CN114547417A

CN114547417A - Media resource ordering method and electronic equipment

Info

Publication number: CN114547417A
Application number: CN202210179843.7A
Authority: CN
Inventors: 戴文恺; 杨钊; 王慕天; 马丽芬; 张瑞昌; 白云龙; 韩艳
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2022-02-25
Filing date: 2022-02-25
Publication date: 2022-05-27

Abstract

The disclosure provides a media resource ordering method and electronic equipment, relates to an artificial intelligence technology, and particularly relates to the field of big data. The specific scheme is as follows: acquiring the association characteristics of a first user; inputting the associated features into a pre-acquired target click rate prediction model for prediction to obtain a first predicted click rate vector, and inputting the associated features into a pre-acquired target multitask learning model for prediction to obtain a first predicted stay time vector and a first predicted completion rate vector; for each candidate media resource, obtaining a second predicted click rate vector, a second predicted stay time vector and a second predicted completion rate vector, and determining a first score of the candidate media resource based on the first predicted click rate vector, the second predicted click rate vector, the first predicted stay time vector, the second predicted stay time vector, the first predicted completion rate vector and the second predicted completion rate vector; the first scoring accuracy can be improved, and therefore resource sequencing is carried out, and the resource sequencing accuracy is improved.

Description

Media resource ordering method and electronic equipment

Technical Field

The disclosure relates to the technical field of artificial intelligence such as big data, and particularly relates to a media resource sorting method and electronic equipment.

Background

With the development of the internet, media resources are recommended to users more and more frequently, in the recommendation process, media resources which are interested in the users need to be recommended to the users in massive resources, and the media resource sequencing is an important part in the recommendation process and directly influences the recommendation effect.

Currently, in the process of ranking media resources, a common method is to rank the media resources by using historical behavior data of a user to be recommended.

Disclosure of Invention

The disclosure provides a media resource sorting method and electronic equipment.

In a first aspect, an embodiment of the present disclosure provides a media resource sorting method, where the method includes:

acquiring the association characteristics of a first user;

inputting the correlation characteristics into a pre-obtained target click rate prediction model for prediction to obtain a first predicted click rate vector, and inputting the correlation characteristics into a pre-obtained target multitask learning model for prediction to obtain a first predicted stay time vector and a first predicted completion rate vector;

aiming at each candidate media resource of a plurality of candidate media resources, obtaining a second predicted click rate vector, a second predicted stay time vector and a second predicted completion rate vector of the candidate media resource; and determining a first score for the candidate media resource based on the first predicted click rate vector, the second predicted click rate vector, the first predicted length of dwell vector, the second predicted length of dwell vector, the first predicted completion rate vector, and the second predicted completion rate vector;

ranking the plurality of candidate media resources based on the first scores of the plurality of candidate media resources.

In this embodiment, in the process of determining the first score of the candidate media resource, the associated feature of the first user is used, and the predicted click rate vector predicted by the target click rate prediction model based on the associated feature of the first user and the first predicted staying time vector and the first predicted completing rate vector predicted by the target multitask learning model based on the associated feature of the first user are taken into consideration, the second predicted click rate vector, the second predicted staying time vector and the second predicted completing rate vector of the candidate media resource are used, and the first score of the candidate media resource is determined based on the first predicted click rate vector, the second predicted click rate vector and the first predicted staying time vector, so that the accuracy of the first score of the candidate media resource can be improved, and the ranking of the candidate media resources is realized through the first score of the candidate media resources, thereby improving the accuracy of the ranking of the candidate media assets.

In a second aspect, an embodiment of the present disclosure provides a click rate prediction model training method, where the method includes:

screening a plurality of users according to attribute characteristics of the users in the application to determine a first user set;

screening the first user set to obtain a second user set, wherein any user in the second user set has a first behavior of associating media resources with a first project of the application;

screening a first media resource set with a preset identification from a plurality of media resources of the application, wherein the preset identification is used for indicating that the media resources can be published and distributed in a second item;

screening the first media resource set to obtain a second media resource set, wherein any media resource in the second media resource set has associated user behavior in the first project of the application;

constructing a first sample data set based on the association characteristics of the second user set, the association characteristics of the second item association user of the application, the resource characteristics of the second media resource set and the resource characteristics of a third media resource set, wherein the third media resource set comprises media resources of a second behavior performed by the second item association user of the application;

and training an initial click rate prediction model by using the first sample data set to obtain a target click rate prediction model.

In this embodiment, a first sample data set may be constructed through the association features of the second user set, the association features of the applied second item association users, the resource features of the second media resource set, and the resource features of the third media resource set, and then the initial click rate prediction model may be trained by using the first sample data set to obtain a target click rate prediction model, so as to improve the performance of the target click rate prediction model.

In a third aspect, an embodiment of the present disclosure further provides a method for training a multi-task learning model, including:

screening a first media resource set with a preset identification from a plurality of media resources of the application, wherein the preset identification is used for indicating that the media resources can be distributed in a second item;

constructing a second sample data set based on the associated features of the second user set, the associated features of the second item associated user of the application, the resource features of the second media resource set and the resource features of a third media resource set, wherein the third media resource set comprises media resources for performing a second behavior by the second item associated user of the application;

and training the initial multi-task learning model by using the second sample data set to obtain a target multi-task learning model.

In this embodiment, a second sample data set may be constructed through the association features of the second user set, the association features of the applied second item association user, the resource features of the second media resource set, and the resource features of the third media resource set, and the initial multi-task learning model may be trained subsequently using the second sample data set to obtain the target multi-task learning model, so as to improve the performance of the target multi-task learning model.

In a fourth aspect, an embodiment of the present disclosure provides a media resource sorting apparatus, including:

the first acquisition module is used for acquiring the association characteristics of the first user;

the prediction module is used for inputting the associated features into a pre-acquired target click rate prediction model for prediction to obtain a first predicted click rate vector, and inputting the associated features into a pre-acquired target multitask learning model for prediction to obtain a first predicted stay time vector and a first predicted completion rate vector;

the score determining module is used for acquiring a second predicted click rate vector, a second predicted stay time vector and a second predicted completion rate vector of the candidate media resources aiming at each candidate media resource of the candidate media resources; and determining a first score for the candidate media resource based on the first predicted click rate vector, the second predicted click rate vector, the first predicted length of dwell vector, the second predicted length of dwell vector, the first predicted completion rate vector, and the second predicted completion rate vector;

a ranking module to rank the plurality of candidate media resources based on the first scores of the plurality of candidate media resources.

In a fifth aspect, an embodiment of the present disclosure provides a click rate prediction model training apparatus, including:

the first user determining module is used for screening the users according to the attribute characteristics of the users in the application to determine a first user set;

a second user determination module, configured to filter the first user set to obtain a second user set, where any user in the second user set has a first behavior of associating a media resource with the first project of the application;

the first resource screening module is used for screening a first media resource set with a preset identifier from a plurality of media resources of the application, wherein the preset identifier is used for indicating that the media resources can be distributed in a second item in a publishing way;

a second resource screening module, configured to screen the first media resource set to obtain a second media resource set, where any media resource in the second media resource set has an associated user behavior in the first item of the application;

the first sample structure modeling block is used for constructing a first sample data set based on the association characteristics of the second user set, the association characteristics of the second item association user of the application, the resource characteristics of the second media resource set and the resource characteristics of a third media resource set, wherein the third media resource set comprises media resources for performing a second action on the second item association user of the application;

and the first training module is used for training the initial click rate prediction model by utilizing the first sample data set to obtain a target click rate prediction model.

In a sixth aspect, an embodiment of the present disclosure provides a multitask learning model training device, including:

the third user determining module is used for screening the plurality of users according to the attribute characteristics of the plurality of users in the application to determine a first user set;

a fourth user determination module, configured to filter the first user set to obtain a second user set, where any user in the second user set has a first behavior on a media resource associated with the first project of the application;

a third resource screening module, configured to screen a first media resource set having a preset identifier from the multiple media resources of the application, where the preset identifier is used to indicate that a media resource may be distributed in a second item;

a fourth resource screening module, configured to screen the first media resource set to obtain a second media resource set, where any media resource in the second media resource set has an associated user behavior in the first item of the application;

a second sample construction module, configured to construct a second sample data set based on the association characteristics of the second user set, the association characteristics of the second item-associated user of the application, the resource characteristics of the second media resource set, and the resource characteristics of a third media resource set, where the third media resource set includes media resources for performing a second behavior by the second item-associated user of the application;

and the second training module is used for training the initial multi-task learning model by utilizing the second sample data set to obtain a target multi-task learning model.

In a seventh aspect, an embodiment of the present disclosure further provides an electronic device, including:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a media asset ranking method as provided by the first aspect of the disclosure or a click-through rate prediction model training method as provided by the second aspect or a multi-task learning model training method as provided by the third aspect.

In an eighth aspect, an embodiment of the present disclosure further provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method for ranking media assets as provided in the first aspect of the present disclosure or the method for training a click-through rate prediction model as provided in the second aspect or the method for training a multi-task learning model as provided in the third aspect of the present disclosure.

In a ninth aspect, an embodiment of the present disclosure provides a computer program product comprising a computer program which, when executed by a processor, implements the media resource ranking method provided by the first aspect of the present disclosure or the click-through rate prediction model training method provided by the second aspect or the multi-task learning model training method provided by the third aspect of the present disclosure.

It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present disclosure, nor do they limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description.

Drawings

The drawings are included to provide a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

fig. 1 is a schematic flow chart diagram of a media resource ranking method according to an embodiment provided in the present disclosure;

FIG. 2 is a schematic flow chart diagram illustrating a click-through rate prediction model training method according to an embodiment of the present disclosure;

FIG. 3 is a schematic flow chart diagram of a method for multi-task learning model training according to an embodiment provided by the present disclosure;

FIG. 4 is a block diagram of a media asset ranking device of one embodiment provided by the present disclosure;

FIG. 5 is a block diagram of a click-through rate prediction model training device according to an embodiment of the present disclosure;

FIG. 6 is a block diagram of a multitask learning model training device according to one embodiment of the present disclosure;

FIG. 7 is a block diagram of an electronic device for implementing a media resource ranking method, a click-through rate prediction model training method, or a multi-task learning model training method of embodiments of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, in which various details of the embodiments of the disclosure are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

As shown in fig. 1, according to an embodiment of the present disclosure, the present disclosure provides a media resource ranking method, including:

step S101: and acquiring the association characteristics of the first user.

The first user may be understood as a user to be recommended, and after the plurality of candidate resources are ranked, media resource recommendation may be performed to the first user, and the first user may be a user in a certain application. As an example, the associated features of the user may include, but are not limited to, attribute features, current scene features, historical behavior features, and the like, the attribute features may include, but are not limited to, age, gender, city, and the like, and the current scene features may be understood as features of the scene where the user is currently located, for example, features including a currently used network, such as a network type and a network signal strength. The historical behavior feature may be a behavior feature of the user over a preset historical period of time.

Step S102: inputting the correlation characteristics into a pre-acquired target click rate prediction model for prediction to obtain a first predicted click rate vector, and inputting the correlation characteristics into a pre-acquired target multitask learning model for prediction to obtain a first predicted stay time vector and a first predicted completion rate vector.

Step S103: aiming at each candidate media resource of the candidate media resources, acquiring a second predicted click rate vector, a second predicted stay time vector and a second predicted completion rate vector of the candidate media resource; and determining a first score for the candidate media resource based on the first predicted click rate vector, the second predicted click rate vector, the first predicted dwell duration vector, the second predicted dwell duration vector, the first predicted completion rate vector, and the second predicted completion rate vector.

The media resources may include, but are not limited to, videos, pictures and texts, and the videos may also be classified into a first type of video and a second type of video, where the duration of the first type of video is different from the duration of the second type of video, for example, the first type of video may be understood as a video with a duration greater than a preset duration, and the second type of video may be understood as a video with a duration less than or equal to the preset duration, which may also be referred to as a small video. The plurality of candidate media assets are media assets of the application.

As one example, the resource characteristics may include, but are not limited to, characteristics of a resource Identification (ID), a resource type (e.g., video type, teletext type, etc.), a resource category (i.e., a resource classification, e.g., science, entertainment, education, etc.), and a resource size.

It is understood that the first predicted click rate vector, the first predicted dwell time vector and the first predicted completion rate vector are predicted vectors measured by users, the first predicted click rate vector may be a feature vector obtained by feature extraction of associated features of the first user by a target click rate prediction model, may be a vector representing a possibility of user behavior of the first user, the second predicted click rate vector, the second predicted dwell time vector and the second predicted completion rate vector may be predicted vectors measured by resources, the second predicted click rate vector may be a feature vector obtained by feature extraction of associated features by a target multitask learning model, may represent a possibility of user behavior of a candidate media resource, and the possibility of user behavior of the candidate media resource by the first user may be obtained by the first predicted click rate vector and the second predicted click rate vector, the predicted click rate may be obtained, for example, by inner-product the first predicted click rate vector with the second predicted click rate vector, and the predicted click rate may be obtained, for example, the predicted click rate may include predicted click or predicted no click.

The first predicted dwell time vector may represent a vector of predicted dwell time associated with dwell time of the first user after performing user actions, and the first predicted completion rate vector may represent a vector of completion rate associated with completion rate of the first user after performing user actions. The second predicted dwell time vector may characterize a predicted dwell time related vector of the candidate media resource after being acted upon by the user, and the second predicted completion rate vector may characterize a predicted completion rate related vector of the candidate media resource after being acted upon by the user. The predicted stay time of the first user after performing the user action on the candidate media resource can be obtained through the first predicted time vector and the second predicted time vector, that is, the predicted stay time can be obtained, for example, the predicted stay time is obtained by performing an inner product on the first predicted time vector and the second predicted time vector. The predicted completion rate of the first user after performing the user action on the candidate media resource can be obtained through the first predicted completion rate vector and the second predicted completion rate vector, that is, the predicted completion rate can be obtained, for example, after the user clicks a certain resource, the time is 10 seconds, the duration of the resource is 20 seconds, and the completion rate is 50%. In one example, the first vector of completion rates of prediction is inner-multiplied with the second vector of completion rates of prediction to obtain a completion rate of prediction.

After a first predicted click rate vector, a second predicted click rate vector, a first predicted stay time vector, a second predicted stay time vector, a first predicted completion rate vector and a second predicted completion rate vector of a certain candidate media resource are obtained, a first score of the candidate media resource is determined accordingly. In this embodiment, the corresponding first score may be calculated for each candidate media resource, so that the first scores of the plurality of candidate media resources may be obtained, that is, the plurality of first scores may be obtained.

It should be noted that the first score of the candidate media resource may be used to characterize the matching degree of the candidate media resource to the first user, and the higher the score is, the higher the matching degree is, the higher the possibility that the candidate media resource is subsequently recommended to the first user is.

Step S104: the plurality of candidate media resources are ranked based on their first scores.

I.e., ranking the plurality of candidate media assets is achieved by the first scores of the plurality of candidate media assets.

In the embodiment, the first predicted click rate vector is obtained by inputting the associated characteristics of the first user into a pre-acquired target click rate prediction model for prediction, inputting the correlation characteristics into a pre-acquired target multi-task learning model for prediction to obtain a first prediction stay time vector and a first prediction completion rate vector, and for each candidate media resource of the plurality of candidate media resources, obtaining a second predicted click rate vector, a second predicted stay time vector and a second predicted completion rate vector of the candidate media resource, determining a first score of the candidate media resource by using the first predicted click rate vector, the second predicted click rate vector, the first predicted stay time vector, the second predicted stay time vector, the first predicted completion rate vector and the second predicted completion rate vector, ranking the candidate media resources is then achieved by a first scoring of the plurality of candidate media resources. Namely, in the process of determining the first score of the candidate media resource, the associated feature of the first user is adopted, and the predicted click rate vector obtained by predicting the target click rate prediction model based on the associated feature of the first user and the first predicted stay time vector and the first predicted completion rate vector obtained by predicting the target multitask learning model based on the associated feature of the first user are taken into consideration, the second predicted click rate vector, the second predicted stay time vector and the second predicted completion rate vector of the candidate media resource are taken into consideration, and the first score of the candidate media resource is determined based on the first predicted click rate vector, the second predicted click rate vector and the first predicted stay time vector, so that the accuracy of the first score of the candidate media resource can be improved, the ranking of the candidate media resources is realized through the first score of the candidate media resources, thereby improving the accuracy of the ranking of the candidate media assets.

In one embodiment, determining a first score for a candidate media resource based on a first predicted click rate vector, a second predicted click rate vector, a first predicted length of dwell vector, a second predicted length of dwell vector, a first predicted completion rate vector, and a second predicted completion rate vector comprises:

determining a target predicted click rate based on the first predicted click rate vector and the second predicted click rate vector;

determining a target predicted duration based on the first predicted duration vector and the second predicted duration vector;

determining a target completion prediction rate based on the first completion prediction rate vector and the second completion prediction rate vector;

determining a first score of the candidate media resource using the target predicted click rate, the target predicted duration, and the target predicted completion rate.

That is, in this embodiment, in the process of determining the first score of the candidate media resource, the target predicted click rate determined by the first predicted click rate vector and the second predicted click rate vector, the target predicted duration determined by the first predicted duration vector and the second predicted duration vector, and the target predicted completion rate determined by the first predicted completion rate vector and the second predicted completion rate vector are considered, so as to improve the accuracy of the obtained first score, and thus improve the accuracy of resource ranking.

As one example, an inner product of the first predicted dwell time period vector and the second predicted dwell time period vector may be taken as the target predicted time period, and an inner product of the first predicted completion rate vector and the second predicted completion rate vector may be taken as the target predicted completion rate. In one example, determining a first score for a candidate media asset using a target predicted click rate, a target predicted duration, and a target predicted completion rate may include: and multiplying a first preset numerical power of the target predicted click rate, a second preset numerical power of the target predicted stay time and a third preset numerical power of the target predicted completion rate to obtain a first score of the candidate media resource. The first preset numerical value, the second preset numerical value and the third preset numerical value can be preset in advance through experience and are numbers larger than 0, the first score is obtained through the mode that the first preset numerical value power of the target prediction click rate, the second preset numerical value power of the target prediction stay time and the third preset numerical value power of the target prediction completion rate are multiplied, and the accuracy of the first score can be improved.

In one embodiment, after ranking the plurality of candidate media resources based on the first scores of the plurality of candidate media resources, the method further comprises:

and recommending the top N media resources in the candidate media resources to the first user, wherein N is a positive integer.

And sequencing the candidate media resources in a descending order of scores, wherein N media resources before sequencing are N media resources with larger scores, and recommending the N media resources with larger scores to the first user, so that the recommended media resources are more adaptive to the first user, have higher pertinence, and improve the accuracy of resource recommendation.

In one embodiment, obtaining a second predicted click rate vector, a second predicted dwell time vector, and a second predicted completion rate vector for a candidate media asset comprises:

and searching a second predicted click rate vector, a second predicted stay time vector and a second predicted completion rate vector of the candidate media resource from the cache, wherein the second predicted click rate vector of the candidate media resource is obtained by predicting for a target click rate prediction model based on the resource characteristics of the candidate media resource, and the second predicted stay time vector and the second predicted completion rate vector of the candidate media resource are obtained by predicting for a target multi-task learning model based on the resource characteristics of the candidate media resource.

The resource characteristics of each media resource in the first media resource set with preset identification can be input into a target click rate prediction model in advance for prediction to obtain a second predicted click rate vector of each media resource in the first media resource set and cached, a plurality of candidate media resources belong to the first media resource set, and the resource characteristics of each media resource in the first media resource set can be input into a target multitask learning model in advance for prediction to obtain a second predicted stay time vector and a second predicted completion rate vector of each media resource in the first media resource set and cached, so that in the process of sorting the candidate media resources, the second predicted click rate vector, the second predicted stay time vector and the second predicted completion rate vector of the candidate media resources can be searched from the cache, and the second predicted click rate vector, the second predicted stay time vector and the second predicted completion rate vector of the candidate media resources can be obtained in an improved mode, The speed of the second prediction stay time length vector and the second prediction completion rate vector, so that the overall resource sequencing efficiency is improved.

In one embodiment, the target click rate prediction model is determined by:

screening a plurality of users according to attribute characteristics of the plurality of users in the application to determine a first user set;

screening the first user set to obtain a second user set, wherein any user in the second user set has a first behavior of associating media resources with the first project of the application;

screening a first media resource set with a preset identifier from a plurality of applied media resources, wherein the preset identifier is used for indicating that the media resources can be published and distributed in a second item;

screening the first media resource set to obtain a second media resource set, wherein any media resource in the second media resource set has associated user behavior in the applied first project;

constructing a first sample data set based on the association characteristics of the second user set, the association characteristics of the second item association user of the application, the resource characteristics of the second media resource set and the resource characteristics of a third media resource set, wherein the third media resource set comprises the media resources of the second item association user of the application for performing a second action;

and training the initial click rate prediction model by using the first sample data set to obtain a target click rate prediction model.

In the process of acquiring the first sample data set, the multiple users are screened according to the attribute features of the multiple users in the application to determine a first user set, as an example, the attribute features may include but are not limited to at least one of gender, age, city, and the like, the first user set whose attribute features satisfy the preset attribute conditions may be screened from the multiple users, the preset attribute conditions may include but are not limited to that the matching degree of the attribute features and the attribute features of the first user is greater than the preset matching degree, that is, the attribute features of the first user are similar, and it can be understood that the second user set is a population group similar to the first user in the first item-associated users. In this embodiment, the behaviors of the crowd similar to the current first user in the first project and the behaviors of the users associated with the second project are aggregated to generate sample data, so as to train the click rate prediction model and improve the model training effect, thereby improving the performance of the obtained target click rate prediction model.

Since the screened first set of users are users who satisfy the attribute condition but may not have a first behavior on the media resource associated with the first item of the application, the screened first set of users are filtered to screen a second set of users, and the first behavior includes but is not limited to at least one of clicking, collecting, sharing, commenting and agreeing. It should be noted that, in the above application, a plurality of items (channels or themes) may be included, and different categories of resources may be displayed under different items, for example, hot spot items (mainly displaying hot spot resources), entertainment items (mainly displaying resources related to entertainment), mood items (mainly displaying resources such as trending, etc.), recommended items (mainly displaying resources recommended by an application), attention items, education items, science and technology items, international items, sports items, etc. may be included. The first item may be any at least one item among a plurality of items, and may be, for example, a recommended item. The second item may be any of a plurality of items, different from the first item, and may be a mood item, for example.

Any user in the second user set has a first behavior on the first project-associated media resource of the application, and the first project-associated media resource may be understood as a resource distributed in the first project or a resource shown in the first project, that is, the user in the second user set is a user of the first user set having the first behavior on the first project-associated media resource.

In the process of generating the first sample data set, not only the user but also the resource need to be considered, a first media resource set with a preset identifier may be screened from a plurality of media resources of the application, where the preset identifier is used to indicate that the media resources may be distributed in a second item, that is, the resources in the first media resource set are resources that may be distributed in the second item, and then the resources of the first media resource set that are not associated with user behaviors in a second item of the application are filtered to obtain a second media resource set, where any media resource in the second media resource set has associated user behaviors in the first item of the application, and the user behaviors may include, but are not limited to, at least one of clicking, collecting, sharing, commenting, and agreeing.

It should be noted that the second item associated users may be all users in the second item, and these users have an past behavior on resources in the second item in the application. The second behavior may include, but is not limited to, at least one of clicking, collecting, sharing, commenting, and agreeing, and may be the same as or different from the first behavior.

Therefore, a first sample data set can be constructed through the association characteristics of the second user set, the association characteristics of the applied second item association user, the resource characteristics of the second media resource set and the resource characteristics of the third media resource set, the initial click rate prediction model can be trained subsequently through the first sample data set to obtain a target click rate prediction model, so that the performance of the target click rate prediction model is improved, click rate prediction is performed subsequently through the target click rate prediction model, and the accuracy of the predicted click rate can be improved.

In one embodiment, the first sample data set includes associated features of the second user set, associated features of the second item-associated user of the application, resource features of the second media resource set, and resource features of the third media resource set, any piece of sample data in the first sample data set includes associated features of a user, resource features of a media resource, and a corresponding first tag, and the first tag includes an actual click tag.

The actual click tag may be understood as an actual click result of the user on the media resource, for example, one piece of sample data includes an associated feature of the user a1 and a resource feature of the media resource M1, and then the corresponding actual click tag is an actual click result of the user a1 on the media resource M1, for example, the user a1 clicks on the media resource M1, and then the actual click tag is a click tag, for example, the actual click tag may be set to 1, and if the user a1 does not click on the media resource M1, then the actual click tag is an un-click tag, for example, the actual click tag may be set to 0. The associated features of the users in any sample data in the first sample data set belong to the associated features of the total users, the total users include a second user set and a second project associated user, the resource features of the media resources in any sample data in the first sample data set belong to the resource features of the total resources, and the total resources include a second resource sample set and a third resource sample set.

In this embodiment, the initial click rate prediction model is trained by the association features of the second user set, the association features of the applied second item association users, the resource features of the second media resource set, and the resource features of the third media resource set, where the first label is the first sample data set of the actual click label, so that the model training effect can be improved, and thus the performance of the obtained target click rate prediction model is improved.

In one embodiment, the target multitask learning model is determined by:

screening a first media resource set with a preset identification from a plurality of applied media resources, wherein the preset identification is used for indicating that the media resources can be distributed in a second item;

constructing a second sample data set based on the association characteristics of the second user set, the association characteristics of the second item association user of the application, the resource characteristics of the second media resource set and the resource characteristics of a third media resource set, wherein the third media resource set comprises media resources for performing a second behavior by the second item association user of the application;

It should be noted that the target multitask learning model may be a target CGC (Customized Gate Control) model, the building process of the second sample data set is similar to the building process of the first sample data set, and is not described herein again, the feature data between the second sample data set and the first sample data set are the same, but the difference is that the former tag includes an actual dwell time length tag and an actual completion rate tag, and the latter tag is an actual click tag. The initial multi-task learning model is trained through the second sample data set which comprises the association characteristics of the second user set, the association characteristics of the applied second item association user, the resource characteristics of the second media resource set and the resource characteristics of the third media resource set, wherein the second label comprises the actual stay time label and the actual completion rate label, the model training effect can be improved, and therefore the performance of the obtained target multi-task learning model is improved.

In one embodiment, the second sample data set includes associated features of the second user set, associated features of the second item associated user applied, resource features of the second media resource set, and resource features of the third media resource set, any sample data in the second sample data set includes associated features of a user, resource features of a media resource, and a corresponding second tag, and the second tag includes an actual dwell time duration tag and an actual completion rate tag.

In this embodiment, the initial multi-task learning model is trained through the second sample data set including the association features of the second user set, the association features of the applied second item association users, the resource features of the second media resource set, and the resource features of the third media resource set, where the second label includes the actual dwell time label and the actual completion rate label, so that the model training effect can be improved, and thus the performance of the obtained target multi-task learning model is improved.

As shown in fig. 2, according to an embodiment of the present disclosure, the present disclosure further provides a click-through rate prediction model training method, including:

step S201: screening a plurality of users according to attribute characteristics of the plurality of users in the application to determine a first user set;

step S202: screening the first user set to obtain a second user set, wherein any user in the second user set has a first behavior of associating media resources with the first project of the application;

step S203: screening a first media resource set with a preset identifier from a plurality of applied media resources, wherein the preset identifier is used for indicating that the media resources can be published and distributed in a second item;

step S204: screening the first media resource set to obtain a second media resource set, wherein any media resource in the second media resource set has associated user behavior in the applied first project;

step S205: constructing a first sample data set based on the association characteristics of the second user set, the association characteristics of the second item association users of the application, the resource characteristics of the second media resource set and the resource characteristics of the third media resource set;

the third media resource set comprises media resources of a second action performed by a second item associated user of the application;

step S206: and training the initial click rate prediction model by using the first sample data set to obtain a target click rate prediction model.

In this embodiment, a first sample data set may be constructed through the association features of the second user set, the association features of the applied second item association user, the resource features of the second media resource set, and the resource features of the third media resource set, and then the initial click rate prediction model may be trained by using the first sample data set to obtain a target click rate prediction model, so as to improve the performance of the target click rate prediction model, and then click rate prediction may be performed by using the target click rate prediction model, so that the accuracy of the predicted click rate may be improved.

As shown in fig. 3, according to an embodiment of the present disclosure, the present disclosure further provides a method for training a multi-task learning model, including:

step S301: screening a plurality of users according to attribute characteristics of the plurality of users in the application to determine a first user set;

step S302: screening the first user set to obtain a second user set, wherein any user in the second user set has a first behavior of associating media resources with the first project of the application;

step S303: screening a first media resource set with a preset identification from a plurality of applied media resources, wherein the preset identification is used for indicating that the media resources can be distributed in a second item;

step S304: screening the first media resource set to obtain a second media resource set, wherein any media resource in the second media resource set has an associated user behavior in the applied first project;

step S305: constructing a second sample data set based on the association characteristics of the second user set, the association characteristics of the second item association users of the application, the resource characteristics of the second media resource set and the resource characteristics of the third media resource set;

step S306: and training the initial multi-task learning model by using the second sample data set to obtain a target multi-task learning model.

In this embodiment, a second sample data set may be constructed through the association features of the second user set, the association features of the applied second item association user, the resource features of the second media resource set, and the resource features of the third media resource set, and the initial multi-task learning model may be subsequently trained by using the second sample data set to obtain a target multi-task learning model, so as to improve the performance of the target multi-task learning model, and the prediction accuracy of the dwell time and the completion rate may be improved by subsequently performing the dwell time and the completion rate prediction by using the target multi-task learning model.

The process of the above method is described in detail below with an embodiment.

An off-line part:

(1) user collection: screening a first user set through user attribute characteristics (gender, age, city and the like); screening users with first project behaviors from the first user set to obtain a second user set; acquiring all user behaviors related to the current second item;

(2) resource collection: screening by resource characteristics (special identification which can be distributed in the current second item) to obtain a first media resource set; screening resources with first project behaviors from the first media resource set to obtain a second media resource set; acquiring resources of all user behaviors associated with the current second item, namely acquiring a third media resource set;

(3) sample generation:

clicking shows a sample: all actual clicks of the user on the resources in (2) in (1) above reveal behavior. And taking the actual click behavior with too short time as a negative sample, and filtering invalid clicks. Eliminating samples which are refreshed for the first time and are not clicked by a user, and filtering invalid displays;

duration sample: (1) the actual stay time of the user in (2) after all clicks on the resource in (2);

building a model: establishing an initial click rate prediction model for learning the click behavior of a user on a target resource, and establishing an initial multi-task learning model for learning the time duration behavior of the user on the target resource; for the initial click rate prediction model: in the training process, inputting the associated characteristics of a user, obtaining a click rate vector of a user side through a plurality of layers of fully-connected networks, inputting resource characteristics, obtaining a click rate vector of a resource side through a plurality of layers of fully-connected networks, fusing (for example, inner product) the click rate vector of the user side and the click rate vector of the resource side into a predicted click rate, and calculating cross entropy with a label on which the user clicks; for the multitask learning model: the method comprises the steps of inputting correlation characteristics of a user, obtaining a duration vector and a completion rate vector of a user side through a CGC structure, inputting resource characteristics, obtaining a duration vector and a completion rate vector of a resource side through the CGC structure, fusing the duration vector of the user side and the duration vector of the resource side into a predicted stay duration, and calculating mean square error (mse) loss with a normalized actual stay duration. And fusing the completion rate vector of the user side and the completion rate vector of the resource side into a predicted completion rate, and calculating mse with the actual completion rate.

Model training: and training the generated samples to obtain a target click rate prediction model and a target multi-task learning model, and storing parameters of a bottom layer embedded layer and an upper layer model (full connection network) generated by the models in a streaming manner to a cluster. In addition, the white list resource (for example, the first media resource set) may be processed through the trained target click rate prediction model and the target multitask learning model to obtain a second predicted click rate vector of the white list resource, a second predicted stay time vector of the white list resource, and a second predicted completion rate vector of the white list resource, and may be cached.

And an online part:

after training is finished, a target click rate prediction model and a target multi-task learning model are obtained, in the process of online application, the associated characteristic target click rate prediction model of the first user can be predicted to obtain a first predicted click rate, the associated characteristic of the first user can be input into the target multi-task learning model to be predicted to obtain a first predicted stay time vector and a first predicted completion rate vector; searching the cache of the resource measurement to obtain a second predicted click rate vector, a second predicted stay time vector and a second predicted completion rate vector of each candidate media resource, and fusing the corresponding second predicted click rate vector, the corresponding second predicted stay time vector and the corresponding second predicted completion rate vector of each candidate media resource through a fusion formula to obtain a first score of each candidate media resource, so that a plurality of first scores of the candidate media resources can be obtained, and the candidate media resources are ranked by the first scores of the candidate media resources.

According to the scheme of the embodiment of the disclosure, the behaviors of the first project associated users, which are similar to the current first user, and the behaviors of the second project are combined to construct a sample, model training is performed, a model training effect is provided, and the performance of the obtained model is improved. And the CGC structure is adopted to simultaneously optimize the duration and the completion rate, so that the commonality of two targets which are originally closely related can be modeled.

As shown in fig. 4, the present disclosure also provides a media resource ranking apparatus 400 according to an embodiment of the present disclosure, the apparatus including:

a first obtaining module 401, configured to obtain an association characteristic of a first user;

a score determining module 402, configured to obtain, for each candidate media resource of the multiple candidate media resources, a second predicted click rate vector, a second predicted stay time vector, and a second predicted completion rate vector of the candidate media resource; and determining a first score for the candidate media resource based on the first predicted click rate vector, the second predicted click rate vector, the first predicted dwell duration vector, the second predicted dwell duration vector, the first predicted completion rate vector, and the second predicted completion rate vector;

a ranking module 403 for ranking the plurality of candidate media resources based on the first scores of the plurality of candidate media resources.

In one embodiment, the scoring module includes:

a first determination module to determine a target predicted click rate based on the first predicted click rate vector and the second predicted click rate vector;

a second determination module to determine a target predicted duration based on the first predicted duration vector and the second predicted duration vector;

a third determination module to determine a target completion prediction rate based on the first completion prediction rate vector and the second completion prediction rate vector;

and the resource score determining module is used for determining a first score of the candidate media resource by using the target prediction click rate, the target prediction duration and the target prediction completion rate.

In one embodiment, the apparatus further comprises:

and the recommending module is used for recommending the top N media resources in the candidate media resources to the first user, wherein N is a positive integer.

The media resource sorting device in each of the embodiments is a device for implementing the media resource sorting method in each of the embodiments, and the technical features correspond to those of the media resource sorting device in each of the embodiments, and the technical effects correspond to those of the media resource sorting device in each of the embodiments, and are not described herein again.

As shown in fig. 5, according to an embodiment of the present disclosure, the present disclosure further provides a click-through rate prediction model training apparatus 500, including:

a first user determining module 501, configured to filter multiple users according to attribute features of the multiple users in the application, and determine a first user set;

a second user determining module 502, configured to filter the first user set to obtain a second user set, where any user in the second user set has a first behavior of associating a media resource with the first project of the application;

a first resource screening module 503, configured to screen a first media resource set with a preset identifier from a plurality of media resources of an application, where the preset identifier is used to indicate that a media resource can be distributed in a second item;

a second resource screening module 504, configured to screen the first media resource set to obtain a second media resource set, where any media resource in the second media resource set has an associated user behavior in the applied first project;

a first sample structure modeling block 505, configured to construct a first sample data set based on the association features of the second user set, the association features of the second item-associated users of the application, the resource features of the second media resource set, and the resource features of a third media resource set, where the third media resource set includes media resources for performing a second action by the second item-associated users of the application;

and the first training module 506 is configured to train the initial click rate prediction model by using the first sample data set to obtain a target click rate prediction model.

The click rate prediction model training device in each embodiment is a device for implementing the click rate prediction model training method in each embodiment, and has corresponding technical features and technical effects, which are not described herein again.

As shown in fig. 6, the present disclosure also provides a multitask learning model training device 600 according to an embodiment of the present disclosure, including:

a third user determining module 601, configured to filter, according to attribute features of multiple users in an application, the multiple users to determine a first user set;

a fourth user determining module 602, configured to filter the first user set to obtain a second user set, where any user in the second user set has a first behavior of associating a media resource with the first project of the application;

a third resource screening module 603, configured to screen a first media resource set having a preset identifier from the multiple media resources of the application, where the preset identifier is used to indicate that the media resource may be distributed in the second item;

a fourth resource screening module 604, configured to screen the first media resource set to obtain a second media resource set, where any media resource in the second media resource set has an associated user behavior in the applied first project;

a second sample construction module 605, configured to construct a second sample data set based on the associated features of the second user set, the associated features of the second item-associated user of the application, the resource features of the second media resource set, and the resource features of a third media resource set, where the third media resource set includes media resources for performing a second behavior by the second item-associated user of the application;

and a second training module 606, configured to train the initial multi-task learning model by using a second sample data set, so as to obtain a target multi-task learning model.

In the technical scheme of the disclosure, the acquisition, storage, application and the like of the personal information of the related user all accord with the regulations of related laws and regulations, and do not violate the good customs of the public order.

The multi-task learning model training device of each embodiment is a device for implementing the multi-task learning model training method of each embodiment, and has corresponding technical features and technical effects, which are not described herein again.

The present disclosure also provides an electronic device, a readable storage medium, and a computer program product according to embodiments of the present disclosure.

A non-transitory computer-readable storage medium of an embodiment of the present disclosure stores computer instructions for causing a computer to perform a media asset ranking method provided by the present disclosure.

The computer program product of the embodiments of the present disclosure includes a computer program for causing a computer to execute the media resource ranking method provided by the embodiments of the present disclosure.

FIG. 7 illustrates a schematic block diagram of an example electronic device 700 that can be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 7, the electronic device 700 includes a computing unit 701, which may perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM)702 or a computer program loaded from a storage unit 708 into a Random Access Memory (RAM) 703. In the RAM703, various programs and data required for the operation of the device 700 can also be stored. The computing unit 701, the ROM 702, and the RAM703 are connected to each other by a bus 704. An input/output (I/O) interface 705 is also connected to bus 704.

A number of components in the electronic device 700 are connected to the I/O interface 705, including: an input unit 706 such as a keyboard, a mouse, or the like; an output unit 707 such as various types of displays, speakers, and the like; a storage unit 708 such as a magnetic disk, optical disk, or the like; and a communication unit 709 such as a network card, modem, wireless communication transceiver, etc. The communication unit 709 allows the electronic device 700 to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.

Computing unit 701 may be a variety of general purpose and/or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various dedicated artificial intelligence (I) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, and so forth. The calculation unit 701 performs the respective methods and processes described above, such as a media resource ranking method, a click-through rate prediction model training method, a multitask learning model training method, and the like. For example, in some embodiments, the media asset ranking method, the click-through rate prediction model training method, or the multi-task learning model training method may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as storage unit 708. In some embodiments, part or all of a computer program may be loaded onto and/or installed onto device 700 via ROM 702 and/or communications unit 709. When loaded into RAM703 and executed by the computing unit 701, may perform one or more steps of the media asset ranking method, click-through rate prediction model training method, or multi-task learning model training method described above. Alternatively, in other embodiments, the computing unit 701 may be configured by any other suitable means (e.g., by means of firmware) to perform a media asset ranking method, a click-through rate prediction model training method, or a multi-task learning model training method. Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions/operations specified in the flowchart and/or block diagram to be performed. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), the internet, and blockchain networks.

The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The Server can be a cloud Server, also called a cloud computing Server or a cloud host, and is a host product in a cloud computing service system, so as to solve the defects of high management difficulty and weak service expansibility in the traditional physical host and VPS service ("Virtual Private Server", or simply "VPS"). The server may also be a server of a distributed system, or a server incorporating a blockchain.

It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present disclosure may be executed in parallel or sequentially or in different orders, and are not limited herein as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved.

The above detailed description should not be construed as limiting the scope of the disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present disclosure should be included in the scope of protection of the present disclosure.

Claims

1. A method of media asset ranking, the method comprising:

acquiring the association characteristics of a first user;

2. The method of claim 1, wherein the determining a first score for the candidate media asset based on the first predicted click rate vector, the second predicted click rate vector, the first predicted length of dwell vector, the second predicted length of dwell vector, the first predicted completion rate vector, and the second predicted completion rate vector comprises:

determining a target predicted duration based on the first predicted duration of dwell vector and the second predicted duration of dwell vector;

determining a target completion rate of prediction based on the first completion rate of prediction vector and the second completion rate of prediction vector;

determining a first score for the candidate media resource using the target predicted click rate, the target predicted duration, and the target predicted completion rate.

3. The method of claim 1, wherein after ranking the plurality of candidate media assets based on their first scores, further comprising:

4. The method of claim 1, wherein said obtaining a second predicted click rate vector, a second predicted dwell time vector, and a second predicted completion rate vector for the candidate media asset comprises:

searching the second predicted click rate vector, the second predicted stay time vector and the second predicted completion rate vector of the candidate media resource from a cache, wherein the second predicted click rate vector of the candidate media resource is obtained by predicting the target click rate prediction model based on the resource characteristics of the candidate media resource, and the second predicted stay time vector and the second predicted completion rate vector of the candidate media resource are obtained by predicting the target multitask learning model based on the resource characteristics of the candidate media resource.

5. A click-through rate prediction model training method comprises the following steps:

6. The method of claim 5, wherein the first sample data set comprises associated features of the second set of users, associated features of the second item-associated users of the application, resource features of the second set of media resources, and resource features of a third set of media resources, and wherein any sample data in the first sample data set comprises associated features of a user, resource features of a media resource, and a corresponding first tag, and wherein the first tag comprises an actual click tag.

7. A multi-task learning model training method comprises the following steps:

screening the first media resource set to obtain a second media resource set, wherein any media resource in the second media resource set has an associated user behavior in the first project of the application;

8. The method of claim 7, wherein the second sample data set comprises associated features of the second set of users, associated features of a second item-associated user of the application, resource features of the second set of media resources, and resource features of a third set of media resources, any sample data in the second sample data set comprises associated features of a user, resource features of a media resource, and a corresponding second label, the second label comprising an actual dwell time label and an actual completion rate label.

9. An apparatus for ranking media assets, the apparatus comprising:

10. The apparatus of claim 9, wherein the scoring module comprises:

and the resource score determining module is used for determining a first score of the candidate media resource by using the target predicted click rate, the target predicted duration and the target predicted completion rate.

11. The apparatus of claim 9, wherein the apparatus further comprises:

and the recommending module is used for recommending the media resources of N in the candidate media resources to the first user, wherein N is a positive integer.

12. The apparatus of claim 9, wherein said obtaining a second predicted click rate vector, a second predicted dwell time vector, and a second predicted completion rate vector for the candidate media asset comprises:

13. A click-through rate prediction model training device comprises:

14. The apparatus of claim 13, wherein the first sample data set comprises associated features of the second set of users, associated features of a second item-associated user of the application, resource features of the second set of media resources, and resource features of a third set of media resources, and wherein any sample data in the first sample data set comprises associated features of a user, resource features of a media resource, and a corresponding first tag, and wherein the first tag comprises an actual click tag.

15. A multitask learning model training device comprising:

a fourth user determination module, configured to filter the first user set to obtain a second user set, where any user in the second user set has a first behavior of associating a media resource with the first project of the application;

a second sample construction module, configured to construct a second sample data set based on the associated features of the second user set, the associated features of the second item-associated user of the application, the resource features of the second media resource set, and the resource features of a third media resource set, where the third media resource set includes media resources for performing a second behavior by the second item-associated user of the application;

16. The apparatus of claim 15, wherein the second sample data set comprises associated features of the second set of users, associated features of a second item-associated user of the application, resource features of the second set of media resources, and resource features of a third set of media resources, any sample data in the second sample data set comprises associated features of a user, resource features of a media resource, and a corresponding second label, the second label comprising an actual duration of stay label and an actual completion rate label.

17. An electronic device, comprising:

at least one processor; and

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of ranking media assets of any one of claims 1-4 or the method of training a click-through rate prediction model of any one of claims 5-6 or the method of training a multi-task learning model of any one of claims 7-8.

18. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method for media asset ranking of any of claims 1-4 or the method for click-through rate prediction model training of any of claims 5-6 or the method for multitask learning model training of any of claims 7-8.

19. A computer program product comprising a computer program which, when executed by a processor, implements a media asset ranking method according to any of claims 1-4 or a click-through rate prediction model training method according to any of claims 5-6 or a multi-task learning model training method according to any of claims 7-8.