CN115757933A

CN115757933A - Recommendation information generation method, device, equipment, medium and program product

Info

Publication number: CN115757933A
Application number: CN202211194196.3A
Authority: CN
Inventors: 刘银星; 阮涛; 张政; 吕晶晶
Original assignee: Beijing Jingdong Century Trading Co Ltd; Beijing Wodong Tianjun Information Technology Co Ltd
Current assignee: Beijing Jingdong Century Trading Co Ltd; Beijing Wodong Tianjun Information Technology Co Ltd
Priority date: 2022-09-28
Filing date: 2022-09-28
Publication date: 2023-03-07
Also published as: WO2024067568A1

Abstract

Embodiments of the present disclosure disclose recommendation information generation methods, apparatuses, devices, media, and program products. One embodiment of the method comprises: acquiring a historical browsing creative image sequence, a user sparse feature information set, a main image and an article sparse feature information set; carrying out image coding processing on the main image to obtain a main image coding vector, and carrying out image coding processing on each historical browsing creative image to generate a historical image coding vector; performing information coding on each user sparse feature information to generate a user sparse feature vector, and performing information coding on each article sparse feature information to generate an article sparse feature vector; performing visual preference adjustment on each historical image coding vector to obtain an adjusted historical image coding vector sequence; and generating a recommended image set which is to be pushed to the target user and corresponds to the target recommended article. The implementation mode is related to artificial intelligence, accurate recommendation information is generated, and a good recommendation effect is obtained.

Description

Recommendation information generation method, device, equipment, medium and program product

Technical Field

Embodiments of the present disclosure relate to the field of computer technologies, and in particular, to a recommendation information generation method, apparatus, device, medium, and program product.

Background

Currently, there are often multiple pre-designed materials for an item (e.g., an item host, creative image, etc.). For selecting the material that the target user may prefer from a plurality of materials, the following methods are generally adopted: firstly, a plurality of materials are input into a pre-trained user material preference model to generate a score set corresponding to the plurality of materials. Then, at least one material that the target user may prefer is screened out from the plurality of materials by using the score set. And finally, recommending at least one material to the target user.

However, the inventor has found that when the above-described manner is adopted to screen out material that may be preferred by a target user from a plurality of materials, there are often technical problems as follows:

the user material preference model is trained only by utilizing the historical browse creative image sequence aiming at the target user, so that the characteristic information which can be learned by the user material preference model is limited, and the user material preference model is not accurate enough. The side results in poor recommendations.

The above information disclosed in this background section is only for enhancement of understanding of the background of the inventive concept and, therefore, it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art in this country.

Disclosure of Invention

This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

Some embodiments of the present disclosure propose recommendation information generation methods, apparatuses, devices, media, and program products to solve the technical problems mentioned in the background section above.

In a first aspect, some embodiments of the present disclosure provide a recommendation information generating method, including: acquiring a historical browsing creative image sequence of a target user, a user sparse feature information set of the target user, a main image of a target recommended article and an article sparse feature information set of the target recommended article, wherein the article sparse feature information set comprises a creative image feature information set aiming at the target recommended article; performing image coding processing on the main image to obtain a main image coding vector, and performing image coding processing on each historical browsing creative image in the historical browsing creative image sequence to generate a historical image coding vector to obtain a historical image coding vector sequence; performing information coding on each user sparse feature information in the user sparse feature information set to generate a user sparse feature vector to obtain a user sparse feature vector set, and performing information coding on each article sparse feature information in the article sparse feature information set to generate an article sparse feature vector to obtain an article sparse feature vector set; performing visual preference adjustment on each historical image coding vector in the historical image coding vector sequence to obtain an adjusted historical image coding vector sequence; and generating a recommended image set corresponding to the target recommended article to be pushed to the target user by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set.

Optionally, the performing a visual preference adjustment on each historical image coding vector in the historical image coding vector sequence to obtain an adjusted historical image coding vector sequence includes: determining visual preference information corresponding to the historical image coding vector sequence as target visual preference information; and adjusting each historical image coding vector in the historical image coding vector sequence according to the target visual preference information to obtain an adjusted historical image coding vector sequence.

Optionally, the generating, by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set, and the article sparse feature vector set, a recommended image set corresponding to the target recommended article to be pushed to the target user includes: carrying out vector splicing on the user sparse feature vector set and the article sparse feature vector set to obtain spliced sparse feature vectors; inputting the adjusted historical image coding vector sequence, the main image coding vector and the splicing sparse feature vector into the multi-head attention mechanism model to obtain a feature information fusion vector; inputting the spliced sparse feature vector into a full-connection model to obtain an output vector; splicing the characteristic information fusion vector and the output vector to obtain a spliced vector; and generating the recommended image set by using a preset loss function.

Optionally, the performing a graph coding process on each history browsing creative image in the history browsing creative image sequence to generate a history image coding vector includes: and inputting the historical browse creative image into a pre-trained image coding model to generate a historical image coding vector.

Optionally, the graph coding model includes: a residual network model and a plurality of fully connected layers; and the above-mentioned picture coding model that inputs the above-mentioned historical browse creative image to train in advance, in order to produce the historical picture coding vector, including: inputting the historical browse creative image into the residual error network model to obtain a model output result; and inputting the output result of the model to the plurality of full-connection layers to obtain the historical image coding vector.

Optionally, the generating the recommended image set by using a preset loss function includes: generating a creative image score set aiming at the splicing vector by using the preset loss function; generating a recommended article score corresponding to the target recommended article by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set; in response to determining that the recommended item score is greater than a predetermined recommended item value, determining a creative image score with a score value in the creative image score set greater than a predetermined recommended creative image value to obtain a creative image score subset; and determining the creative image set corresponding to the creative image score subset as the recommended image set.

Optionally, the method further includes: generating a recommended main image score corresponding to the target recommended article by utilizing a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set; and in response to the fact that the recommended item score is larger than a preset recommended item value and the recommended main image score is larger than a preset recommended main image value, pushing the main image to a terminal corresponding to the target user.

In a second aspect, some embodiments of the present disclosure provide a recommendation information generating apparatus, including: an acquisition unit configured to acquire a history browsing creative image sequence of a target user, a user sparse feature information set of the target user, a main image of a target recommended item, and an item sparse feature information set of the target recommended item, wherein the item sparse feature information set includes a creative image feature information set for the target recommended item; a graph coding unit configured to perform graph coding processing on the main image to obtain a main graph coding vector, and perform graph coding processing on each history browsing creative image in the history browsing creative image sequence to generate a history image coding vector to obtain a history image coding vector sequence; an information encoding unit configured to perform information encoding on each user sparse feature information in the user sparse feature information set to generate a user sparse feature vector to obtain a user sparse feature vector set, and perform information encoding on each article sparse feature information in the article sparse feature information set to generate an article sparse feature vector to obtain an article sparse feature vector set; the adjusting unit is configured to perform visual preference adjustment on each historical image coding vector in the historical image coding vector sequence to obtain an adjusted historical image coding vector sequence; and a generating unit configured to generate a recommended image set corresponding to the target recommended item to be pushed to the target user by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the item sparse feature vector set.

Optionally, the adjusting unit may be configured to: determining visual preference information corresponding to the historical image coding vector sequence as target visual preference information; and adjusting each historical image coding vector in the historical image coding vector sequence according to the target visual preference information to obtain an adjusted historical image coding vector sequence.

Optionally, the generating unit may be configured to: carrying out vector splicing on the user sparse feature vector set and the article sparse feature vector set to obtain spliced sparse feature vectors; inputting the adjusted historical image coding vector sequence, the main image coding vector and the splicing sparse feature vector into the multi-head attention mechanism model to obtain a feature information fusion vector; inputting the spliced sparse feature vector into a full-connection model to obtain an output vector; splicing the characteristic information fusion vector and the output vector to obtain a spliced vector; and generating the recommended image set by using a preset loss function.

Optionally, the graph encoding unit may be configured to: and inputting the historical browse creative image into a pre-trained image coding model to generate a historical image coding vector.

Optionally, the graph coding model includes: a residual network model and a plurality of fully connected layers; and the graph encoding unit may be configured to: inputting the historical browsing creative image into the residual error network model to obtain a model output result; and inputting the output result of the model to the plurality of full-connection layers to obtain the historical image coding vector.

Optionally, the generating unit may be configured to: generating a creative image score set aiming at the splicing vector by using the preset loss function; generating a recommended article score corresponding to the target recommended article by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set; in response to determining that the recommended item score is greater than a predetermined recommended item value, determining a creative image score with a score value in the creative image score set greater than a predetermined recommended creative image value to obtain a creative image score subset; and determining the creative image set corresponding to the creative image score subset as the recommended image set.

Optionally, the apparatus further comprises: generating a recommended main image score corresponding to the target recommended article by utilizing a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set; and in response to the fact that the recommended item score is larger than a preset recommended item value and the recommended main image score is larger than a preset recommended main image value, pushing the main image to a terminal corresponding to the target user.

In a third aspect, some embodiments of the present disclosure provide an electronic device, comprising: one or more processors; a storage device having one or more programs stored thereon, which when executed by one or more processors, cause the one or more processors to implement a method as described in any implementation of the first aspect.

In a fourth aspect, some embodiments of the disclosure provide a computer readable medium having a computer program stored thereon, where the program when executed by a processor implements a method as described in any of the implementations of the first aspect.

In a fifth aspect, some embodiments of the present disclosure provide a computer program product comprising a computer program that, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

The above embodiments of the present disclosure have the following beneficial effects: by the recommendation information generation method of some embodiments of the disclosure, accurate recommendation information is generated, and a better recommendation effect is obtained. Specifically, the reasons for the poor effect of the relevant recommendations are: the user material preference model is trained only by utilizing the historical browse creative image sequence aiming at the target user, so that the characteristic information which can be learned by the user material preference model is limited, and the user material preference model is not accurate enough. The side results in poor recommendations. Based on this, in the recommendation information generation method of some embodiments of the present disclosure, first, a history browsing creative image sequence of a target user, a user sparse feature information set of the target user, a main image of a target recommended item, and an item sparse feature information set of the target recommended item are obtained, where the item sparse feature information set includes a creative image feature information set for the target recommended item. For subsequent retrieval of more feature information to facilitate subsequent generation of more accurate recommendation information (i.e., multiple recommended creative image scores and recommended primary image scores). Then, image coding processing is carried out on the main image to obtain a main image coding vector, and image coding processing is carried out on each history browsing creative image in the history browsing creative image sequence to generate a history image coding vector to obtain a history image coding vector sequence. Here, the main image is subjected to image encoding processing to extract feature information of the main image. In addition, the generation of recommendation information using the main image encoding vector can effectively solve the problem of large calculation amount due to large pixel dimension of the image, compared with the generation of recommendation information by the main image. Similarly, compared with the method for adjusting the visual preference through the history browsing creative image sequence, the method for adjusting the visual preference of the history image coding vector sequence can effectively solve the problem of large calculation amount. And then, carrying out information coding on each user sparse feature information in the user sparse feature information set so as to convert the user sparse feature information into a vector form, so that the feature information of the user sparse feature can be conveniently used. Similarly, each article sparse feature information in the article sparse feature information set is subjected to information coding to be converted into a vector form, so that the feature information of the article sparse features can be conveniently used. Furthermore, visual preference adjustment is carried out on each historical image coding vector in the historical image coding vector sequence, so that the visual preference which can be embodied by the historical image coding vector sequence after adjustment is more obvious, and more accurate recommendation information can be generated subsequently. Finally, by means of the multi-head attention mechanism model, the feature information in multiple aspects of the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set can be extracted. Thus, the generated recommendation information (i.e., the recommended image set) is more accurate.

Drawings

The above and other features, advantages and aspects of various embodiments of the present disclosure will become more apparent by referring to the following detailed description when taken in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numbers refer to the same or similar elements. It should be understood that the drawings are schematic and that elements and elements are not necessarily drawn to scale.

1-2 are schematic diagrams of one application scenario of a recommendation information generation method according to some embodiments of the present disclosure;

FIG. 3 is a flow diagram of some embodiments of a recommendation information generation method according to the present disclosure;

FIG. 4 is a flow diagram of further embodiments of a recommendation information generation method according to the present disclosure;

FIG. 5 is a schematic block diagram of some embodiments of a recommendation information generating device according to the present disclosure;

FIG. 6 is a schematic block diagram of an electronic device suitable for use in implementing some embodiments of the present disclosure.

Detailed Description

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it is to be understood that the disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and the embodiments of the disclosure are for illustration purposes only and are not intended to limit the scope of the disclosure.

It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings. The embodiments and features of the embodiments in the present disclosure may be combined with each other without conflict.

It should be noted that the terms "first", "second", and the like in the present disclosure are only used for distinguishing different devices, modules or units, and are not used for limiting the order or interdependence of the functions performed by the devices, modules or units.

It is noted that references to "a", "an", and "the" modifications in this disclosure are intended to be illustrative rather than limiting, and that those skilled in the art will recognize that "one or more" may be used unless the context clearly dictates otherwise.

The names of messages or information exchanged between devices in the embodiments of the present disclosure are for illustrative purposes only, and are not intended to limit the scope of the messages or information.

The operations involved in the present disclosure of collecting, storing, using, etc. creative materials (e.g., creative images, main images) include, as far as possible, the relevant organization or individual undertaking material security impact assessment, fulfilling notification obligations to material creating agents, soliciting prior authorization consent of material creating agents, etc., before performing the respective operations.

The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

Fig. 1-2 are schematic diagrams of an application scenario of a recommendation information generation method according to some embodiments of the present disclosure.

In the application scenarios of fig. 1-2, first, the electronic device 101 may acquire a history browsing creative image sequence 103 of the target user 102, a user sparse feature information set 104 of the target user 102, a main image 106 of the target recommended item 105, and an item sparse feature information set 107 of the target recommended item 105. Wherein the item sparse feature information set 107 includes a creative image feature information set for the target recommended item 105. In the context of the present application, the target user 102 may be "li. The historically viewed creative image sequence 103 may include: a history browsing creative image 1031, and a history browsing creative image 1032. The user sparse feature information set 104 may be: { "sex: 1"," age: 18"," height: 184"}. The target recommended item 105 may be an "apple. The item sparse feature information set 107 may be: { "click rate: 0.4"," price: 5"," place of origin: 02"}. Wherein gender "1" may characterize a male. The production area "02" may characterize the Shanghai. The electronic device 101 may then perform a graph encoding process on the primary image 106 to obtain a primary graph encoding vector 110, and perform a graph encoding process on each of the historically viewed creative images in the sequence of historically viewed creative images 103 to generate a historical image encoding vector to obtain a sequence of historical image encoding vectors 108. In this application scenario, the historical image coding vector sequence 108 may include: a history image code vector 1081 corresponding to the history browsing creative image 1031 and a history image code vector 1082 corresponding to the history browsing creative image 1032. Next, the electronic device 101 may perform information encoding on each user sparse feature information in the user sparse feature information set 104 to generate a user sparse feature vector, obtain a user sparse feature vector set 109, and perform information encoding on each item sparse feature information in the item sparse feature information set 107 to generate an item sparse feature vector, obtain an item sparse feature vector set 111. In the present application scenario, the user sparse feature vector set 109 may include: "sex: 1", corresponding user sparse feature vector 1091," age: 18 "corresponding user sparse feature vector 1092 and" height: 184 "corresponding to the user sparse feature vector 1093. The sparse feature vector set 111 of articles may include: "click rate: 0.4 "corresponding sparse feature vector 1111," price: 5 "corresponding article sparse feature vector 1112 and" origin: 02 "corresponding to the sparse feature vector 1113. Furthermore, the electronic device 101 may perform a visual preference adjustment on each historical image encoding vector in the historical image encoding vector sequence 108 to obtain an adjusted historical image encoding vector sequence 112. In this application scenario, the adjusted historical image coding vector sequence 112 includes: an adjusted history image coding vector 1121 corresponding to the history image coding vector 1081 and an adjusted history image coding vector 1122 corresponding to the history image coding vector 1082. Finally, the electronic device 101 may generate, by using a multi-head attention mechanism model 113, a recommended image set 114 corresponding to the target recommended item 105 to be pushed to the target user 102 according to the adjusted historical image encoding vector sequence 112, the main image encoding vector 110, the user sparse feature vector set 109, and the item sparse feature vector set 111. The recommended image set 114 is an image subset of the creative image set corresponding to the creative image feature information set. In the present application scenario, the recommended image set 114 includes: a recommendation image 1141 and a recommendation image 1142.

The electronic device 101 may be hardware or software. When the electronic device is hardware, the electronic device may be implemented as a distributed cluster formed by a plurality of servers or terminal devices, or may be implemented as a single server or a single terminal device. When the electronic device is embodied as software, it may be installed in the above-listed hardware devices. It may be implemented, for example, as multiple software or software modules to provide distributed services, or as a single software or software module. And is not particularly limited herein.

It should be understood that the number of electronic devices in fig. 1-2 is merely illustrative. There may be any number of electronic devices, as desired for implementation.

With continued reference to fig. 3, a flow 300 of some embodiments of a recommendation information generation method in accordance with the present disclosure is shown. The recommendation information generation method comprises the following steps:

step 301, obtaining a history browsing creative image sequence of a target user, a user sparse feature information set of the target user, a main image of a target recommended item, and an item sparse feature information set of the target recommended item.

In some embodiments, an execution subject (e.g., the electronic device 101 shown in fig. 1) of the recommendation information generation method may obtain the sequence of the historically-viewed creative images of the target user, the user sparse feature information set of the target user, the main image of the target recommended item, and the item sparse feature information set of the target recommended item through a wired connection manner or a wireless connection manner. The item sparse feature information set comprises a creative image feature information set aiming at the target recommended item. The creative image feature information may be ID identification information of the creative image of the target recommended item. The ID identification information and the creative image have a one-to-one correspondence. The target user can be a user of the material to be recommended. In practice, the material may be presented in the form of pictures or videos. For the form that the material is a video, the key frames extracted from the video can be used as the content recommended to the user. For the e-market scenes, the material can be a main picture of the article and can also be a creative image of the article. The creative image may be a pre-designed image for the characteristics of the item. The historically viewed creative image sequence may be a sequence of creative images viewed by the target user on respective applications (apps) over a historical period of time. For example, the current time is 2022 years, month 4. The sequence of historically viewed creative images may be a sequence of user viewed creative images from 1 month 2021 to 4 months 2021. The primary image of the target recommended item may be a primary image of the target recommended item. The user sparse feature information of the target user may be numerical information of a sparse feature of the target user. The sparse feature of the target user may be an ID class feature of the target user. The ID feature may be a unique ID of the user feature, or may be an identity of the user. For example, the sparse feature of the target user may be one of: the gender ID identification of the target user, the age ID identification of the target user, the ID card ID identification of the target user and the height ID identification of the target user. As another example, the gender ID is identified as "1," indicating that the target user is a male. The gender ID is identified as "0" indicating that the target user is female. The item sparse feature information of the target recommended item may be numerical information of a sparse feature of the target recommended item. The sparse feature of the target recommended item may be an ID class feature of the target recommended item. The ID type feature can be a unique ID identification of the article feature, and can also be an identification of the article identity. For example, the sparse characteristic of the target recommended item may be one of: click rate of the target recommended item, price of the target recommended item, and origin of the target recommended item. As another example, the origin ID is identified as "02" and the origin characterizing the target recommended item is Shanghai.

It should be noted that the information that can be recommended to the target user may include, but is not limited to, one of the following: the creative image of the target recommended item and the main image of the target recommended item.

Step 302, performing image coding processing on the main image to obtain a main image coding vector, and performing image coding processing on each history browsing creative image in the history browsing creative image sequence to generate a history image coding vector to obtain a history image coding vector sequence.

In some embodiments, the execution subject may perform a graph coding process on the main image to obtain a main graph coding vector, and perform a graph coding process on each of the sequence of historically viewed creative images to generate a historical image coding vector to obtain a sequence of historical image coding vectors. The primary image encoding vector may characterize image feature information of the primary image. The historical image encoding vectors may characterize image feature information of the historical browse creative images.

As an example, first, the execution body may input the main image into a plurality of serially connected Convolutional Neural Networks (CNNs), resulting in a first model output result. And then, inputting the output result of the first model into the Bert coding model to obtain a main image coding vector.

Likewise, first, the execution agent can input each historically viewed creative image in the sequence of historically viewed creative images into a plurality of serially connected convolutional neural networks to generate a second model output result, resulting in a second sequence of model output results. And then, inputting each second model output result in the second model output result sequence into the Bert coding model to obtain a historical image coding vector sequence.

The subsequent generation of the plurality of recommended creative image scores and recommended primary image scores using the primary image encoding vector is less computationally intensive than the subsequent generation of the plurality of recommended creative image scores and recommended primary image scores using the primary image. Compared with the method for adjusting the visual preference through the historical browsing creative image sequence, the method for adjusting the visual preference of the historical image coding vector sequence can effectively solve the problem of large calculation amount.

In some optional implementations of some embodiments, the execution subject may input the historical browse creative image to a pre-trained graph coding model to generate a historical image coding vector.

The graph coding model may be a model that codes the creative image to generate a coding vector. For example, the graph coding model may be a plurality of serially connected convolutional neural networks.

Optionally, the graph coding model includes: a residual network model and a plurality of fully connected layers. The above inputting the historical browsing creative image into a pre-trained graph coding model to generate a historical image coding vector may include the following steps:

firstly, inputting the historical browse creative image into a Residual network (Resnets) model to obtain a model output result.

And secondly, inputting the output result of the model to the plurality of full-connection layers to obtain the historical image coding vector. Wherein the plurality of fully connected layers may be a plurality of fully connected layers connected in series.

Step 303, performing information coding on each user sparse feature information in the user sparse feature information set to generate a user sparse feature vector to obtain a user sparse feature vector set, and performing information coding on each article sparse feature information in the article sparse feature information set to generate an article sparse feature vector to obtain an article sparse feature vector set.

In some embodiments, the execution subject may perform information encoding on each user sparse feature information in the user sparse feature information set to generate a user sparse feature vector, obtain a user sparse feature vector set, and perform information encoding on each item sparse feature information in the item sparse feature information set to generate an item sparse feature vector, obtain an item sparse feature vector set. The user sparse feature vector can represent feature information of the user sparse feature information. The sparse feature vector of the article can represent feature information of sparse feature information of the article.

As an example, the executing entity may input each user sparse feature information in the user sparse feature information set to the Bert coding model to generate a user sparse feature vector, resulting in a user sparse feature vector set. Similarly, the executing entity may input each article sparse feature information in the article sparse feature information set to a Bert coding model to generate an article sparse feature vector, so as to obtain an article sparse feature vector set.

And 304, performing visual preference adjustment on each historical image coding vector in the historical image coding vector sequence to obtain an adjusted historical image coding vector sequence.

In some embodiments, the execution subject may perform a visual preference adjustment on each historical image encoding vector in the historical image encoding vector sequence to obtain an adjusted historical image encoding vector sequence. And the visual preference information corresponding to each adjusted historical image coding vector included in the adjusted historical image coding vector sequence is the same as the visual preference information corresponding to each historical image coding vector included in the historical image coding vector sequence. The visual preference information may characterize a visual preference characteristic of the user. In particular, the video preference characteristic of the user may be, but is not limited to, one of: the user can only beautify the video preference feature, and the user can make the video preference feature funny. The visual preference information corresponding to each adjusted historical image coding vector may be a visual preference characteristic of the user embodied by each adjusted historical image coding vector. The visual preference information corresponding to each historical image coding vector may be a visual preference characteristic of the user embodied by each historical image coding vector. The number of the historical image coding vectors included in the historical image coding vector sequence is the same as the number of the adjusted historical image coding vectors included in the adjusted historical image coding vector sequence.

As an example, the execution subject may directly input each historical image encoding vector in the historical image encoding vector sequence to the transform model to generate the adjusted historical image encoding vector sequence.

It should be noted that the visual preference feature that can be embodied by each adjusted historical image coding vector included in the adjusted historical image coding vector sequence is stronger than the visual preference feature that can be embodied by each historical image coding vector included in the historical image coding vector sequence.

In some optional implementation manners of some embodiments, the performing a visual preference adjustment on each historical image coding vector in the historical image coding vector sequence to obtain an adjusted historical image coding vector sequence may include:

firstly, visual preference information corresponding to the historical image coding vector sequence is determined and used as target visual preference information.

As an example, the execution subject may input the above-described Sequence of history image coding vectors to a Sequence to Sequence (Sequence to Sequence) model to output the target visual preference information.

And secondly, adjusting each historical image coding vector in the historical image coding vector sequence according to the target visual preference information to obtain an adjusted historical image coding vector sequence.

As an example, the execution subject may input the target visual preference information and each historical image coding vector in the historical image coding vector sequence into a Generative adaptive neural Network (GAN) model, so as to obtain an adjusted historical image coding vector sequence.

And 305, generating a recommended image set corresponding to the target recommended article to be pushed to the target user by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set.

In some embodiments, the execution subject may generate, by using a Multi-head attention-oriented (Multi-head-attention) model, a recommended image set corresponding to the target recommended item to be pushed to the target user according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set, and the item sparse feature vector set. And the recommended image set is an image subset of the creative image set corresponding to the creative image characteristic information set. . The diversified feature information that the multi-head attention mechanism model can learn may include, but is not limited to, at least one of the following: the method comprises the steps of adjusting a vector incidence relation between all adjusted historical image coding vectors in a historical image coding vector sequence after adjustment, a vector incidence relation between all user sparse feature vectors in a user sparse feature vector set, a vector incidence relation between all article sparse feature vectors in an article sparse feature vector set, a vector incidence relation between the historical image coding vector sequence after adjustment and a main image coding vector, a vector incidence relation between the historical image coding vector sequence after adjustment and the historical image coding vector sequence after adjustment, a vector incidence relation between the historical image coding vector sequence after adjustment and the article sparse feature vector set, a vector incidence relation between the main image coding vector and the user sparse feature vector set, a vector incidence relation between the main image coding vector and the article sparse feature vector set, and a vector incidence relation between the user sparse feature vector set and the article sparse feature vector set.

As an example, the execution subject may input the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set, and the article sparse feature vector set to a multi-attention mechanism model, and generate a recommended image set corresponding to the target recommended article to be pushed to the target user.

The above embodiments of the present disclosure have the following advantages: by the recommendation information generation method of some embodiments of the disclosure, accurate recommendation information is generated, and a better recommendation effect is obtained. Specifically, the reasons for the poor recommendation effect of the correlation are: the user material preference model is trained only by utilizing the historical browsing creative image sequence aiming at the target user, so that the characteristic information which can be learned by the user material preference model is limited, and the user material preference model is not accurate enough. The side results in poor recommendation. Based on this, in the recommendation information generation method of some embodiments of the present disclosure, first, a history browsing creative image sequence of a target user, a user sparse feature information set of the target user, a main image of a target recommended item, and an item sparse feature information set of the target recommended item are obtained, where the item sparse feature information set includes a creative image feature information set for the target recommended item. For subsequent retrieval of more feature information to facilitate subsequent generation of more accurate recommendation information (i.e., multiple recommended creative image scores and recommended primary image scores). Then, image coding processing is carried out on the main image to obtain a main image coding vector, and image coding processing is carried out on each history browsing creative image in the history browsing creative image sequence to generate a history image coding vector to obtain a history image coding vector sequence. Here, the main image is subjected to image encoding processing to extract feature information of the main image. In addition, the generation of recommendation information using the main image encoding vector can effectively solve the problem of large calculation amount due to large pixel dimension of the image, compared with the generation of recommendation information by the main image. Similarly, compared with the method for adjusting the visual preference through the history browsing creative image sequence, the method for adjusting the visual preference of the history image coding vector sequence can effectively solve the problem of large calculation amount. Then, each user sparse feature information in the user sparse feature information set is subjected to information coding to be converted into a vector form, so that the feature information of the user sparse features can be conveniently used. Similarly, each article sparse feature information in the article sparse feature information set is subjected to information coding to be converted into a vector form, so that the feature information of the article sparse features can be conveniently used. Furthermore, visual preference adjustment is carried out on each historical image coding vector in the historical image coding vector sequence, so that the visual preference which can be embodied by the historical image coding vector sequence after adjustment is more obvious, and more accurate recommendation information can be generated subsequently. Finally, by means of the multi-head attention mechanism model, multi-aspect feature information of the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set can be extracted. Thus, the generated recommendation information (i.e., the recommended image set) is more accurate.

With further reference to FIG. 4, a flow 400 of further embodiments of a recommendation information generation method according to the present disclosure is shown. The recommendation information generation method comprises the following steps:

step 401, obtaining a history browsing creative image sequence of a target user, a user sparse feature information set of the target user, a main image of a target recommended item, and an item sparse feature information set of the target recommended item.

Step 402, performing image coding processing on the main image to obtain a main image coding vector, and performing image coding processing on each history browsing creative image in the history browsing creative image sequence to generate a history image coding vector to obtain a history image coding vector sequence.

Step 403, performing information coding on each user sparse feature information in the user sparse feature information set to generate a user sparse feature vector to obtain a user sparse feature vector set, and performing information coding on each article sparse feature information in the article sparse feature information set to generate an article sparse feature vector to obtain an article sparse feature vector set.

And step 404, performing visual preference adjustment on each historical image coding vector in the historical image coding vector sequence to obtain an adjusted historical image coding vector sequence.

In some embodiments, specific implementations of steps 401 to 404 and technical effects brought by the same may refer to steps 301 to 304 in the embodiment corresponding to fig. 3, and are not described herein again.

And 405, carrying out vector splicing on the user sparse feature vector set and the article sparse feature vector set to obtain spliced sparse feature vectors.

In some embodiments, an executing entity (e.g., the electronic device 101 shown in fig. 1) may perform vector splicing on the user sparse feature vector set and the article sparse feature vector set to obtain a spliced sparse feature vector.

And step 406, inputting the adjusted historical image coding vector sequence, the main image coding vector and the spliced sparse feature vector into the multi-head attention mechanism model to obtain a feature information fusion vector.

In some embodiments, the execution subject may input the adjusted historical image coding vector sequence, the main image coding vector, and the stitched sparse feature vector into the multi-head attention mechanism model to obtain a feature information fusion vector. The feature information fusion vector is obtained by fusing various feature information. Wherein the feature information fusion vector may include, but is not limited to, at least one of: the vector relationship information between the historical image coding vectors in the adjusted historical image coding vector sequence, the vector relationship information between the adjusted historical image coding vector sequence and the main picture coding vector, the vector relationship information between the splicing sparse feature vector and the adjusted historical image coding vector sequence, and the vector relationship information between the splicing sparse feature vector and the main picture coding vector.

In addition, the diversified feature information that the multi-head attention mechanism model can learn can further include: importance degree information of sparse feature information (i.e., a user sparse feature vector set and an article sparse feature vector set) and importance degree information of material content information (i.e., the adjusted historical image coding vector sequence and the main image coding vector).

And 407, inputting the spliced sparse feature vector into a full-connection model to obtain an output vector.

In some embodiments, the execution subject may input the stitched sparse feature vector to a fully connected model to obtain an output vector. Wherein, the fully connected model may include a plurality of fully connected layers.

And step 408, generating the recommended image set by using a preset loss function. In some embodiments, the execution subject may generate the recommended image set in various ways by using a preset loss function. For example, the preset loss function may be a square loss function (square loss function).

In some optional implementations of some embodiments, the generating the recommended image set by using a preset loss function may include:

firstly, generating a creative image score set aiming at the splicing vector by utilizing the preset loss function. And the creative image score represents the interest degree of the target user in the creative image of the target recommended item. The creative image score sets in the creative image score sets are in one-to-one correspondence with the creative images in the creative image sets.

As an example, the execution subject may input the stitching vector to a preset loss function, resulting in a creative image score set for the stitching vector.

And secondly, generating a recommended article score corresponding to the target user by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set. The recommended item score can represent the preference degree of the target user for the target recommended item.

For example, the execution body may input the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set, and the item sparse feature vector set to a multi-attention mechanism model, and generate a recommended item score corresponding to the target user.

And thirdly, in response to the fact that the recommended item score is larger than a preset recommended item value, the execution main body can determine the creative image score of which the creative image score set score value is larger than the preset recommended creative image value, and a creative image score subset is obtained.

For example, the predetermined recommended item value is 75 points.

Fourth, the executing main body may determine the creative image set corresponding to the creative image score subset as the recommended image set.

Optionally, the steps further comprise:

and a first step of generating a recommended main image score corresponding to the target recommended article by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set. Wherein the recommended main image score represents the interest degree of the target user in the main image.

For example, the execution subject may input the adjusted historical image encoding vector sequence, the main image encoding vector, the user sparse feature vector set, and the item sparse feature vector set to a multi-attention mechanism model, and generate a recommended main image score corresponding to the target recommended item.

And secondly, in response to the fact that the recommended item score is larger than a preset recommended item value and the recommended main image score is larger than a preset recommended main image value, pushing the main image to a terminal corresponding to the target user. The terminal corresponding to the target user may be a display terminal.

For example, the predetermined recommended item value may be 70. The predetermined recommended main image numerical value may be 75.

As can be seen from fig. 4, compared with the description of some embodiments corresponding to fig. 3, the process 400 of the recommendation information generation method in some embodiments corresponding to fig. 4 may learn diversified feature information of the importance information of the sparse feature information (i.e., the user sparse feature vector set and the article sparse feature vector set) and the importance information of the material content information (i.e., the adjusted historical image coding vector sequence and the main image coding vector) by using the multi-head attention mechanism model, so as to generate a more accurate creative image set.

With further reference to fig. 5, as an implementation of the methods shown in the above figures, the present disclosure provides some embodiments of a recommendation information generating apparatus, which correspond to those shown in fig. 3, and which may be applied in various electronic devices in particular.

As shown in fig. 5, a recommendation information generating apparatus 500 includes: an acquisition unit 501, a map encoding unit 502, an information encoding unit 503, an adjustment unit 504, and a generation unit 505. The acquiring unit 501 is configured to acquire a history browsing creative image sequence of a target user, a user sparse feature information set of the target user, a main image of a target recommended item, and an item sparse feature information set of the target recommended item, where the item sparse feature information set includes a creative image feature information set for the target recommended item; a graph encoding unit 502 configured to perform graph encoding processing on the main image to obtain a main graph encoding vector, and perform graph encoding processing on each of the history browsing creative images in the history browsing creative image sequence to generate a history image encoding vector to obtain a history image encoding vector sequence; an information encoding unit 503, configured to perform information encoding on each user sparse feature information in the user sparse feature information set to generate a user sparse feature vector, obtain a user sparse feature vector set, and perform information encoding on each article sparse feature information in the article sparse feature information set to generate an article sparse feature vector, obtain an article sparse feature vector set; an adjusting unit 504, configured to perform visual preference adjustment on each historical image coding vector in the historical image coding vector sequence to obtain an adjusted historical image coding vector sequence; a generating unit 505, configured to generate, by using a multi-head attention mechanism model, a recommended image set corresponding to the target recommended item to be pushed to the target user according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set, and the item sparse feature vector set.

In some optional implementations of some embodiments, the adjusting unit 504 in the apparatus 500 may be further configured to: determining visual preference information corresponding to the historical image coding vector sequence as target visual preference information; and adjusting each historical image coding vector in the historical image coding vector sequence according to the target visual preference information to obtain an adjusted historical image coding vector sequence.

In some optional implementations of some embodiments, the generating unit 505 in the apparatus 500 described above may be further configured to: carrying out vector splicing on the user sparse feature vector set and the article sparse feature vector set to obtain spliced sparse feature vectors; inputting the adjusted historical image coding vector sequence, the main image coding vector and the splicing sparse feature vector into the multi-head attention mechanism model to obtain a feature information fusion vector; inputting the spliced sparse feature vector into a full-connection model to obtain an output vector; splicing the characteristic information fusion vector and the output vector to obtain a spliced vector; and generating the recommended image set by using a preset loss function.

In some optional implementations of some embodiments, the graph coding model includes: a residual network model and a plurality of fully connected layers; the graph encoding unit 502 in the apparatus 500 described above may be further configured to: inputting the historical browsing creative image into the residual error network model to obtain a model output result; and inputting the output result of the model to the plurality of full-connection layers to obtain the historical image coding vector.

In some optional implementations of some embodiments, the generating unit 505 in the apparatus 500 may be further configured to: generating a creative image score set aiming at the splicing vector by using the preset loss function; generating a recommended article score corresponding to the target recommended article by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set; in response to determining that the recommended item score is greater than a predetermined recommended item value, determining a creative image score with a score value in the creative image score set greater than a predetermined recommended creative image value to obtain a creative image score subset; and determining the creative image set corresponding to the creative image score subset as the recommended image set.

In some optional implementations of some embodiments, the generating unit 505 in the apparatus 500 described above may be further configured to: generating a recommended main image score corresponding to the target recommended article by utilizing a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set; and in response to the fact that the recommended item score is larger than a preset recommended item value and the recommended main image score is larger than a preset recommended main image value, pushing the main image to a terminal corresponding to the target user.

It will be understood that the units described in the apparatus 500 correspond to the various steps in the method described with reference to fig. 3. Thus, the operations, features and resulting advantages described above with respect to the method are also applicable to the apparatus 500 and the units included therein, and are not described herein again.

Referring now to FIG. 6, a block diagram of an electronic device (e.g., electronic device 101 of FIG. 1) 600 suitable for use in implementing some embodiments of the present disclosure is shown. The electronic device shown in fig. 6 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.

As shown in fig. 6, the electronic device 600 may include a processing means (e.g., central processing unit, graphics processor, etc.) 601 that may perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM) 602 or a program loaded from a storage means 608 into a Random Access Memory (RAM) 603. In the RAM603, various programs and data necessary for the operation of the electronic apparatus 600 are also stored. The processing device 601, the ROM 602, and the RAM603 are connected to each other via a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.

Generally, the following devices may be connected to the I/O interface 605: input devices 606 including, for example, a touch screen, touch pad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 607 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 608 including, for example, magnetic tape, hard disk, etc.; and a communication device 609. The communication means 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. While fig. 6 illustrates an electronic device 600 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided. Each block shown in fig. 6 may represent one device or may represent multiple devices as desired.

In particular, according to some embodiments of the present disclosure, the processes described above with reference to the flow diagrams may be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In some such embodiments, the computer program may be downloaded and installed from a network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. The computer program, when executed by the processing device 601, performs the above-described functions defined in the methods of some embodiments of the present disclosure.

It should be noted that the computer readable medium described above in some embodiments of the present disclosure may be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having 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. In some embodiments of the disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In some embodiments of the present disclosure, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.

In some embodiments, the clients, servers may communicate using any currently known or future developed network Protocol, such as HTTP (Hyper Text Transfer Protocol), and may interconnect with any form or medium of digital data communication (e.g., a communications network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the Internet (e.g., the Internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed network.

The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device. The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: acquiring a historical browsing creative image sequence of a target user, a user sparse feature information set of the target user, a main image of a target recommended article and an article sparse feature information set of the target recommended article, wherein the article sparse feature information set comprises a creative image feature information set aiming at the target recommended article; carrying out image coding processing on the main image to obtain a main image coding vector, and carrying out image coding processing on each historical browsing creative image in the historical browsing creative image sequence to generate a historical image coding vector to obtain a historical image coding vector sequence; performing information coding on each user sparse feature information in the user sparse feature information set to generate a user sparse feature vector to obtain a user sparse feature vector set, and performing information coding on each article sparse feature information in the article sparse feature information set to generate an article sparse feature vector to obtain an article sparse feature vector set; performing visual preference adjustment on each historical image coding vector in the historical image coding vector sequence to obtain an adjusted historical image coding vector sequence; and generating a recommended image set corresponding to the target recommended article to be pushed to the target user by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set.

Computer program code for carrying out operations for embodiments of the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units described in some embodiments of the present disclosure may be implemented by software or hardware. The described units may also be provided in a processor, and may be described as: a processor includes an acquisition unit, a graph encoding unit, an information encoding unit, an adjustment unit, and a generation unit. Where the names of these units do not constitute a limitation on the unit itself in some cases, for example, the acquiring unit may also be described as a "unit that acquires a sequence of historically-viewed creative images of a target user, a user sparse feature information set of the target user, a main image of a target recommended item, and an item sparse feature information set of the target recommended item".

The functions described herein above may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems on a chip (SOCs), complex Programmable Logic Devices (CPLDs), and the like.

Some embodiments of the present disclosure also provide a computer program product comprising a computer program that, when executed by a processor, implements any of the recommendation information generation methods described above.

The foregoing description is only exemplary of the preferred embodiments of the disclosure and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the invention in the embodiments of the present disclosure is not limited to the specific combination of the above-mentioned features, but also encompasses other embodiments in which any combination of the above-mentioned features or their equivalents is made without departing from the inventive concept as defined above. For example, the above features and (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure are mutually replaced to form the technical solution.

Claims

1. A recommendation information generation method includes:

acquiring a historical browsing creative image sequence of a target user, a user sparse feature information set of the target user, a main image of a target recommended article and an article sparse feature information set of the target recommended article, wherein the article sparse feature information set comprises a creative image feature information set aiming at the target recommended article;

carrying out image coding processing on the main image to obtain a main image coding vector, and carrying out image coding processing on each historical browsing creative image in the historical browsing creative image sequence to generate a historical image coding vector to obtain a historical image coding vector sequence;

performing information coding on each user sparse feature information in the user sparse feature information set to generate a user sparse feature vector to obtain a user sparse feature vector set, and performing information coding on each article sparse feature information in the article sparse feature information set to generate an article sparse feature vector to obtain an article sparse feature vector set;

performing visual preference adjustment on each historical image coding vector in the historical image coding vector sequence to obtain an adjusted historical image coding vector sequence;

and generating a recommended image set corresponding to the target recommended article to be pushed to the target user by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set, wherein the recommended image set is an image subset of an creative image set corresponding to the creative image feature information set.

2. The method of claim 1, wherein the performing a visual preference adjustment on each historical image encoding vector in the sequence of historical image encoding vectors to obtain a sequence of adjusted historical image encoding vectors comprises:

determining visual preference information corresponding to the historical image coding vector sequence as target visual preference information;

and adjusting each historical image coding vector in the historical image coding vector sequence according to the target visual preference information to obtain an adjusted historical image coding vector sequence.

3. The method according to claim 1, wherein the generating, by using a multi-head attention mechanism model, a recommended image set corresponding to the target recommended item to be pushed to the target user according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set, and the item sparse feature vector set includes:

carrying out vector splicing on the user sparse feature vector set and the article sparse feature vector set to obtain spliced sparse feature vectors;

inputting the adjusted historical image coding vector sequence, the main image coding vector and the splicing sparse feature vector into the multi-head attention mechanism model to obtain a feature information fusion vector;

inputting the spliced sparse feature vector into a full-connection model to obtain an output vector;

splicing the characteristic information fusion vector and the output vector to obtain a spliced vector;

and generating the recommended image set by using a preset loss function.

4. The method of claim 1, wherein the graph coding each historically viewed creative image of the sequence of historically viewed creative images to generate a historically image coded vector comprises:

and inputting the historical browse creative image to a pre-trained image coding model to generate a historical image coding vector.

5. The method of claim 4, wherein the graph coding model comprises: a residual network model and a plurality of fully connected layers; and

the inputting the historical browse creative image into a pre-trained image coding model to generate a historical image coding vector comprises:

inputting the historical browse creative image into the residual error network model to obtain a model output result;

and inputting the model output result to the plurality of full-connection layers to obtain the historical image coding vector.

6. The method of claim 3, wherein said generating the recommended set of images using a preset loss function comprises:

generating a creative image score set aiming at the splicing vector by utilizing the preset loss function;

generating a recommended article score corresponding to the target recommended article by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set;

in response to determining that the recommended item score is greater than a predetermined recommended item value, determining a creative image score for which the creative image score set score value is greater than a predetermined recommended creative image value, resulting in a creative image score subset;

and determining the creative image set corresponding to the creative image score subset as the recommended image set.

7. The method of claim 6, wherein the method further comprises:

generating a recommended main image score corresponding to the target recommended article by utilizing a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set;

and in response to the fact that the recommended item score is larger than a preset recommended item numerical value and the recommended main image score is larger than a preset recommended main image numerical value, pushing the main image to a terminal corresponding to the target user.

8. A recommendation information generating apparatus comprising:

an obtaining unit configured to obtain a sequence of historical browsing creative images of a target user, a user sparse feature information set of the target user, a main image of a target recommended item, and an item sparse feature information set of the target recommended item, wherein the item sparse feature information set includes a creative image feature information set for the target recommended item;

a graph encoding unit configured to perform graph encoding processing on the main image to obtain a main graph encoding vector, and to perform graph encoding processing on each of the history browsing creative images in the history browsing creative image sequence to generate a history image encoding vector to obtain a history image encoding vector sequence;

an information encoding unit configured to perform information encoding on each user sparse feature information in the user sparse feature information set to generate a user sparse feature vector to obtain a user sparse feature vector set, and perform information encoding on each article sparse feature information in the article sparse feature information set to generate an article sparse feature vector to obtain an article sparse feature vector set;

the adjusting unit is configured to perform visual preference adjustment on each historical image coding vector in the historical image coding vector sequence to obtain an adjusted historical image coding vector sequence;

and the generating unit is configured to generate a recommended image set corresponding to the target recommended article to be pushed to the target user by using a multi-head attention mechanism model according to the adjusted historical image coding vector sequence, the main image coding vector, the user sparse feature vector set and the article sparse feature vector set.

9. An electronic device, comprising:

one or more processors;

a storage device having one or more programs stored thereon,

when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-7.

10. A computer-readable medium, on which a computer program is stored, wherein the program, when executed by a processor, implements the method of any one of claims 1-7.

11. A computer program product comprising a computer program which, when executed by a processor, carries out the method of any one of claims 1-7.