B in the advertisement column 321 is different from the advertisement copy 411B in the advertisement column 411. The advertisement copy 321B is a copy fragment related to the keyword 1 "practical sun hat" in the advertisement information of "Brand A sun hat", and the advertisement copy 411B is a copy fragment related to the keyword 2 "handsome sun hat" in the advertisement information of "Brand A sun hat". In other words, the present application can generate and recommend personalized advertisement copy online according to the keywords searched by the user. The advertisement copy can meet the real-time needs of the user when searching for keywords, thereby increasing the probability of the user operating the advertisement copy to view the detailed advertisement content and improving the efficiency of advertisement delivery.

Not limited to the examples in Figures 3 and 4, in other examples, scenario 1 may also be an advertising scenario for the product "Brand B mobile phone". Assuming that the advertising information of "Brand B mobile phone" includes "Brand B full-screen mobile phone, incredibly powerful performance, in-screen fingerprint support, 90Hz refresh rate, 1440Hz high-frequency dimming", when keyword 1 is "high-performance full-screen mobile phone", the advertising copy of "Brand B mobile phone" displayed by the electronic device 100 may be a copy fragment related to keyword 1 in the advertising information: "Brand B full-screen mobile phone, incredibly powerful performance"; when keyword 2 is "fingerprint payment mobile phone", the advertising copy of "Brand B mobile phone" displayed by the electronic device 100 may be a copy fragment related to keyword 2 in the advertising information: "Brand B full-screen mobile phone, in-screen fingerprint support".

Please refer to Figure 5, which is a flowchart of a method for generating advertising copy provided by the present application. The method may include but is not limited to the following steps:

S101: The network device 300 trains to obtain a rough selection model.

In one embodiment, S101 is an optional step. In one embodiment, S101 is an offline operation, that is, it is performed before advertisement delivery/display.

In one embodiment, the network device 300 can train a rough selection model based on the searched keywords and advertising information. In some examples, when the network device 300 trains the rough selection model, it can obtain the vector representation of the keyword and the vector representation of the text (including multiple sentences/statements) in the advertising information, respectively, and then pass these vector representations through the subsequent attention module (such as Cross Attention) and the fully connected layer (such as a multilayer perceptron (MLP)) to predict the probability of each sentence in the advertising information being selected by the keyword, and this probability can characterize the degree of relevance between the sentence and the keyword. Among them, when the network device 300 trains the rough selection model, the vector representation of the keyword and the vector representation of the sentence in the advertising information can be decoupled, that is, each vector representation can exist independently, and there may be no interaction between these vector representations, so as to implement S103. The vector representation in the present application can be a vector representation of the current sentence or phrase in a certain semantic space, such as a word embedding vector (embedding), for example, obtained through a word encoder (sentence encoder).

S102: The network device 300 is trained to obtain a selected model.

In one embodiment, S102 is an optional step. In one embodiment, S102 is an offline operation.

In one embodiment, the network device 300 can train a selection model based on the searched keywords and advertising information. In some examples, when the network device 300 trains the selection model, it can first use prompts to concatenate sentences and keywords in the advertising information and obtain corresponding vector representations, and then send them into a new prediction model for training to predict the probability of each sentence in the advertising information being selected by the keyword. The probability can represent the degree of relevance between the sentence and the keyword.

It can be understood that the training data of the coarse selection model and the fine selection model can be the same or different, but the model structures are not the same. The time consumption of the fine selection model is often greater than that of the coarse selection model. At the same time, the probability predicted by the fine selection model is often more accurate than that predicted by the coarse selection model.

S103: The network device 300 stores the vector representation of the advertisement information in the advertisement library.

In one embodiment, S103 is an optional step. In one embodiment, S103 is an offline operation.

In one implementation, the network device 300 may obtain the vector representation of each sentence of the advertisement information in the advertisement library. For example, when training the rough selection model in S101, the vector representation of some or all sentences in the advertisement information may be obtained. Then, the network device 300 may store these vector representations, and the stored vector representations of the sentences may be collectively referred to as an advertisement material library.

In one embodiment, the network device 300 may store the identity document (ID) of each sentence in the advertisement information and the vector representation of the sentence. This may be understood as identifying the vector representation of the sentence by the ID of the sentence, so as to obtain the vector representation of the sentence when subsequently generating advertisement copy online, such as obtaining the vector representation of the sentence according to the ID of the sentence in the following S107.

The order of any two steps in S101-S103 is not limited.

S104: The electronic device 100 obtains a first keyword input by a first user.

In one implementation, the electronic device 100 receives a first keyword and a corresponding search request input by a first user, and in response to the search request, obtains the first keyword and obtains search results related to the first keyword, for example, executing S105 to obtain related advertisements. In some examples, in the scenario shown in Figure 3, the electronic device 100 can receive the keyword 1 "practical sun hat" entered by the user in the search box 311 of the user interface 310 shown in (A) of Figure 3, and then receive the user operation on the search control 312 (which can be understood as receiving a search request), and in response to the user operation, obtain keyword 1, and obtain web page results related to keyword 1, as well as advertising content related to keyword 1.

S105: The electronic device 100 sends a first advertisement request (carrying a first keyword) to the network device 300.

In one implementation, the first advertisement request may be used to request to obtain advertisement content, the first advertisement request may carry a first keyword, and the first keyword may be used to obtain advertisement content related to the first keyword.

S106: The network device 300 obtains a first vector representation of the first keyword.

In one implementation, the network device 300 may use a rough selection model trained offline to obtain a vector representation of the first keyword (referred to as a first vector representation).

S107: The network device 300 determines a first advertisement to be displayed, and obtains a second vector representation of the advertisement information.

In one implementation, the network device 300 may first use the first keyword to recall and sort multiple advertisements to obtain a first advertisement related to the first keyword. In some examples, it is assumed that the multiple advertisements maintained by the network device 300 include advertisements for products such as "Brand A Sun Hat", "Brand B Mobile Phone", "Brand X Car", and "C Financial Services". The first keyword is keyword 1 "Practical Sun Hat" shown in FIG3. The network device 300 may obtain an advertisement for the product "Brand A Sun Hat" related to keyword 1 from the above multiple advertisements. Optionally, the first advertisement may be characterized by a corresponding ID, and the IDs of advertisements of different objects are different.

In one embodiment, after the network device 300 determines the first advertisement to be displayed, it can obtain the vector representation of each sentence in the advertisement information of the first advertisement from the stored advertisement material library (collectively referred to as the second vector representation). In some examples, the network device 300 can first obtain the advertisement information of the first advertisement (including multiple sentences) and/or the ID of the sentence in the advertisement information according to the ID of the first advertisement, and then obtain the vector representation of the sentence of the first advertisement according to the sentence and/or the ID of the sentence. It can be understood that the stored, offline-calculated advertisement material library is directly used, rather than the network device 300 online calculating the vector representation of the sentence in the advertisement information, which effectively reduces the processing pressure.

The order of S106 and S107 is not limited.

S108: The network device 300 uses the coarse selection model to process the first vector representation and the second vector representation to obtain K sentences of the first advertisement (K is a positive integer, for example, a positive integer greater than 1).

In one implementation, the network device 300 may use the first vector representation of the keyword and the second vector representation of the sentence of the first advertisement as inputs of the rough selection model, and obtain output: K sentences in the advertisement information of the first advertisement. In some examples, the rough selection model may score each sentence in the advertisement information of the first advertisement, and the obtained score may indicate the relevance between the corresponding sentence and the keyword, and the higher the score, the higher the relevance, and the rough selection model may select K sentences with scores ranked in the top K positions, where K is a hyperparameter and may be determined by manual experience or engineering experiment results.

S109: The network device 300 uses the selection model to process the first keyword and the K sentences of the first advertisement to obtain scores of the K sentences.

In one embodiment, the network device 300 can use the first keyword and the K sentences obtained in S108 as inputs of the selection model, and obtain outputs: K scores corresponding to the K sentences. It can be understood that, like the scores of the advertisement sentences obtained by the rough selection model, the scores of the advertisement sentences obtained by the selection model can also indicate the relevance between the corresponding sentences and the keywords. The higher the score, the higher the relevance. However, the structure of the selection model is different from that of the rough selection model. The selection model will perform interactive fusion operations (such as prompt) on the keywords and advertisement sentences, so it can more accurately predict the relevance between the keywords and advertisement sentences, that is, the scores of the advertisement sentences obtained by the selection model are more accurate.

S110: The network device 300 selects N sentences (ie, short-term interests) whose scores are greater than or equal to a first threshold from the K sentences (N is a positive integer less than or equal to K).

In one implementation, the network device 300 may use the first threshold value set by the hyperparameter, based on the K scores obtained in S109, to select: N sentences with scores greater than or equal to the first threshold value from the K sentences of the first advertisement obtained in S108. Optionally, these N sentences may be the most relevant content in the advertisement information of the first advertisement and the first keyword currently searched, and may be referred to as the user's short-term interest expression in the current advertisement information.

S111: The network device 300 generates a first copy according to N sentences of the first advertisement.

In one implementation, the network device 300 may generate a first copy of the first advertisement based on the N sentences of the first advertisement obtained in S110, which may be understood as language organization of the N sentences into a smooth copy.

S112 : The network device 300 sends the first document to the electronic device 100 .

S113: The electronic device 100 displays the first text.

In one embodiment, after the electronic device 100 sends the first advertisement request to the network device 300, the network device 300 may receive a response based on the first advertisement request. The first advertisement requests the generated advertisement text (ie, the first text) and displays the first text.

A specific example of the method shown in Figure 5 can be found in Figure 3, wherein the first keyword is keyword 1 "practical sun hat" shown in Figure 3, and the K sentences of the first advertisement are 9 sentences of the product "Brand A sun hat": "Brand A sun hat", "Super practical summer sun hat", "Made of high-quality fabrics", "Can effectively prevent ultraviolet invasion", "Give you a refreshing summer", "A must-have sun hat for trendsetters", "Novel style", "Rich colors", "Show your handsomeness", and the N sentences screened by the network device 300 are 3 sentences out of the above 9 sentences: "Super practical summer sun hat", "Can effectively prevent ultraviolet invasion", "Give you a refreshing summer", and the advertisement content of "Brand A sun hat" sent by the network device 300 and displayed by the electronic device 100 can be found in the advertisement column 321 in the user interface 320 shown in (B) of Figure 3, and the advertisement copy 321B in the advertisement column 321 ("Super practical summer sun hat, can effectively prevent ultraviolet invasion, give you a refreshing summer") is the first copy.

Not limited to the above examples, in other examples, some sentences included in the first copy may be sentences that are different from the sentences in the short-term interest but have similar meanings. For example, in FIG4 , the first keyword is keyword 2 “handsome sun hat” shown in FIG4 , and the K sentences of the first advertisement are 9 sentences of “Brand A sun hat”: “Brand A sun hat”, “Super practical summer sun hat”, “Made of high-quality fabrics”, “Can effectively prevent ultraviolet invasion”, “Give you a cool summer”, “A sun hat that must be chosen by fashion gurus”, “Novel style”, “Rich colors”, “Show your handsomeness”, network device 30 0 The N sentences obtained by screening are 3 sentences among the above 9 sentences: "Sun hats that must be chosen by trendsetters", "Rich colors", "Show your handsomeness", and the advertisement content of "Brand A sun hats" sent by network device 300 and displayed by electronic device 100 can be found in advertisement column 411 in user interface 410 shown in (B) of FIG. 4. Advertisement copy 411B in advertisement column 411 ("Sun hats that must be chosen by trendsetters, colorful and show your handsomeness") is the first copy. Although "colorful and colorful" in the first copy is different from "rich colors" obtained by the above screening, they have similar meanings. In other examples, some sentences included in the first copy may be obtained by expanding and supplementing sentences in short-term interests. The present application does not limit the specific implementation method of generating advertisement copy (i.e., S111) according to short-term interests.

In one implementation, S104-S113 are online operations, that is, operations performed when an advertisement is recommended to the user online after the user performs a triggering operation in real time.

In the method shown in Figure 5, the network device 300 can extract sentences with strong relevance from the advertising information online based on the keywords searched by the user online (i.e., explore the user's current short-term interests), and then generate a first copy based on these sentences for the electronic device 100 to recommend to the user for viewing, rather than simply combining popular keywords and advertising information to obtain advertising copy through offline data processing. After subsequently obtaining the keywords searched by the user online, the keywords searched by the user online are found from the above-mentioned popular keywords stored offline, and the advertising copy related to the keyword is returned to the user for viewing. This not only avoids the situation where some keywords searched by the user are not within the above-mentioned popular keywords, resulting in the inability to provide the user with suitable advertising copy, but also the first copy of the present application is more in line with the content of the advertising information itself while taking into account the keywords searched by the user online, and is therefore more in line with the needs and preferences of the user, thereby improving the efficiency of advertising delivery and user experience.

In addition, the network device 300 uses offline storage (vector representation of advertisement sentences) and online prediction (probability of advertisement sentences being selected by keywords) (i.e., rough selection model) to perform rough selection of sentences in the first stage, and then further selects sentences based on the rough selected sentences through a subsequent interactive model (i.e., refined model), thereby avoiding the situation where the single use of the rough selection model results in a small delay but low accuracy, and also avoiding the situation where the single use of the refined model results in a high accuracy but large delay. In this way, the coordination of the rough selection model and the refined model not only realizes online model reasoning and ensures the accuracy of the model, but also alleviates the pressure of online delay.

Scenario 2: The multiple advertising copies operated by user 1 while using electronic device 100 are different from the multiple advertising copies operated by user 2 while using electronic device 200. Subsequently, when user 1 and user 2 use electronic device 100 and electronic device 200 respectively to search for keyword 1, the advertising copies of the same object displayed on electronic device 100 and electronic device 200 are different. Specific examples can be seen in Figures 6 and 7 below.

Figures 6 and 7 illustrate the advertising scenario of the product "Brand A sun hat" as an example. It is assumed that the advertising information of "Brand A sun hat" includes "Brand A sun hat, a super practical summer sun hat, made of high-quality fabrics, can effectively prevent the invasion of ultraviolet rays and give you a refreshing summer. It is a must-have sun hat for trendsetters. It has novel styles and rich colors to show your handsomeness." Figure 6 shows the process of user 1 using an electronic device 100 to operate multiple advertising copies and search for keyword 1 (assuming it is "practical sun hat"). Figure 7 shows the process of user 2 using an electronic device 200 to operate multiple advertising copies and search for keyword 1.

In one embodiment, as shown in (A) of FIG. 6 , the electronic device 100 may display a user interface 610 of a social media application. The user interface 610 may include a search box 611, a search control 612, and multiple published contents (e.g., content 613 and content 614). The search box 611 may be used to input search keywords, and the search control 612 may be used to trigger a search for published contents related to the search box 611. Content 614 may be used to display the content of advertisement a, and may include advertising copy 614A (i.e., "This fish tank is cleverly designed and is simply a work of art"). The user interface 610 may be a home page, and the published contents displayed on the home page of the social media application may change after refreshing (e.g., in response to a user operation of sliding from top to bottom, or jumping from other pages to the home page). The electronic device 100 may receive a user operation (e.g., a touch operation, such as a click operation) of user 1 on content 614, and in response to the user operation, display content related to advertisement a (e.g., a purchase page for a product corresponding to advertisement a, not shown in FIG. 6 ). In the above process, user 1 clicks on the advertising copy 614A of advertisement a and And the electronic device 100 displays the corresponding advertising content.

In another embodiment, as shown in (B) of FIG. 6 , the user interface 620 of the social media application may include a search box 611, a search control 612, and multiple published contents (e.g., content 621 and content 622). The user interface 620 is similar to the user interface 610, except that the displayed published contents are different. Content 622 may be used to display the content of advertisement b, and may include advertisement copy 622A (i.e., "a new floral long skirt, soft and comfortable to the touch, simple and elegant design style, allowing you to stand out from the crowd"). The electronic device 100 may receive a user operation (e.g., a touch operation, such as a click operation) of user 1 on content 622, and in response to the user operation, display content related to advertisement b (e.g., a purchase page of a product corresponding to advertisement b, not shown in FIG. 6 ). In the above process, user 1 clicks on advertisement copy 622A of advertisement b and the electronic device 100 displays the corresponding advertisement content.

After the implementation shown in (A) and (B) of FIG. 6 , the electronic device 100 may display the user interface 630 shown in (C) of FIG. 6 . As shown in (C) of FIG. 6 , the user interface 630 of the social media application may include a search box 611, a search control 612, and multiple published contents (e.g., content 631 and content 632), and the user interface 630 is similar to the user interface 610, except that the displayed published contents are different. Content 632 may be used to display the content of advertisement c (i.e., the advertisement of the product “Brand A sun hat”), and may include an advertising manufacturer 632A (i.e., “Manufacturer ccc”), an advertising instruction 632B (i.e., “advertisement”), an advertising copy 632C (i.e., “Brand A sun hat is made of high-quality fabrics and cleverly designed to make you stand out from the crowd”) and an advertising video 632D.

In one embodiment, as shown in (A) of FIG. 7 , the electronic device 200 may display a user interface 710 of a social media application, and the user interface 710 may be a home page. The user interface 710 may include a search box 711, a search control 712, and multiple published contents (e.g., content 713 and content 714). The search box 711 may be used to input search keywords, and the search control 712 may be used to trigger the search and the published contents related to the search box 711. Content 713 may be used to display the content of advertisement d, and may include advertisement copy 713A (i.e., "sleeping pants are specially designed for pregnant women, comfortable and skin-friendly, allowing you to feel comfortable all night"). The electronic device 200 may receive a user operation (e.g., a touch operation, such as a click operation) of user 2 on content 713, and in response to the user operation, display content related to advertisement d (e.g., a purchase page of a product corresponding to advertisement d, not shown in FIG. 7 ). In the above process, user 2 clicks on the advertisement copy 713A of advertisement d and the electronic device 200 displays the corresponding advertisement content.

In one embodiment, as shown in (B) of FIG. 7 , the user interface 720 of the social media application is similar to the user interface 710, except that the displayed published content is different, and the multiple published content in the user interface 720 includes, for example, content 721 and content 722. Content 722 can be used to display the content of advertisement e, and can include advertising copy 722A (i.e., "This shirt is novel in style, breathable and comfortable, allowing you to easily cope with various occasions"). The electronic device 200 can receive a user operation (e.g., a touch operation, such as a click operation) of user 2 on content 722, and in response to the user operation, display content related to advertisement e (e.g., a purchase page of a product corresponding to advertisement e, not shown in FIG. 7 ). In the above process, user 2 clicks on the advertising copy 722A of advertisement e and the electronic device 200 displays the corresponding advertising content.

After the implementation shown in (A) and (B) of FIG. 7 , the electronic device 200 may display the user interface 730 shown in (C) of FIG. 7 . As shown in (C) of FIG. 7 , the user interface 730 of the social media application is similar to the user interface 710, except that the displayed published content is different, and the multiple published contents in the user interface 730 include, for example, content 731 and content 732. Content 732 may be used to display the content of advertisement c (i.e., the advertisement of the product “Brand A sun hat”), and may include an advertising manufacturer 732A (i.e., “Manufacturer ccc”), an advertising instruction 732B (i.e., “advertisement”), an advertising copy 732C (i.e., “Brand A sun hat is made of high-quality fabrics, comfortable to wear, allowing you to easily cope with various occasions”) and an advertising video 732D.

Among them, the content 632 in the user interface 630 shown in (C) of FIG. 6 and the content 732 in the user interface 730 shown in (C) of FIG. 7 are both used to display the advertisement content of the product "Brand A Sun Hat". The advertisement manufacturer 632A and the advertisement video 632D in the content 632 are respectively the same as the advertisement manufacturer 732A and the advertisement video 732D in the content 732, but the advertisement copy 632C in the content 632 is different from the advertisement copy 732C in the content 732. The advertisement copy 632C may be an advertisement copy related to the user's historical behavior generated based on the operation behavior of the user 1 on the advertisement copy 614A of advertisement a and the advertisement copy 622A of advertisement b. The advertisement copy 632C includes the words "cleverly designed" in the advertisement copy 614A and "making you stand out from the crowd" in the advertisement copy 622A. Advertisement copy 732C may be an advertisement copy related to the user's historical behavior generated based on the operation behavior of user 2 on advertisement copy 713A of advertisement d and advertisement copy 722A of advertisement e. Advertisement copy 732C includes "comfortable and skin-friendly" in advertisement copy 713A and "allowing you to easily cope with various occasions" in advertisement copy 722A. In other words, the present application may generate and recommend personalized advertisement copy based on the user's historical behavior, and the advertisement copy may meet the user's preference for the copy type/style (reflected by historical behavior), thereby increasing the probability of the user operating the advertisement copy to view detailed advertisement content and improving the efficiency of advertisement delivery.

In one embodiment, the generation of advertising copy not only utilizes the user's historical behavior, but also utilizes the user's static features. In some examples, such as the scenarios shown in FIG6 and FIG7 , the static features of user 1 include: age 30, gender female, and user profile as a fashionista. Therefore, the advertising copy 632C displayed for user 1 includes the historical copy "cleverly designed to make you stand out from the crowd" instead of the historical copy "soft and comfortable to the touch". The static features of user 2 include: age 50, gender female, and user profile as practical and cost-effective. Therefore, the advertising copy 732C displayed for user 2 includes the historical copy "comfortable and skin-friendly, allowing you to easily cope with various In other examples, the static features of the user include: age 30, gender female, and user profile as a fashionista. The copy that the user has clicked on in the past includes: the advertising copy of "M brand watches""M brand watches have new colors, why not take action?" and the advertising copy of "N brand bags""New autumn bags, keep up with the trend, without losing personality." The advertising copy that is subsequently displayed to the user may include "new colors" and/or "keep up with the trend, without losing personality."

Not limited to the above examples, in other examples, the user interface 630 shown in (C) of FIG6 and the user interface 730 shown in (C) of FIG7 may both be obtained by searching for keyword 1 “practical sun hat”. It can be understood that, when different users have different historical behaviors, even if users search for the same keyword and are recommended advertisements for the same object, the displayed advertising copy may also be different (related to historical behavior).

FIG6 and FIG7 illustrate different historical behaviors of different users by taking them as examples. In other examples, FIG7 may not be executed by electronic device 200 but by electronic device 100. User 1 may implement the process shown in FIG6 in period 1 and the process shown in FIG7 in period 2. The historical behaviors (which may reflect the copywriting preference) of user 1 in different periods in the same scenario may be different, or the historical behaviors of user 1 in different scenarios may also be different. The electronic device may obtain relevant advertising copywriting based on the historical behaviors to meet the user's preference for copywriting type/style in various scenarios.

Please refer to Figure 8, which is a flowchart of another method for generating advertising copy provided by the present application. The method may include but is not limited to the following steps:

S201: The network device 300 obtains a long-term interest model through training.

In one embodiment, S201 is an optional step. In one embodiment, S201 is an offline operation, that is, it is performed before advertisement delivery/display.

In one embodiment, the network device 300 can train a long-term interest model based on the static features and behavior data of multiple users, and abstract the user's preferences for copy type/style, etc. in a model-generated manner. The specific training process can be seen in Figure 9 below and will not be described in detail for the time being.

S202: The electronic device 100 uploads the historical behavior data and static features of the first user to the network device 300.

S203: The network device 300 obtains historical behavior data and static features of the first user.

In one implementation, the network device 300 may receive the historical behavior data and static features of the first user uploaded by the electronic device 100 and store these data. These data may be obtained through the ID of the corresponding first user.

In the present application, the static characteristics of the user may include, but are not limited to, age, gender, city, type of electronic device used, user portrait, etc. The user's behavior data/historical behavior data may include, but are not limited to, data on operations and browsing behaviors for advertisements and information related to advertisements, such as, but not limited to, clicks on advertisements, browses advertisements, closes advertisements, purchases corresponding goods/services through links in advertisements, downloads applications or other forms of software through links in advertisements, duration of browsing advertisements, types of operated/browsed advertisements, etc. The types of operated/browsed advertisements are, for example, the types of advertisements preferred by the user recorded by the electronic device, such as information flow advertisements, opening screen advertisements, or interstitial advertisements.

The order of S201 and S202-S203 is not limited.

S204: The electronic device 100 receives a first operation of a first user.

In one embodiment, the electronic device 100 may receive a first operation, and in response to the first operation, obtain advertising content, and the first operation may be understood as a triggering operation for advertising delivery/display. In some examples, in the scenario shown in FIG6 , the electronic device 100 may receive a user operation for refreshing the homepage of a social media application, such as a sliding operation from top to bottom, and in response to the user operation (i.e., the first operation), display the advertising content 632 in the user interface 630 shown in (C) of FIG6 . An example of the first operation may also be found in the description of S104 of FIG5 .

S205: The electronic device 100 sends a second advertisement request (carrying the identifier of the first user) to the network device 300.

In one implementation, the second advertisement request may be used to request to obtain advertisement content. The second advertisement request may carry the ID of the first user, and the ID of the first user may be used by the network device 300 to determine the corresponding static features and historical behavior data of the first user.

S206: The network device 300 determines a second advertisement to be displayed, and obtains advertisement information.

In one implementation, the network device 300 may determine the second advertisement to be displayed from the maintained multiple advertisements according to a preset rule. In some examples, the network device 300 may pre-set the display order of multiple advertisements and determine the advertisement to be displayed according to the order. In other examples, examples of preset rules may also refer to the description of S107 of FIG. 5. The present application does not limit the specific content of the preset rules. In some examples, the network device 300 may obtain the advertisement information of the second advertisement according to the ID of the second advertisement. The advertisement information may include the text of the advertisement, as well as non-text content such as pictures, videos, and links. The text of the advertisement may include multiple sentences.

S207: The network device 300 obtains historical behavior data and static features of the first user according to the identifier of the first user.

In one implementation, the network device 300 may determine the historical behavior data and static features of the first user from the historical behavior data and static features of at least one user stored according to the ID of the first user.

The order of S206 and S207 is not limited.

S208: The network device 300 uses the long-term interest model to process the historical behavior data of the first user, the static features of the first user, and the advertisement information of the second advertisement to obtain a second copy of the second advertisement.

In one embodiment, the network device 300 may use the historical behavior data of the first user, the static features of the first user, and the text in the advertising information of the second advertisement as inputs to the long-term interest model, and obtain the output: the second copy of the second advertisement. The text in the advertising information may include at least one of the following: the advertisement name, theme, content summary, description, and advertising copy (which may be referred to as offline advertising copy) input by the advertiser, and the text obtained after crawling the landing page of the advertisement. Optionally, the network device 300 may modify/rewrite the offline advertising copy based on the historical behavior data and static features of the first user to obtain the second copy of the second advertisement.

S209 : The network device 300 sends the second text to the electronic device 100 .

S210: The electronic device 100 displays the second text.

In one implementation, after sending the second advertisement request to the network device 300 , the electronic device 100 may receive the advertisement text (ie, the second text) generated by the network device 300 based on the second advertisement request, and display the second text.

A specific example of the method shown in Figure 8 can be found in Figure 6, wherein the historical behavior data of the first user includes: the user's behavior of clicking on the advertising copy 614A in the user interface 610 shown in Figure 6 (A) and browsing the corresponding advertising content, and the user's behavior of clicking on the advertising copy 622A in the user interface 620 shown in Figure 6 (B) and browsing the corresponding advertising content. The static characteristics of the first user include age 30, gender female, and user portrait of a fashion expert. The advertising content of "Brand A Sun Hat" sent by the network device 300 and displayed on the electronic device 100 can be found in content 632 in the user interface 630 shown in Figure 6 (C), and the advertising copy 632C in the content 632 is the second copy.

A specific example of the method shown in Figure 8 can also be found in Figure 7, wherein the historical behavior data of the first user includes: the user's behavior of clicking on the advertising copy 713A in the user interface 710 shown in Figure 7 (A) and browsing the corresponding advertising content, and the user's behavior of clicking on the advertising copy 722A in the user interface 720 shown in Figure 7 (B) and browsing the corresponding advertising content. The static characteristics of the first user include age of 50, gender of female, user portrait of practicality and cost-effectiveness. The advertising content of "Brand A Sun Hat" sent by the network device 300 and displayed on the electronic device 100 can be found in content 732 in the user interface 730 shown in Figure 7 (C), and the advertising copy 732C in the content 732 is the second copy.

In one implementation, S204 - S210 are online operations, that is, operations performed when an advertisement is recommended to the user online after the user performs a trigger operation in real time.

Not limited to the implementation shown in Figure 8, in another implementation, the static features of the first user may not be obtained, and the network device 300 uses the long-term interest model to process the historical behavior data of the first user and the text in the advertising information of the second advertisement to obtain the advertising copy of the second advertisement, which does not include processing the static features of the first user.

In the method shown in FIG8 , after receiving the operation of the online first user, the network device 300 can generate the second copy online based on the text in the advertising information of the second advertisement in combination with the static features and historical behavior data of the online user. The above-mentioned static features and historical behavior data can reflect the online user's preference for the copy type/style, rather than simply classifying different users and recommending different advertising copies to users of different categories for the same advertisement (recommending the same advertising copy to users of the same category). Classification cannot reflect the different preferences of the same user at different time periods in the same scenario, nor can it reflect the different preferences of the same user in different scenarios. The second copy obtained in the present application can meet the user's preference for the copy type/style in these cases, and can also meet the different preferences of different users in the same scenario or different scenarios, thereby improving the efficiency of advertising delivery and user experience.

Please refer to Figure 9, which is a flow chart of a model training process provided by the present application. The method may be a specific implementation process of S201 shown in Figure 8. The method may include but is not limited to the following steps:

S301: The network device 300 obtains static characteristics of the user.

In one embodiment, the network device 300 can perform offline mining and cleaning on the data accumulated by the advertising platform to obtain static features of multiple users. The static features may include but are not limited to age, gender, city, type of electronic device used, and user portrait (such as obtained through analysis by the advertising platform).

S302: The network device 300 obtains historical behavior data of the user.

In one implementation, the network device may mine the historical behavior data of the above-mentioned multiple users, for example, mining the historical behavior data within a preset time period (for example, 30 days).

The order of S301 and S302 is not limited.

In one embodiment, after obtaining the static features and historical behavior data of at least one user in S301 and S302, the model for generating advertising copy can be trained based on these data, that is, the following S303-S311 is executed. Among them, these data can be used to construct copy generation signals and copy control signals. The copy generation signal is a direct input that the generation result depends on, and the copy control signal is an indirect input model that affects the generation result. The following S303-S304 can be used to construct the copy control signal, and the following 305 can be used to construct the copy generation signal.

For the convenience of explanation, the following S303-S311 takes the static characteristics and historical behavior data of a user (assuming it is user 3) as an example to illustrate the model. Type training process.

S303: The network device 300 constructs an advertisement copy sample (including a next hop sample and a corresponding historical sample) based on the user's historical behavior data.

In one embodiment, the network device 300 can randomly extract an advertisement copy of an advertisement 3 operated by user 3 from the historical behavior data of user 3 as the next-hop sample, and set at least one advertisement copy operated by user 3 before operating the next-hop sample as the corresponding historical sample (one advertisement copy is one historical sample), which can be understood as constructing a data pair of <historical sample, next-hop sample> (i.e., an advertisement copy sample).

S304: The network device 300 encodes the user's static features and the advertisement text sample to obtain a text control signal.

In one embodiment, the network device 300 can encode the static features of user 3 and obtain the corresponding vector representation (e.g., an embedding matrix). The network device 300 can also encode each sample of the advertising copy sample of user 3 (i.e., the advertising copy operated by user 3) (e.g., using Transformers Encoder for sequence encoding) and obtain the corresponding vector representation (e.g., an embedding sequence corresponding to multiple samples). Then, the network device 300 can fuse the vector representation of the static features and the vector representation of the advertising copy sample (e.g., through Cross Attention) to obtain a new vector expression (which can be called a copy control signal). Optionally, the copy control signal can be sent to the decoder (Decoder) as the output of the Encoder for processing, and the Decoder is used to execute S310.

S305: The network device 300 encodes the offline advertisement text and the first task mark to obtain a first generated signal.

In one implementation, the first task tag may indicate a text generation task, and the offline advertising copy may include advertising copy of at least one advertisement generated offline (eg, advertisement 3 corresponding to the next hop sample). The present application does not limit the specific method of generating advertising copy offline.

In some examples, the network device 300 may use the Prompt paradigm to fuse the offline advertisement text and the first task tag, and then encode the fused content to obtain a first generated signal.

In one implementation, the network device 300 may also obtain a text generation signal other than the first generation signal, for example, executing S306-S307 to obtain a second generation signal, and/or executing S308-S309 to obtain a third generation signal.

S306: The network device 300 sets the next hop sample as a positive sample, and obtains a corresponding negative sample.

In one implementation, S306 is an optional step.

In one embodiment, the network device 300 may adopt an exposure sampling method, taking the next-hop sample operated by the user as a positive sample, and taking the advertisement copy that is exposed in the same time window as the next-hop sample but not operated by the user as a negative sample, so as to construct data for model training.

S307: The network device 300 encodes the positive sample/negative sample and the second task label to obtain a second generated signal.

In one implementation, S307 is an optional step.

In one embodiment, the second task tag indicates a predicted click task, that is, based on the historical record (i.e., historical sample) of the current operation advertisement copy, whether the user will click on the next jump sample. For example, the training target corresponding to the positive sample is a "positive sample" (characterized by the fact that the user will click on the positive sample based on the historical sample), and the training target corresponding to the negative sample is a "negative sample" (characterized by the fact that the user will not click on the negative sample based on the historical sample). The network device 300 can encode the positive sample and the second task tag, and/or encode the negative sample and the second task tag, and obtain a second generated signal. In some examples, the network device 300 can use the Prompt paradigm to fuse the positive sample/negative sample and the second task tag, and then encode the fused content to obtain a second generated signal.

S308: The network device 300 performs random masking on the next hop sample.

In one implementation, S308 is an optional step.

In one embodiment, the network device 300 can perform word-granular random masking on the advertisement text of the next-hop sample, such as using a special <MASK> symbol to replace some words/characters in the advertisement text, or adding a special <MASK> symbol to the advertisement text, and obtain the next-hop sample after random masking.

S309: The network device 300 encodes the randomly masked next hop sample and the third task tag to obtain a third generated signal.

In one implementation, S309 is an optional step.

In one embodiment, the third task mark indicates a text recovery task, that is, restoring/recovering the next hop sample after random masking. For example, the original text of the next hop sample is "AABAB", and the text of the next hop sample after random masking is "AA<MASK>AB". The training goal is to restore "AA<MASK>AB" to "AABAB", where A and B are both a result of text segmentation. In some examples, the network device 300 can use the Prompt paradigm to fuse the next hop sample after random masking and the third task mark, and then encode the fused content to obtain a third generated signal.

S310: The network device 300 decodes the input signal in a non-autoregressive manner and obtains a pre-trained model.

In one implementation, the input signal includes the text control signal obtained in S304 and the first generated signal obtained in S305, and optionally also includes the second generated signal obtained in S307, and optionally also includes the third generated signal obtained in S309. In some examples, the network device 300 can obtain the pre-trained model by using a non-autoregressive generation method based on the control generation, that is, the decoder uses a fully masked input (for example For example, each word/phrase is replaced by a <MASK> symbol) as a reserved slot, the goal of model training is to predict the words/phrases in these slots, and also to predict the special character <EOS>, which indicates the end position of the text generation. In some examples, the network device 300 can also adopt a Deep multi-layer supervised training method. Compared with the usual Decoder which is composed of multiple identical network layers stacked together and the supervisory signal is only back-propagated in the last network layer, the Deep multi-layer supervised training method adds a supervisory signal to the output of each network layer for back-propagation, effectively improving the non-autoregressive generation accuracy.

S311: The network device 300 performs fine-tuning training on the pre-trained model to obtain a long-term interest model.

In one embodiment, the network device 300 can use the crowd-testing platform to manually verify the advertising copy samples, correct the copy content, and obtain model training data. Then, the network device 300 can fine-tune the pre-selected model according to the model training data. Fine-tuning training is a general method in the field of natural language processing for adapting the pre-trained model to downstream tasks. Through fine-tuning, the capabilities of the original pre-trained model can be fully utilized, and the current model can be made more suitable for downstream tasks, thereby improving the reasoning accuracy of downstream tasks. The long-term interest model obtained by fine-tuning training can be stored on the network device 300 side for online service call reasoning, such as implementing the method shown in Figure 8.

Not limited to the above implementation, in another implementation, the input signal of S310 may be a second generation signal and a text control signal, and the pre-trained model obtained at this time may be used to implement the predicted click task indicated by the second task mark. In another implementation, the input signal of S310 may be a third generation signal and a text control signal, and the pre-trained model obtained at this time may be used to implement the text recovery task indicated by the third task mark.

Not limited to the implementation shown in FIG. 9 , in another implementation, the offline advertisement text in S305 may also be replaced by other text in the advertisement information, such as the advertisement description input by the advertiser.

Not limited to the implementation shown in FIG. 9 , in another implementation, the user's static features may not be obtained as data for training the long-term interest model.

In the process shown in FIG9 , the network device 300 can combine the user's static features and the user's historical behavior data to abstractly extract the user's preferences at the copywriting level, that is, abstract the user's historical copywriting style preferences through the pre-trained model, and the long-term interest model obtained according to the pre-trained model can be used to rewrite/generate advertising copy that conforms to the user's historical copywriting style preferences. In addition, the model training adopts a multi-layer supervision method to construct non-autoregressive generation, which not only improves the accuracy of the model, but also alleviates the pressure of online delay.

Scenario 3: User 1 first uses electronic device 100 to operate multiple advertising copies, and then searches for keyword 3. Electronic device 100 can display advertising copies related to user 1’s static features, keyword 3, and the above multiple advertising copies. For specific examples, see Figure 10 below.

Figure 10 takes the advertising scenario of the product "Brand B mobile phone" as an example to illustrate. It is assumed that the advertising information of "Brand B mobile phone" includes "Brand B mobile phone, full screen, in-screen fingerprint support, multiple color schemes, 90Hz refresh rate, 1440Hz high-frequency dimming", and the offline advertising copy is "Brand B mobile phone, supports in-screen fingerprint, 90Hz refresh rate, 1440Hz high-frequency dimming", and the static features of user 1 include: age 30, gender is female, user portrait is a fashion expert, and keyword 3 is "high-performance full-screen mobile phone".

In one embodiment, as shown in (A) of FIG. 10 , the electronic device 100 may display a user interface 1010 of a social media application, and the user interface 1010 may be a home page. The user interface 1010 may include a search box 1011, a search control 1012, and multiple published contents (e.g., contents 1013 and 1014). The search box 1011 may be used to input search keywords, and the search control 1012 may be used to trigger a search for published contents related to the search box 1011. Content 1014 may be used to display the content of advertisement h, and may include advertisement copy 1014A (i.e., “M brand watches have new colors, why don’t you take action yet”). The electronic device 100 may receive a user operation (e.g., a touch operation, such as a click operation) of user 1 on content 1014, and in response to the user operation, display content related to advertisement h (e.g., a purchase page of a product corresponding to advertisement h, not shown in FIG. 10 ). In the above process, user 1 clicks on advertisement copy 1014A of advertisement h and the electronic device 100 displays the corresponding advertisement content.

In one embodiment, as shown in (B) of FIG. 10 , the user interface 1020 of the social media application may include a search box 1011, a search control 1012, and multiple published contents (e.g., content 1021 and content 1022). The user interface 1020 is similar to the user interface 1010, except that the displayed published contents are different. Content 1022 may be used to display the content of advertisement i, and may include advertising copy 1022A (i.e., "New autumn bags, keep up with the trend, without losing individuality"). The electronic device 100 may receive a user operation (e.g., a touch operation, such as a click operation) of user 1 on content 1022, and in response to the user operation, display content related to advertisement i (e.g., a purchase page of a product corresponding to advertisement i, not shown in FIG. 10 ). In the above process, user 1 clicks on the advertising copy 1022A of advertisement i and the electronic device 100 displays the corresponding advertising content.

After the implementation shown in (A) and (B) of Figure 10, the electronic device 100 can receive the keyword 3 "high-performance full-screen mobile phone" input by the user in the search box 1011, and then receive the user operation on the search control 1012 (for example, a touch operation, and the touch operation is, for example, a click operation). In response to the user operation, the electronic device 100 can obtain and display the published content related to the keyword 3 (which may include advertising content). For specific examples, see the user interface 1040 shown in (C) of Figure 10.

As shown in (C) of FIG. 10 , the user interface 1040 of the social media application may include a search box 1011 and multiple published contents (e.g., content 1041 and content 1042). The user interface 1040 is similar to the user interface 1010, except that the displayed published contents are different, and the search box 1011 in the user interface 1040 displays the currently searched keyword 3 "high-performance full-screen mobile phone". Content 1041 may be "What is the full-screen mobile phone you are most looking forward to" published by user kkk. Content 1042 may be used to display the content of advertisement j (i.e., the advertisement of the product "B-brand mobile phone"), and may include an advertising manufacturer 1042A (i.e., "manufacturer jjj"), an advertising instruction 1042B (i.e., "advertisement"), an advertising copy 1042C (i.e., "incredibly powerful performance, rich color matching without losing individuality, advanced in-screen fingerprint recognition, this is the B-brand full-screen mobile phone, just one glance will make you excited, take action quickly") and an advertising video 1042D.

Among them, the advertisement copy 1042C of "Brand B mobile phone" in the user interface 1040 shown in (C) of FIG10 may be an advertisement copy related to the keyword searched by the user and the user's historical behavior, generated according to the keyword 3 searched by the user 3, the user 3's operation behavior on the advertisement copy 1014A of advertisement h and the advertisement copy 1022A of advertisement i. Compared with the offline advertisement copy of "Brand B mobile phone" "Brand B mobile phone, supports in-screen fingerprint, 90Hz refresh rate, 1440Hz high-frequency dimming", the advertisement copy 1042C includes: the copy fragments "incredibly powerful performance" and "full screen" related to the keyword 3 "high-performance full-screen mobile phone" searched by the user 3, and also includes: the copy fragments "rich color matching without losing personality" and "only one glance will make you excited, take action quickly" related to the static characteristics of the user 3, the advertisement copy 1014A of the user's historical operation, and the advertisement copy 1022A of the user's historical operation. Optionally, the copy snippets "Rich colors without losing personality" and "Just one glance will make you excited and take action quickly" can be obtained based on the results of long-term interest mining and long-term interest guidance. The result of long-term interest mining can be the labels obtained by identifying the static characteristics of the user and the advertising copy of the user's historical operations: fashion, personality, trend, new style, and the result of long-term interest guidance can be the copy snippet "multiple colors" found from the advertising information based on the results of long-term interest mining. In other words, this application can combine the user's historical behavior and the keywords searched by the user to generate/rewrite and recommend personalized advertising copy online, so that the style of the recommended advertising copy tends to the user's preference on the basis of explicitly meeting the user's current needs, thereby increasing the probability of the user operating the advertising copy to view the detailed advertising content and improving the efficiency of advertising delivery.

Please refer to Figure 11, which is a flowchart of another method for generating advertising copy provided by the present application. The method may include but is not limited to the following steps:

S401: The network device 300 obtains corresponding short-term interests based on user behavior data collected offline.

In one embodiment, S401 is an optional step. In one embodiment, S401 is an offline operation, that is, it is performed before the advertisement is delivered/displayed. In one embodiment, the user behavior data collected offline may include the user's search behavior (which may include search keywords), exposure behavior (for example, including advertisement text that is exposed but not clicked) and operation behavior/browsing behavior for advertisements and advertisement-related information in at least one advertisement delivery scenario. For the convenience of explanation, the user behavior data in the delivery scenario of an advertisement (assuming it is advertisement 4) is used as an example for explanation. In one embodiment, the user behavior data collected offline may include at least one keyword. Based on any one of the keywords, the network device 300 can extract the copy fragment related to the keyword from the advertisement information of advertisement 4, that is, the short-term interest in advertisement 4. This process is similar to S106-S110 of Figure 5 and will not be repeated.

S402: The network device 300 obtains corresponding long-term interest training data based on the user behavior data collected offline.

In one embodiment, S402 is an optional step. In one embodiment, S402 is an offline operation, that is, it is performed before advertisement delivery/display.

In one embodiment, the user behavior data collected offline in S402 is the same as the user behavior data collected offline in S401. For this set of user behavior data, the network device 300 can simulate the process of online users requesting advertising content, and construct the user's static characteristics and historical behavior data in the delivery scenario of advertisement 4, that is, training data of long-term interests.

The order of S401 and S402 is not limited.

S403: The network device 300 trains and obtains a short- and long-term interest model based on the training data of the short-term interest and the long-term interest.

In one embodiment, S403 is an optional step. In one embodiment, S403 is an offline operation, that is, it is performed before advertisement delivery/display.

In one embodiment, the network device 300 can train the long-term and short-term interest models based on the training data of the short-term interest obtained in S401 and the long-term interest obtained in S402. The specific training process is similar to the process shown in Figure 9, except that in S305, the short-term interest obtained in S401 will also be used. In some examples, the network device 300 can use the Prompt paradigm to fuse the short-term interest obtained in S401, the offline advertising copy of advertisement 4, and the first task marker, and then encode the fused content to obtain the corresponding copy generation signal.

S404: The electronic device 100 uploads the historical behavior data and static features of the first user to the network device 300.

S405: The network device 300 obtains historical behavior data and static features of the first user.

S404-S405 are similar to S202-S203 in FIG. 8 , and please refer to the description of S202-S203 in FIG. 8 for details.

The order of any one of S404-S405 and S401-S403 is not limited.

S406: The electronic device 100 obtains the second keyword input by the first user.

In one embodiment, the electronic device 100 receives the second keyword and the corresponding search request input by the first user, and in response to the search request, obtains the second keyword, and obtains search results related to the second keyword, for example, executing S407 to obtain related advertisements. In some examples, in the scenario shown in FIG10 , the electronic device 100 can receive the keyword 3 "high-performance full-screen mobile phone" input by the user in the search box 1011 of the user interface 1030 shown in (C) of FIG10 , and then receives the user operation on the search control 1012 (which can be understood as receiving the search request), and in response to the user operation, obtains the keyword 3, and obtains the published content (including advertising content) related to the keyword 3.

S407: The electronic device 100 sends a third advertisement request (carrying the second keyword and the ID of the first user) to the network device 300.

In one embodiment, the third advertisement request may be used to request to obtain advertisement content. The third advertisement request may carry the second keyword and the ID of the first user, the second keyword may be used to obtain advertisement content related to the second keyword, and the ID of the first user may be used by the network device 300 to determine the static features and historical behavior data of the corresponding first user.

S408: The network device 300 determines a third advertisement to be displayed, and obtains advertisement information.

In one implementation, the network device 300 may use the second keyword to recall and sort from multiple advertisements to obtain a third advertisement related to the second keyword. In some examples, the network device 300 may obtain the advertisement information of the third advertisement according to the ID of the third advertisement. The advertisement information may include the text of the advertisement, as well as non-text content such as pictures, videos, and links. The text of the advertisement may include multiple sentences.

S409: The network device 300 obtains the first user's short-term interest in the third advertisement.

In one implementation, the network device 300 can extract text segments related to the second keyword from the advertising information of the third advertisement, that is, the first user's short-term interest in the third advertisement. This process is similar to S106-S110 of FIG. 5 and will not be described in detail.

S410: The network device 300 obtains historical behavior data and static features of the first user according to the ID of the first user.

S410 is similar to S207 in FIG. 8 , and details may refer to the description of S207 in FIG. 8 .

The order of S410, S408 and S409 is not limited.

S411: The network device 300 uses the long-term and short-term interest model to process the historical behavior data of the first user, the static characteristics of the first user, the short-term interests of the first user, and the advertisement information of the third advertisement to obtain a third copy of the third advertisement.

In one embodiment, the network device 300 may use the historical behavior data of the first user, the static characteristics of the first user, the short-term interest of the first user in the second advertisement, and the text in the advertisement information of the third advertisement as inputs of the long-short interest model, and obtain the output: the third copy of the third advertisement. For the description of the text in the advertisement information, please refer to the description of S208 of FIG. 8. Optionally, the network device 300 may modify/rewrite the offline advertisement copy based on the historical behavior data of the first user, the static characteristics of the first user, and the short-term interest of the first user in the second advertisement to obtain the third copy of the third advertisement.

S412 : The network device 300 sends the third text to the electronic device 100 .

S413: The electronic device 100 displays the third text.

In one implementation, after sending the third advertisement request to the network device 300 , the electronic device 100 may receive the advertisement text (ie, the third text) generated by the network device 300 based on the third advertisement request, and display the third text.

A specific example of the method shown in Figure 11 can be found in Figure 10, wherein the second keyword is keyword 3 "high-performance full-screen mobile phone", and the historical behavior data of the first user includes: the user's behavior of clicking on the advertising copy 1014A in the user interface 1010 shown in Figure 10 (A) and browsing the corresponding advertising content, and the user's behavior of clicking on the advertising copy 1022A in the user interface 1020 shown in Figure 10 (B) and browsing the corresponding advertising content. The static characteristics of the first user include age 30, gender female, and user portrait of a fashion expert. The advertising content of "Brand B mobile phone" sent by the network device 300 and displayed on the electronic device 100 can be found in content 1042 in the user interface 1040 shown in Figure 10 (C), and the advertising copy 1042C in the content 1042 is the third copy.

In one implementation, S406-S413 are online operations, that is, operations performed when an advertisement is recommended to the user online after the user performs a trigger operation in real time.

Not limited to the implementation shown in Figure 11, in another implementation, the static features of the first user may not be obtained, and the network device 300 uses the long-term and short-term interest models to process the historical behavior data of the first user, the first user's short-term interest in the third advertisement, and the text in the advertising information of the second advertisement to obtain the advertising copy of the third advertisement, and does not include processing the static features of the first user.

In the method shown in FIG. 11 , the network device 300 can generate a third copy online based on the text in the advertising information of the third advertisement based on the keywords searched online by the user, the static features of the online user, and the historical behavior data. The third advertisement can not only meet the current needs of the user (which can be reflected by keywords), but also meet the online user's preference for the copy type/style (which can be reflected by historical behavior data), effectively improving the probability of the user operating the advertising copy to view detailed advertising content, improving the efficiency and effectiveness of advertising delivery, and also improving the user experience.

Scenario 4: Advertiser 1 can input configuration content for placing advertisement 1 on electronic device 400. For specific examples, see FIG. 12A and FIG. 12B, the electronic device 400 can upload the received configuration content to the network device 300, so that the network device 300 can implement the delivery of advertisement 1 according to the configuration content.

FIG. 12A is a schematic diagram showing an exemplary advertisement configuration interface.

As shown in FIG. 12A , the electronic device 400 may display a user interface 1210. The user interface 1210 may be used by advertisers to set options such as advertisement name, advertisement theme, advertisement classification, advertisement label, advertisement content summary, advertisement uniform resource locator (URL) address, advertisement description, and online personalized copy. That is, the user interface 1210 is an advertisement configuration interface. Among them, the advertisement classification option is shown in FIG. 12A , for example, including one or more of the options such as “news”, “technology”, “society”, “entertainment”, “education”, “finance”, “sports”, “real estate”, “game”, “business”, and “car”. The advertisement label option is shown in FIG. 12A , for example, including one or more of the options such as “region”, “user’s gender”, “user’s age”, “device used by the user”, and “network connection method”. The advertisement URL address may be the address of the landing page of the advertisement, which may be used for crawlers to obtain advertisement information. The advertisement description option may be used to input a text description of the advertisement. The online personalized copy option may include an on option and an off option for the online personalized copy function. The online personalized copy function is used to generate advertisement copy according to the keywords searched by the user and/or the user’s historical behavior data, and deliver the advertisement copy. In some examples, assuming that an advertiser places an advertisement for promoting the product "Brand A sun hat" to the network device 300, the advertiser can make the settings shown in Figure 12A on the advertisement configuration interface of the electronic device 100. As shown in Figure 12A, the advertiser can edit the following according to the actual situation of the product "Brand A sun hat": advertisement name such as "Advertisement ID of Brand A sun hat", advertisement theme such as "Practical and trendy", advertisement classification such as "Society", advertisement content summary such as "Made of high-quality fabrics", advertisement description such as "Super practical summer sun hat, made of high-quality fabrics, can effectively prevent ultraviolet invasion, and give you a refreshing summer. The sun hat that must be chosen by trendsetters has novel styles and rich colors to show your handsomeness". The advertiser can also set labels such as region (such as unlimited region), gender (such as unlimited gender), age (such as unlimited age), and the device used (such as unlimited device) on the advertisement configuration interface according to the main audience group and/or target delivery group of the product "Brand A sun hat". Advertisers can also edit the URL address of the advertising content on the advertising server, such as "https://ad_platform_server/ad/video1.mp4". Advertisers can also operate the opening option of the online personalized copy option to turn on the online personalized copy function. In some examples, after the electronic device 400 uploads the configuration content shown in Figure 12A set by the advertiser (i.e., the configuration content of the above example) to the network device 300, the network device 300 can obtain the advertising information of "Brand A Sun Hat" from multiple dimensions (such as advertisement name, advertisement theme, advertisement content summary, advertisement URL address, advertisement description) for advertising delivery. In addition, since the advertiser has turned on the online personalized copy function, the network device 300 will generate and recommend personalized "Brand A Sun Hat" advertising copy online based on the keywords searched by the user and/or the user's historical behavior data. The specific process can be seen in the advertising delivery method shown in Figures 5, 8, and 11.

In other embodiments, the advertiser may not enable the online personalized copy function, for example, the advertiser operates the closing option of the online personalized copy option in the user interface 1210 shown in FIG12A, or the online personalized copy function is closed by default, and the advertiser does not operate the opening option of the online personalized copy option in the user interface 1210 shown in FIG12A. In this case, the network device may recommend offline generated good advertisement copy for the main audience group and/or the target delivery group of the current advertisement, such as at least one sentence in the text such as the advertisement name, advertisement theme, advertisement category, advertisement content summary, advertisement content summary, advertisement description, etc. in the user interface 1210 shown in FIG12A. The network device may also recommend non-text advertisement content, such as pictures, videos, links (such as advertisement URL addresses), etc. uploaded by the user based on the user interface 1210 shown in FIG12A.

FIG12B exemplarily shows a schematic diagram of another advertisement configuration interface. As shown in FIG12B , the electronic device 400 may display a user interface 1220. The user interface 1220 is similar to the user interface 1210 shown in FIG12A , except that the user interface 1220 may also be used by the advertiser to set the probability of online personalized copy delivery. In some examples, assuming that the advertiser delivers an advertisement for promoting the product “Brand A sun hat” to the network device 300, the advertiser may perform the settings shown in FIG12B on the advertisement configuration interface of the electronic device 100. After the advertiser turns on the online personalized copy function by clicking the “On” option of the online personalized copy option, the advertiser may also set the probability of online personalized copy delivery. The probability may indicate the proportion of the number of personalized advertisements for “Brand A sun hat” delivered to the total number of advertisements for “Brand A sun hat” delivered. The personalized advertisements are determined based on the keywords searched by the user and/or the user’s historical behavior data. According to the generated advertising copy, the value of the probability is (0%, 100%] (when the probability is 100%, the advertising copy representing the “Brand A Sun Hat” delivered is all personalized advertising copy), and the advertiser sets the probability to “30%”, for example. In some examples, after the electronic device 400 uploads the configuration content shown in FIG. 12B set by the advertiser (i.e., the configuration content of the above example) to the network device 300, the network device 300 can determine whether to deliver personalized advertising copy each time it recommends the advertising copy of “Brand A Sun Hat” to online users. According to the probability of online personalized copy delivery.

Please refer to Figure 13, which is a flowchart of an advertisement delivery method provided by the present application. The method may include but is not limited to the following steps:

S501: The network device 300 obtains configuration content for placing a fourth advertisement input by a first advertiser.

In one embodiment, after receiving the configuration content of placing the fourth advertisement input by the first advertiser, the electronic device of the first advertiser can send the configuration content to the network device 300, and the electronic device can provide an advertisement configuration interface for the first advertiser to place the fourth advertisement. In another embodiment, the network device 300 may receive the configuration content of the fourth advertisement delivered by the first advertiser, and the network device 300 may provide an advertisement configuration interface for the first advertiser to configure the fourth advertisement delivery. Examples of the advertisement configuration interface may be seen in FIG. 12A and FIG. 12B.

In one embodiment, the configuration content for delivering the fourth advertisement may include first indication information of whether the online personalized copy function is turned on, and optionally, a delivery probability of the online personalized copy. In one case, the delivery probability of the online personalized copy is obtained by the electronic device of the first advertiser according to a received user operation of the first advertiser, that is, manually input by the first advertiser. In another case, if the electronic device of the first advertiser does not receive the delivery probability manually input by the first advertiser, the electronic device may use a preset probability as the delivery probability of the online personalized copy.

S502: When the fourth advertisement is delivered online, the network device 300 determines whether to generate personalized advertisement text according to the configuration content of the fourth advertisement.

In one embodiment, after receiving the user's trigger operation, the electronic device used by the user may send an advertisement request to the network device 300 to obtain the advertisement content. For specific examples, see S104-S105 of FIG. 5, S204-S205 of FIG. 8, and S406-S407 of FIG. 11. After receiving the advertisement request, if the network device 300 determines that the advertisement to be displayed is the fourth advertisement, it may determine whether to generate personalized advertisement copy for the fourth advertisement according to the configuration content of the fourth advertisement. For examples of the network device 300 determining the advertisement to be displayed, see S107 of FIG. 5, S206 of FIG. 8, and S408 of FIG. 11.

In one embodiment, the first indication information in the configuration content for placing the fourth advertisement indicates that the online personalized copy function has been turned on, and the configuration content includes the delivery probability of the online personalized copy. The network device 300 can determine whether to generate personalized advertisement copy for the current fourth advertisement according to the delivery probability. If the result is yes, the network device 300 executes S503, and if the result is no, the network device 300 executes S504. Not limited to this, in another embodiment, the configuration content for placing the fourth advertisement includes the first indication information indicating that the online personalized copy function has been turned on, but does not include the delivery probability of the online personalized copy. The network device 300 can use the preset probability as the delivery probability of the online personalized copy.

S503: The network device 300 obtains the personalized advertisement text of the fourth advertisement.

In one implementation, the network device 300 may generate personalized advertisement text for the fourth advertisement, for which specific examples may be found in S106-S111 of FIG. 5 , S206-S208 of FIG. 8 , and S408-S411 of FIG. 11 . Then, the network device 300 may send the advertisement text to the user's electronic device for display, for which specific examples may be found in S112-S113 of FIG. 5 , S209-S210 of FIG. 8 , and S412-S413 of FIG. 11 .

S504: The network device 300 obtains an offline advertisement text of the fourth advertisement.

In one embodiment, the offline advertising copy of the fourth advertisement is generated in advance/offline. This application does not limit the specific generation method. The network device 300 can obtain the offline advertising copy of the fourth advertisement, for example, from a memory, and send the advertising copy to the user's electronic device for display. For example, the electronic device can display the text and/or non-text content such as pictures, videos, links, etc. uploaded by the advertiser in the advertising configuration interface shown in Figures 12A and 12B.

In the method shown in FIG. 13 , advertisers can set the configuration content of the advertisement delivery by themselves, can choose whether to turn on or off the online personalized copy function, and can set the delivery probability of the online personalized copy. Subsequently, they can confirm the improvement of the advertisement delivery efficiency by the online personalized copy function based on the real feedback from online users, confirm its effectiveness and practicality, and enhance the advertiser's experience.

In the methods provided in the embodiments of the present application, all or part of them can be implemented by software, hardware, firmware or any combination thereof. When implemented using software, all or part of them can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user device or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated therein. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a digital video disc (DWD), or a semiconductor medium (e.g., a solid state drive (SSD)). As mentioned above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application is described in detail with reference to the above embodiments, a person skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

A method for generating advertisement copy, characterized in that it is applied to a network device, and the method comprises:

Acquire first advertisement information of a first object, wherein the first advertisement information includes a first copy;

Obtaining a first advertisement request sent by a first device, and obtaining a first keyword and a second keyword from the first advertisement request;

determining the first advertisement information according to the first keyword;

Determine a second copy according to the first keyword, the second keyword, and the first copy;

A first response message is sent to the first device, where the first response message includes the second text, and the first response message is used to instruct the first device to display the second text.
The method according to claim 1, characterized in that the method further comprises:

Obtaining a second advertisement request sent by the first device, and obtaining the first keyword and the third keyword from the second advertisement request;

determining the first advertisement information according to the first keyword;

Determine a third copy according to the first keyword, the third keyword and the first copy;

Sending a second response message to the first device, where the second response message includes the third text, and the second response message is used to instruct the first device to display the third text.
The method according to claim 1 or 2 is characterized in that the first advertising information includes at least one of the following: a first picture, a first video, and a first link of the first object, and the first response information also includes at least one of the following: the first picture, the first video, and the first link.
The method according to any one of claims 1 to 3, characterized in that the second copy includes a first sentence, the first copy includes the first sentence, and the first sentence is associated with the second keyword.
The method according to any one of claims 1 to 4, characterized in that the second copy includes a second sentence, the first copy does not include the second sentence, and the second sentence is associated with the second keyword.
The method according to any one of claims 1 to 5, characterized in that the first advertisement request includes an identifier of a first user; the method further comprises: acquiring feature information and/or historical behavior data of the first user according to the identifier of the first user;

Determining the second copy based on the first keyword, the second keyword and the first copy includes: determining the second copy based on the first keyword, the second keyword, the first copy, and characteristic information and/or historical behavior data of the first user.
The method as claimed in claim 6 is characterized in that before obtaining the first advertisement request sent by the first device, the method further includes: obtaining characteristic information and/or historical behavior data related to the identifier of the first user.
The method according to claim 6 or 7 is characterized in that the historical behavior data of the first user includes a fourth copy of the second object clicked by the first user, the fourth copy includes a third sentence, and the second copy includes the third sentence.
The method according to any one of claims 6 to 8 is characterized in that the characteristic information of the first user includes at least one of the following: age, gender, location, type of electronic device, and user portrait, and the historical behavior data of the first user includes at least one of the following: clicking on advertisements, browsing advertisements, closing advertisements, purchasing the corresponding object through a link in an advertisement, downloading an application through a link in an advertisement, and the duration of browsing advertisements.
The method according to any one of claims 1 to 9, characterized in that the first advertisement information includes first indication information, the first indication information indicating whether the first function is in an on state or in an off state; the determining the first advertisement information according to the first keyword comprises: judging whether the first indication information indicates that the first function is in an on state; when the first indication information indicates that the first function is in an on state, determining the first advertisement information according to the first keyword;

The method further includes: when the first indication information indicates that the first function is in a closed state, sending third response information to the first device, the third response information including the first text, and the third response information is used to instruct the first device to display the first text. case.
A method for generating advertisement copy, characterized in that it is applied to a network device, and the method comprises:

Acquire first advertisement information of a first object, wherein the first advertisement information includes a first copy;

Acquire a first advertisement request sent by a first device, where the first advertisement request includes an identifier of a first user;

Acquiring characteristic information and/or historical behavior data of the first user according to the identifier of the first user;

Determine a second copy based on the first copy, characteristic information and/or historical behavior data of the first user;

A first response message is sent to the first device, where the first response message includes the second text, and the first response message is used to instruct the first device to display the second text.
The method according to claim 11, characterized in that the method further comprises:

Acquire a second advertisement request sent by the second device, where the second advertisement request includes an identifier of a second user;

Acquiring characteristic information and/or historical behavior data of the second user according to the identifier of the second user;

Determine a third copy based on the first copy, characteristic information and/or historical behavior data of the second user;

Sending second response information to the second device, where the second response information includes the third text, and the third response information is used to instruct the second device to display the third text.
The method as claimed in claim 11 or 12 is characterized in that the first advertising information includes at least one of the following: a first picture, a first video, and a first link of the first object, and the first response information also includes at least one of the following: the first picture, the first video, and the first link.
The method according to any one of claims 11-13 is characterized in that, before obtaining the first advertisement request sent by the first device, the method further includes: obtaining characteristic information and/or historical behavior data related to the identifier of the first user.
The method according to any one of claims 11-14 is characterized in that the historical behavior data of the first user includes a fourth copy of a second object clicked by the first user, the fourth copy includes a first sentence, and the second copy includes the first sentence.
The method according to any one of claims 11-15 is characterized in that the characteristic information of the first user includes at least one of the following: age, gender, location, type of electronic device, and user portrait, and the historical behavior data of the first user includes at least one of the following: clicking on advertisements, browsing advertisements, closing advertisements, purchasing the corresponding object through a link in an advertisement, downloading an application through a link in an advertisement, and the length of time of browsing advertisements.
The method according to any one of claims 11 to 16, characterized in that the first advertisement information includes first indication information, the first indication information indicating whether the first function is in an on state or in an off state; the obtaining of the characteristic information and/or historical behavior data of the first user according to the identifier of the first user comprises: determining whether the first indication information indicates that the first function is in an on state; when the first indication information indicates that the first function is in an on state, obtaining the characteristic information and/or historical behavior data of the first user according to the identifier of the first user;

The method further includes: when the first indication information indicates that the first function is in a closed state, sending third response information to the first device, the third response information including the first text, and the third response information is used to instruct the first device to display the first text.
A network device, characterized in that it includes a transceiver, a processor and a memory, the memory is used to store a computer program, and the processor calls the computer program to execute the method described in any one of claims 1-10 or 11-16.
A computer storage medium, characterized in that the computer storage medium stores a computer program, and when the computer program is executed by a processor, it implements the method described in any one of claims 1-10 or 11-16.