Advertising material recommendation method, model training method, device and electronic equipment

Abstract

The present disclosure relates to a method for recommending advertising materials, a model training method, an apparatus, and an electronic device, and relates to the field of computer technology. The method comprises: determining multimodal features of a plurality of first advertising materials; inputting the multimodal features of the first advertising materials into a first target quality recognition model corresponding to the product to be identified to obtain a first quality score for the first advertising material; and determining an advertising material to be recommended from the plurality of first advertising materials based on the first quality score of each first advertising material in the plurality of first advertising materials. In the present disclosure, the electronic device can determine an accurate and effective quality score, which can fully and effectively reflect the differences between different advertising materials, thereby improving the accuracy of advertising material recommendations.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to an advertising material recommendation method, a model training method, a device, and an electronic device. Background Art

[0002] Currently, you can configure quality scores for different creatives, and then recommend or display corresponding creatives to users based on the quality scores.

[0003] However, the existing quality score configuration process uses a single model for all creatives, which can make it difficult to capture the unique characteristics of different products. For example, the quality score of a gaming creative might be related to the intensity of the in-game activities, while the quality score of an educational creative might be related to the quality of the teaching. Therefore, using the same model to assign quality scores to different creatives might not fully and accurately reflect the differences between creatives, which in turn affects the accuracy of creative recommendations. Summary of the Invention

[0004] The present disclosure provides an advertising material recommendation method, model training method, device and electronic device, which solves the technical problem in the prior art that the quality scores configured for different advertising materials using the same model may not fully and accurately reflect the differences between different advertising materials, thereby affecting the accuracy of advertising material recommendations.

[0005] The technical solutions of the embodiments of the present disclosure are as follows:

[0006] According to a first aspect of an embodiment of the present disclosure, a method for recommending advertising materials is provided. The method may include: determining multimodal features of a plurality of first advertising materials; inputting the multimodal features of the first advertising materials into a first target quality recognition model corresponding to a product to be identified, thereby obtaining a first quality score for the first advertising material, wherein the product to be identified is the product corresponding to the first advertising material, and parameters in the first target quality recognition model are determined based on parameters in a second target quality recognition model and an output result corresponding to the product to be identified in a target parameter recognition model; and determining an advertising material to be recommended from the plurality of first advertising materials based on the first quality score of each first advertising material in the plurality of first advertising materials.

[0007] Optionally, the above-mentioned determination of the multimodal features of multiple first advertising materials specifically includes: determining the image features of each first advertising material in the multiple first advertising materials, the text features of each first advertising material, and the identification features of each first advertising material; inputting the image features of each first advertising material, the text features of each first advertising material, and the identification features of each first advertising material into the second target quality recognition model, performing feature recognition, and obtaining the multimodal features of each first advertising material.

[0008] Optionally, the above-mentioned advertising material recommendation method also includes: determining multimodal features of multiple second advertising materials; inputting the multimodal features of the second advertising materials into a second initial quality identification model to obtain a second quality score of the second advertising material; determining a first loss, which is used to characterize the degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material; and updating the parameters in the second initial quality identification model based on the first loss to obtain the second target quality identification model.

[0009] Optionally, the above-mentioned advertising material recommendation method also includes: obtaining the product identification of the product to be identified; inputting the product identification of the product to be identified into the initial parameter identification model and outputting a first parameter; determining the parameters in the first initial quality identification model corresponding to the product to be identified based on the parameters in the second target quality identification model and the first parameter; determining the multimodal features of multiple second advertising materials corresponding to the product identification; inputting the multimodal features of the second advertising materials into the first initial quality identification model corresponding to the product to be identified to obtain the quality prediction score of the second advertising material, and the product identification corresponding to the second advertising material is the same as the product identification corresponding to the first advertising material; determining a second loss, which is used to characterize the degree of inconsistency between the quality prediction score of the second advertising material and the usage rate of the second advertising material; based on the second loss, updating the parameters in the initial parameter identification model to obtain the target parameter identification model.

[0010] Optionally, the above-mentioned advertising material recommendation method also includes: obtaining the product identification of the product to be identified; inputting the product identification of the product to be identified into the target parameter recognition model and outputting a second parameter; determining the parameters in the first target quality recognition model corresponding to the product to be identified based on the parameters in the second target quality recognition model and the second parameter.

[0011] Optionally, the target parameter recognition model includes an embedding layer and a fully connected layer. The product identification of the product to be identified is input into the target parameter recognition model, and outputting the second parameter specifically includes: inputting the product identification into the embedding layer to obtain the identification feature of the product to be identified; inputting the identification feature into the fully connected layer, and outputting the second parameter.

[0012] Optionally, the parameters in the above-mentioned second target quality identification model include a first weight and a first bias, the second parameter includes a second weight and a second bias, and the parameters in the first target quality identification model corresponding to the product to be identified include a third weight and a third bias. The above-mentioned determination of the parameters in the first target quality identification model corresponding to the product to be identified based on the parameters in the second target quality identification model and the second parameter specifically includes: determining the third weight based on the first weight and the second weight; determining the third bias based on the first bias and the second bias.

[0013] Optionally, the above-mentioned determination of the first loss specifically includes: obtaining historical recommendation data, the historical recommendation data including the number of recommendations of each second advertising creative among the multiple second advertising creatives and the number of uses of each second advertising creative, the number of uses of each second advertising creative being used to represent the number of times the target behavior occurs for each second advertising creative; determining the usage rate of each second advertising creative based on the number of recommendations of each second advertising creative and the number of uses of each second advertising creative; determining the first loss based on the usage rate of each second advertising creative and the second quality score of each second advertising creative.

[0014] According to a second aspect of an embodiment of the present disclosure, a model training method is provided. The method may include: obtaining parameters in a second target quality recognition model; determining an output result corresponding to a product to be recognized in the target parameter recognition model; and determining parameters in a first target quality recognition model corresponding to the product to be recognized based on the parameters in the second target quality recognition model and the output result, to obtain the first target quality recognition model.

[0015] Optionally, the above-mentioned model training method also includes: determining the multimodal features of multiple second advertising materials corresponding to the product identification of the product to be identified; inputting the multimodal features of the second advertising material into a second initial quality recognition model to obtain a second quality score of the second advertising material; determining a first loss, which is used to characterize the degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material; and updating the parameters in the second initial quality recognition model based on the first loss to obtain the second target quality recognition model.

[0016] Optionally, the above-mentioned model training method also includes: determining the multimodal features of multiple second advertising materials corresponding to the product identification of the product to be identified; inputting the multimodal features of the second advertising materials into the first initial quality recognition model corresponding to the product to be identified to obtain the quality prediction score of the second advertising material; determining a second loss, which is used to characterize the degree of inconsistency between the first quality score of the second advertising material and the usage rate of the second advertising material; based on the second loss, updating the parameters in the initial parameter recognition model to obtain the target parameter recognition model.

[0017] According to a third aspect of an embodiment of the present disclosure, an advertising material recommendation device is provided. The device may include: a determination module and a processing module; the determination module is configured to determine the multimodal features of a plurality of first advertising materials; the processing module is configured to input the multimodal features of the first advertising material into a first target quality recognition model corresponding to a product to be identified, to obtain a first quality score of the first advertising material, the product to be identified being the product corresponding to the first advertising material, the parameters in the first target quality recognition model being determined based on the parameters in the second target quality recognition model and the output result corresponding to the product to be identified in the target parameter recognition model; the determination module is further configured to determine the advertising material to be recommended from the plurality of first advertising materials based on the first quality score of each first advertising material in the plurality of first advertising materials.

[0018] Optionally, the determination module is specifically configured to determine the image features of each first advertising material among the multiple first advertising materials, the text features of each first advertising material, and the identification features of each first advertising material; the processing module is also configured to input the image features of each first advertising material, the text features of each first advertising material, and the identification features of each first advertising material into the second target quality recognition model to perform feature recognition to obtain the multimodal features of each first advertising material.

[0019] Optionally, the determination module is further configured to determine the multimodal features of multiple second advertising materials; the processing module is further configured to input the multimodal features of the second advertising materials into a second initial quality identification model to obtain a second quality score of the second advertising material; the determination module is further configured to determine a first loss, which is used to characterize the degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material; the processing module is further configured to update the parameters in the second initial quality identification model based on the first loss to obtain the second target quality identification model.

[0020] Optionally, the advertising material recommendation device also includes an acquisition module; the acquisition module is configured to obtain the product identification of the product to be identified; the processing module is further configured to input the product identification of the product to be identified into the initial parameter identification model and output a first parameter; the determination module is further configured to determine the parameters in the first initial quality identification model corresponding to the product to be identified based on the parameters in the second target quality identification model and the first parameter; the determination module is further configured to determine the multimodal features of multiple second advertising materials corresponding to the product identification; the processing module is further configured to input the multimodal features of the second advertising materials into the first initial quality identification model corresponding to the product to be identified to obtain a quality prediction score of the second advertising material, the product identification corresponding to the second advertising material is the same as the product identification corresponding to the first advertising material; the determination module is further configured to determine a second loss, which is used to characterize the degree of inconsistency between the quality prediction score of the second advertising material and the usage rate of the second advertising material; the processing module is further configured to update the parameters in the initial parameter identification model according to the second loss to obtain the target parameter identification model.

[0021] Optionally, the acquisition module is configured to obtain the product identification of the product to be identified; the processing module is also configured to input the product identification of the product to be identified into the target parameter identification model and output a second parameter; the determination module is also configured to determine the parameters in the first target quality identification model corresponding to the product to be identified based on the parameters in the second target quality identification model and the second parameter.

[0022] Optionally, the above-mentioned target parameter recognition model includes an embedding layer and a fully connected layer; the processing module is specifically configured to input the product identification into the embedding layer to obtain the identification feature of the product to be identified; the processing module is also specifically configured to input the identification feature into the fully connected layer and output the second parameter.

[0023] Optionally, the parameters in the above-mentioned second target quality identification model include a first weight and a first bias, the second parameter includes a second weight and a second bias, and the parameters in the first target quality identification model corresponding to the product to be identified include a third weight and a third bias; the determination module is specifically configured to determine the third weight based on the first weight and the second weight; the determination module is also specifically configured to determine the third bias based on the first bias and the second bias.

[0024] Optionally, the acquisition module is configured to acquire historical recommendation data, which includes the number of recommendations of each second advertising material among the multiple second advertising materials and the number of uses of each second advertising material, and the number of uses of each second advertising material is used to represent the number of times the target behavior occurs for each second advertising material; the determination module is specifically configured to determine the usage rate of each second advertising material based on the number of recommendations of each second advertising material and the number of uses of each second advertising material; the determination module is further specifically configured to determine the first loss based on the usage rate of each second advertising material and the second quality score of each second advertising material.

[0025] According to a fourth aspect of an embodiment of the present disclosure, a model training device is provided. The device may include: an acquisition module and a determination module; the acquisition module is configured to acquire parameters in a second target quality recognition model; the determination module is configured to determine an output result corresponding to a product to be recognized in the target parameter recognition model; the determination module is further configured to determine parameters in a first target quality recognition model corresponding to the product to be recognized based on the parameters in the second target quality recognition model and the output result, so as to obtain the first target quality recognition model.

[0026] Optionally, the model training device also includes a processing module; the determination module is further configured to determine the multimodal features of multiple second advertising materials corresponding to the product identification of the product to be identified; the processing module is configured to input the multimodal features of the second advertising material into a second initial quality recognition model to obtain a second quality score of the second advertising material; the determination module is further configured to determine a first loss, which is used to characterize the degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material; the processing module is further configured to update the parameters in the second initial quality recognition model based on the first loss to obtain the second target quality recognition model.

[0027] Optionally, the determination module is further configured to determine the multimodal features of multiple second advertising materials corresponding to the product identification of the product to be identified; the processing module is configured to input the multimodal features of the second advertising materials into the first initial quality identification model corresponding to the product to be identified to obtain the quality prediction score of the second advertising material; the determination module is further configured to determine a second loss, which is used to characterize the degree of inconsistency between the first quality score of the second advertising material and the usage rate of the second advertising material; the processing module is further configured to update the parameters in the initial parameter identification model according to the second loss to obtain the target parameter identification model.

[0028] According to a fifth aspect of an embodiment of the present disclosure, an electronic device is provided, which may include: a processor and a memory configured to store processor-executable instructions; wherein the processor is configured to execute the instructions to implement any one of the optional advertising material recommendation methods in the above-mentioned first aspect, or to implement any one of the optional model training methods in the above-mentioned second aspect.

[0029] According to a sixth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which instructions are stored. When the instructions in the computer-readable storage medium are executed by an electronic device, the electronic device is enabled to execute any one of the optional advertising material recommendation methods in the above-mentioned first aspect, or execute any one of the optional model training methods in the above-mentioned second aspect.

[0030] According to a seventh aspect of an embodiment of the present disclosure, a computer program product is provided, which includes computer instructions. When the computer instructions are executed on an electronic device, the electronic device executes any optional advertising material recommendation method as described in the first aspect, or executes any optional model training method as described in the second aspect.

[0031] The technical solutions provided by the embodiments of the present disclosure bring at least the following beneficial effects:

[0032] Based on any of the above aspects, in the present disclosure, the electronic device can determine the multimodal features of multiple first advertising materials, and input the multimodal features of the first advertising materials into the first target quality recognition model corresponding to the product to be identified to obtain the first quality score of the first advertising material; then the electronic device can determine the advertising material to be recommended from the multiple first advertising materials based on the first quality score of each first advertising material in the multiple first advertising materials. In the embodiment of the present disclosure, since the parameters in the first target quality recognition model are determined based on the parameters in the second target quality recognition model and the output results of the product to be identified in the target parameter recognition model, the output results corresponding to the first target quality recognition model corresponding to the product to be identified (specifically, the first quality scores of multiple first advertising materials) can not only characterize the quality of each first advertising material in the overall advertising market, but also characterize the differences between each first advertising material within the same product, and can determine accurate and effective quality scores, which can fully and effectively reflect the differences between different advertising materials, thereby improving the accuracy of advertising material recommendations.

[0033] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The accompanying drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description are used to explain the principles of the present disclosure, and do not constitute an improper limitation of the present disclosure.

[0035] Figure 1 A schematic diagram showing a flow chart of an advertising material recommendation method provided by an embodiment of the present disclosure;

[0036] Figure 2 A schematic diagram showing a flow chart of another advertising material recommendation method provided by an embodiment of the present disclosure;

[0037] Figure 3 A schematic diagram showing a flow chart of another advertising material recommendation method provided by an embodiment of the present disclosure;

[0038] Figure 4 A schematic diagram showing a flow chart of another advertising material recommendation method provided by an embodiment of the present disclosure;

[0039] Figure 5 A schematic diagram showing a flow chart of another advertising material recommendation method provided by an embodiment of the present disclosure;

[0040] Figure 6 A schematic diagram showing a flow chart of another advertising material recommendation method provided by an embodiment of the present disclosure;

[0041] Figure 7 A schematic diagram showing a flow chart of another advertising material recommendation method provided by an embodiment of the present disclosure;

[0042] Figure 8 A schematic diagram showing a flow chart of another advertising material recommendation method provided by an embodiment of the present disclosure;

[0043] Figure 9 A flow chart of a model training method provided by an embodiment of the present disclosure is shown;

[0044] Figure 10 A flow chart of another model training method provided by an embodiment of the present disclosure is shown;

[0045] Figure 11 A flow chart of another model training method provided by an embodiment of the present disclosure is shown;

[0046] Figure 12 A schematic structural diagram of an advertising material recommendation device provided by an embodiment of the present disclosure is shown;

[0047] Figure 13 A schematic structural diagram of another advertising material recommendation device provided by an embodiment of the present disclosure is shown;

[0048] Figure 14A schematic structural diagram of a model training device provided by an embodiment of the present disclosure is shown;

[0049] Figure 15 A structural diagram of another model training device provided by an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0050] In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0051] It should be noted that the terms "first," "second," and the like in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the numbers used in this manner are interchangeable where appropriate so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure as detailed in the appended claims.

[0052] It will also be understood that the term “comprising” indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements and / or components.

[0053] The data involved in this disclosure may be data authorized by the user or fully authorized by all parties.

[0054] As described in the background technology, since the quality scores configured for different advertising materials using the same model in the prior art may not be able to fully and accurately reflect the differences between different advertising materials, which in turn affects the accuracy of advertising material recommendations. Based on this, the embodiment of the present disclosure provides an advertising material recommendation method. Since the parameters in the first target quality recognition model corresponding to the product to be identified are determined based on the parameters in the second target quality recognition model and the output results of the product to be identified in the target parameter recognition model, the output results corresponding to the first target quality recognition model corresponding to the product to be identified (specifically, the first quality scores of multiple first advertising materials) can not only characterize the quality of each first advertising material in the overall advertising market, but also characterize the differences between each first advertising material within the same product, and can determine accurate and effective quality scores, which can fully and effectively reflect the differences between different advertising materials, thereby improving the accuracy of advertising material recommendations.

[0055] The advertising material recommendation method, model training method, device, and electronic device provided in the embodiments of the present disclosure are applicable to advertising material recommendation scenarios and / or advertising material display scenarios. When the electronic device determines the multimodal features of multiple first advertising materials, the electronic device can determine an advertising material to be recommended from the multiple first advertising materials based on the first quality score of each of the multiple first advertising materials according to the method provided in the embodiments of the present disclosure.

[0056] The following is an exemplary description of the advertising material recommendation method and model training method provided by the embodiments of the present disclosure with reference to the accompanying drawings:

[0057] Exemplarily, the electronic device that executes the advertising material recommendation method and model training method provided in the embodiments of the present disclosure may be a mobile phone, tablet computer, desktop computer, laptop computer, handheld computer, notebook computer, ultra-mobile personal computer (UMPC), netbook computer, as well as cellular phone, personal digital assistant (PDA), augmented reality (AR) or virtual reality (VR) device, etc., which can install and use content community applications. The present disclosure does not place any special restrictions on the specific form of the electronic device. It can interact with the user through one or more methods such as keyboard, touchpad, touch screen, remote control, voice interaction or handwriting device.

[0058] like Figure 1 As shown, the advertising material recommendation method provided by the embodiment of the present disclosure may include S101-S103.

[0059] S101: The electronic device determines multimodal features of a plurality of first advertising materials.

[0060] It should be understood that the multiple first advertising materials may be advertising materials corresponding to at least one product, where the at least one product is a product corresponding to (or produced by) an advertising business entity (or advertiser). The advertising business entity may entrust an advertising agency to recommend or display the advertising materials corresponding to the advertising business entity to users, specifically entrusting the advertising agency to recommend the products included in the advertising materials to users. Specifically, the electronic device may determine the multimodal characteristics of each of the multiple first advertising materials.

[0061] Optionally, the multimodal features of each first advertising material may include image features of each first advertising material and text features of each first advertising material.

[0062] S102: The electronic device inputs the multimodal features of the first advertising material into a first target quality recognition model corresponding to the product to be recognized, to obtain a first quality score of the first advertising material.

[0063] The product to be identified is the product corresponding to the first advertising material, and the parameters in the first target quality identification model are determined according to the parameters in the second target quality identification model and the output result corresponding to the product to be identified in the target parameter identification model.

[0064] It should be understood that the second target quality identification model is used to determine (or output) other quality scores (such as a second quality score) of a certain advertising material (such as the first advertising material). The second quality score can be used to characterize the quality of the first advertising material among all advertising materials (which can also be understood as the overall advertising market).

[0065] It is understood that, for different products (including the product to be identified), the target parameter identification model is used to output (or determine) the output results corresponding to each of the different products, and the different output results can represent the differences between the different products. For the same product (e.g., the product to be identified), the output results corresponding to the product to be identified in the target parameter identification model can represent the relationship between the same product (or the advertising materials corresponding to the same product, such as multiple first advertising materials), that is, they all correspond to the same product.

[0066] It can be understood that since the above-mentioned first target quality identification model is the first target quality identification model corresponding to the product to be identified, and the parameters in the first target quality identification model are determined based on the parameters in the above-mentioned second target quality identification model and the output results of the product to be identified in the target parameter identification model. In this way, the output results corresponding to the first target quality identification model corresponding to the product to be identified (specifically, the first quality scores of multiple first advertising materials) can not only characterize the quality of each first advertising material in the overall advertising market, but also characterize the differences between each first advertising material within the same product, and can accurately and effectively evaluate each first advertising material, thereby improving the effectiveness of advertising material recommendations.

[0067] S103: The electronic device determines an advertisement material to be recommended from the plurality of first advertisement materials according to the first quality score of each first advertisement material in the plurality of first advertisement materials.

[0068] Specifically, the electronic device may determine the advertising material to be recommended from the multiple first advertising materials based on the first quality score of each first advertising material in the multiple first advertising materials, and recommend the advertising material to be recommended to the user, which can also be understood as displaying the advertising material to be recommended to the user.

[0069] In an implementation manner of the embodiment of the present disclosure, the electronic device may determine M first advertising materials with the largest first quality scores among the multiple first advertising materials as advertising materials to be recommended, where M≥1.

[0070] In another implementation of the embodiment of the present disclosure, the electronic device may further determine, as the advertising material to be recommended, a first advertising material having a first quality score greater than or equal to a quality score threshold among the multiple first advertising materials.

[0071] The technical solution provided by the above embodiment can at least bring the following beneficial effects: As can be seen from S101-S103, the electronic device can determine the multimodal features of multiple first advertising materials, and input the multimodal features of the first advertising materials into the first target quality recognition model corresponding to the product to be identified to obtain the first quality score of the first advertising material; then the electronic device can determine the advertising material to be recommended from the multiple first advertising materials based on the first quality score of each first advertising material in the multiple first advertising materials. In the embodiment of the present disclosure, since the parameters in the first target quality recognition model are determined based on the parameters in the second target quality recognition model and the output result of the product to be identified in the target parameter recognition model, the output result corresponding to the first target quality recognition model corresponding to the product to be identified (specifically, the first quality scores of multiple first advertising materials) can not only characterize the quality of each first advertising material in the overall advertising market, but also characterize the differences between each first advertising material within the same product, and can determine an accurate and effective quality score, which can fully and effectively reflect the differences between different advertising materials, thereby improving the accuracy of advertising material recommendation.

[0072] Combine Figure 1 ,like Figure 2 As shown, in one implementation of the embodiment of the present disclosure, the above-mentioned determination of the multimodal features of the plurality of first advertising materials includes S1011 - S1012.

[0073] S1011. The electronic device determines an image feature of each first advertising material, a text feature of each first advertising material, and an identification feature of each first advertising material among a plurality of first advertising materials.

[0074] It should be understood that the image feature of each first advertising material can also be understood as the visual feature of each first advertising material.

[0075] In one implementation of the disclosed embodiment, if each first advertising material is a picture material, the electronic device may input the picture material into a moco model to obtain image features of the picture material. If each first advertising material is a video material, the electronic device may extract multiple frames of images from the video material, input each of the multiple frames of images into the moco model to obtain image features of the multiple frames of images, and then determine the average value of the image features of the multiple frames of images as the image feature of the video material.

[0076] In another implementation of the embodiment of the present disclosure, the electronic device can identify the text information in each first advertising material based on optical character recognition (OCR) technology (and / or identify the audio information in each first advertising material based on automatic speech recognition (ASR) technology to obtain text information corresponding to the audio information), and then input the text information into the Bert model to obtain the text features of each first advertising material.

[0077] In another implementation of the embodiment of the present disclosure, the electronic device may also obtain the product identification of the at least one product mentioned above, and may also obtain the account identification corresponding to each first advertising creative. The electronic device may then determine an identification used to characterize the uniqueness of each first advertising creative (hereinafter referred to as the identification of each first advertising creative) based on the account identification corresponding to each first advertising creative and the product identification of the product (i.e., the product to be identified) corresponding to each first advertising creative. The electronic device may randomly generate a feature vector based on the identification of each first advertising creative, that is, obtain the identification-type features of each first advertising creative. That is, the identification-type features of each first advertising creative may characterize the account identification corresponding to each first advertising creative and the product identification of the product to be identified.

[0078] S1012: The electronic device inputs the image features, text features, and identification features of each first advertising material into a second target quality recognition model to perform feature recognition to obtain multimodal features of each first advertising material.

[0079] It can be understood that the multimodal features of each first advertising material can at least represent the image features, text features and logo features of each first advertising material, that is, each first advertising material can be represented more comprehensively and completely from multiple modes and multiple directions.

[0080] The technical solution provided by the above embodiment can at least bring about the following beneficial effects: As can be seen from S1011-S1012, the electronic device can determine the image features of each first advertising material in a plurality of first advertising materials, the text features of each first advertising material, and the identification features of each first advertising material, and input the image features of each first advertising material, the text features of each first advertising material, and the identification features of each first advertising material into the second target quality recognition model for feature recognition to obtain the multimodal features of each first advertising material. In the embodiment of the present disclosure, since the multimodal features of each first advertising material can at least characterize the image features of each first advertising material, the text features of each first advertising material, and the identification features of each first advertising material, that is, the electronic device can characterize each first advertising material in the plurality of first advertising materials more comprehensively and completely from multiple modalities and multiple directions, and thus can obtain the first quality score of each first advertising material more completely and accurately.

[0081] Combine Figure 1 ,like Figure 3 As shown, the advertising material recommendation method provided by the embodiment of the present disclosure also includes S104-S107.

[0082] S104: The electronic device determines multimodal features of a plurality of second advertising materials.

[0083] It should be noted that the specific process of the electronic device determining the multimodal features of the multiple second advertising materials is the same as or similar to the explanation of the electronic device determining the multimodal features of the multiple first advertising materials, and will not be repeated here.

[0084] S105: The electronic device inputs the multimodal features of the second advertising material into a second initial quality recognition model to obtain a second quality score of the second advertising material.

[0085] In conjunction with the description of the above embodiment, it should be understood that the second quality score of an advertising material (e.g., a second advertising material) can be used to represent the quality of the second advertising material within the overall advertising material pool (which can also be understood as the overall advertising pool). This second initial quality identification model is the initial model of the above-mentioned second target quality identification model.

[0086] S106. The electronic device determines a first loss.

[0087] The first loss is used to represent the degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material.

[0088] S107: The electronic device updates parameters in the second initial quality identification model according to the first loss to obtain a second target quality identification model.

[0089] It is understood that the second target quality identification model (or second initial quality identification model) in the embodiment of the present disclosure may include a deep neural network (DNN) and a base classifier. The DNN can be used to output the multimodal features of the second advertising material, and the base classifier can be used to output the second quality score of the second advertising material. The electronic device updating the parameters in the second initial quality identification model may specifically include updating the parameters in the DNN and the parameters in the base classifier to obtain the second target quality identification model.

[0090] The technical solution provided by the above embodiment can at least bring about the following beneficial effects: As can be seen from S104-S107, the electronic device can determine the multimodal features of multiple second advertising materials, and input the multimodal features of the second advertising materials into the second initial quality recognition model to obtain the second quality score of the second advertising materials, and determine the first loss; then the electronic device can update the parameters in the second initial quality recognition model according to the first loss to obtain the second target quality recognition model. In the embodiment of the present disclosure, since the first loss can characterize the degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material, the electronic device can guide the update of the parameters in the second initial quality recognition model based on the inconsistency to obtain a second target quality recognition model with a higher accuracy in predicting the second quality score, which can improve the training efficiency of the second target quality recognition model and thereby improve the output efficiency of the first quality scores of multiple first advertising materials.

[0091] Combine Figure 3 ,like Figure 4 As shown, in one implementation of the embodiment of the present disclosure, the above-mentioned determination of the first loss includes S1061-S1063.

[0092] S1061. The electronic device obtains historical recommendation data.

[0093] The historical recommendation data includes the number of times each second advertising material in the plurality of second advertising materials is recommended and the number of times each second advertising material is used. The number of times each second advertising material is used is used to represent the number of times the target behavior occurs with each second advertising material.

[0094] It should be understood that after an electronic device recommends an advertisement to a user, the user may perform a target action on the advertisement. This target action can be understood as the action after the user clicks on and views the advertisement. For example, this target action may include filling in user personal information, downloading related applications, and making a payment.

[0095] It is understood that the number of times each second advertising material in the plurality of second advertising materials is recommended is the number of times the electronic device recommends (or displays) the second advertising material to the user. The number of times each second advertising material is used is the number of times the user uses the second advertising material (i.e., performs the target behavior with respect to the second advertising material).

[0096] S1062: The electronic device determines a usage rate of each second advertising material according to the number of times each second advertising material is recommended and the number of times each second advertising material is used.

[0097] It should be understood that the usage rate of each second advertising creative among the multiple advertising creatives is used to represent the conversion rate of each second advertising creative into the target behavior. Specifically, when the usage rate of each second advertising creative is high, it indicates that the conversion rate of each second advertising creative into the target behavior is high, that is, the user is very likely to perform the target behavior with each second advertising creative.

[0098] Optionally, the electronic device may determine the ratio between the number of times each second advertising material is used and the number of times each second advertising material is recommended as the usage rate of each second advertising material.

[0099] S1063: The electronic device determines a first loss according to the usage rate of each second advertising creative and the second quality score of each second advertising creative.

[0100] In one implementation of the embodiment of the present disclosure, the electronic device may determine that the first loss satisfies the following formula:

[0101] loss1=(ab) 2

[0102] Here, loss1 represents the first loss, a represents the usage rate of the second advertising material, and b represents the second quality score of the second advertising material.

[0103] The technical solution provided by the above embodiment can at least bring the following beneficial effects: As can be seen from S1061-S1063, the electronic device can obtain historical recommendation data, which includes the number of recommendations of each second advertising material in multiple second advertising materials and the number of uses of each second advertising material; then the electronic device determines the usage rate of each second advertising material based on the number of recommendations of each second advertising material and the number of uses of each second advertising material, and determines the first loss based on the usage rate of each second advertising material and the second quality score of each second advertising material. In the embodiment of the present disclosure, the electronic device can determine the usage rate of each second advertising material based on the historical recommendation data, that is, determine the conversion situation of the target behavior of each second advertising material, and then determine the first loss. The first loss can be accurately determined based on the actual conversion situation of each second advertising material, and the second target quality recognition model can be accurately determined.

[0104] Combine Figure 1 ,like Figure 5 As shown, the advertising material recommendation method provided by the embodiment of the present disclosure also includes S108-S114.

[0105] S108. The electronic device obtains a product identification of the product to be identified.

[0106] In combination with the description of the above embodiment, it should be understood that the product to be identified is the product corresponding to the above first advertising material.

[0107] S109: The electronic device inputs the product identification of the product to be identified into the initial parameter recognition model and outputs a first parameter.

[0108] It can be understood that the initial parameter identification model is the initial model of the above-mentioned target parameter identification model.

[0109] S110: The electronic device determines parameters in a first initial quality recognition model corresponding to the product to be recognized based on the parameters in the second target quality recognition model and the first parameters.

[0110] In conjunction with the description of the above embodiment, it should be understood that the second target quality identification model is a trained model for outputting (or predicting) a second quality score of a certain advertising material (eg, a second advertising material).

[0111] Optionally, the electronic device may determine the sum of the parameters in the second target quality recognition model and the first parameter as the parameters in the first initial quality recognition model corresponding to the product to be recognized.

[0112] S111. The electronic device determines multimodal features of a plurality of second advertising materials corresponding to a product identifier of a product to be identified.

[0113] In the present disclosure, the multiple second advertising materials corresponding to the product identification can also be understood as the multiple second advertising materials corresponding to the product to be identified.

[0114] S112: The electronic device inputs the multimodal features of the second advertising material into a first initial quality recognition model corresponding to the product to be recognized, to obtain a quality prediction score for the second advertising material.

[0115] The product identifier corresponding to the second advertising material is the same as the product identifier corresponding to the first advertising material.

[0116] It can be understood that the fact that the product identifier corresponding to the second advertising material is the same as the product identifier corresponding to the first advertising material indicates that the second advertising material and the first advertising material correspond to the same product (or product identifier), that is, the second advertising material and the first advertising material both correspond to the product to be identified.

[0117] In an embodiment of the present disclosure, each of the at least one product mentioned above (including the product to be identified) can correspond to a first initial quality recognition model, and the electronic device can input the multimodal features of a certain advertising material (for example, a second advertising material) corresponding to the product to be identified into the first initial quality recognition model corresponding to the product to be identified to obtain a second quality prediction score for the second advertising material.

[0118] In one implementation of the embodiment of the present disclosure, the above-mentioned first initial quality recognition model can be understood as an initial meta-classifier corresponding to the product to be identified. The electronic device inputs the multimodal features of the second advertising material into the initial meta-classifier to obtain a quality prediction score for the second advertising material.

[0119] S113. The electronic device determines a second loss.

[0120] The second loss is used to represent the degree of inconsistency between the quality prediction score of the second advertising material and the usage rate of the second advertising material.

[0121] In conjunction with the description of the above embodiment, it should be understood that the usage rate of the second advertising material represents the conversion rate of the second advertising material into the target behavior. Specifically, when the usage rate of the second advertising material is high, it indicates that the conversion rate of the second advertising material into the target behavior is high, that is, the user is very likely to perform the target behavior with the second advertising material.

[0122] It should be noted that the specific process of the electronic device determining the second loss is the same as or similar to the explanation of the above-mentioned electronic device determining the first loss, and will not be repeated here.

[0123] S114. The electronic device updates the parameters in the initial parameter identification model according to the second loss to obtain a target parameter identification model.

[0124] The technical solution provided by the above embodiment can at least bring the following beneficial effects: As can be seen from S108-S114, the electronic device can obtain the product identification of the product to be identified, and input the product identification of the product to be identified into the initial parameter identification model to output the first parameter; then, based on the parameters in the second target quality identification model and the first parameter, determine the parameters in the first initial quality identification model corresponding to the product to be identified. The electronic device can also determine the multimodal features of multiple second advertising materials corresponding to the product identification, and input the multimodal features of the second advertising materials into the first initial quality identification model corresponding to the product to be identified to obtain the quality prediction score of the second advertising material; then, the electronic device can determine the second loss, and update the parameters in the initial parameter identification model according to the second loss to obtain the target parameter identification model. In the embodiment of the present disclosure, the electronic device can determine the parameters in the first initial quality identification model corresponding to the product to be identified based on the parameters in the second target quality identification model and the output result (i.e., the first parameter) corresponding to the product to be identified in the initial parameter identification model, and can reasonably and effectively determine the first initial quality identification model corresponding to each product. Furthermore, because the second loss can represent the degree of inconsistency between the predicted quality score of the second advertising creative and the usage rate of the second advertising creative, the electronic device can guide the update of parameters in the initial parameter recognition model based on the degree of inconsistency to obtain a target parameter recognition model with a more accurate output result. This can improve the training efficiency of the target parameter recognition model, thereby improving the efficiency of determining the first quality score of each of the multiple first advertising creatives.

[0125] Combine Figure 1 ,like Figure 6 As shown, the advertising material recommendation method provided by the embodiment of the present disclosure also includes S115-S117.

[0126] S115. The electronic device obtains a product identification of the product to be identified.

[0127] In combination with the description of the above embodiment, it should be understood that the product to be identified is the product corresponding to the above first advertising material (or second advertising material).

[0128] S116 : The electronic device inputs the product identification of the product to be identified into the target parameter recognition model and outputs a second parameter.

[0129] In the disclosed embodiment, the target parameter recognition model can also be understood as a DNN. The electronic device inputs the product identification of one of the at least one product (e.g., the product to be recognized) into the target parameter recognition model to obtain an output result corresponding to the product to be recognized (i.e., the second parameter).

[0130] S117: The electronic device determines the parameters in the first target quality recognition model corresponding to the product to be recognized based on the parameters in the second target quality recognition model and the second parameter.

[0131] In an optional implementation, the electronic device may determine the sum of the parameter in the second target quality recognition model and the second parameter as the parameter in the first target quality recognition model corresponding to the product to be recognized.

[0132] The technical solution provided by the above embodiment can at least bring the following beneficial effects: As can be seen from S115-S117, the electronic device can obtain the product identification of the product to be identified, and input the product identification of the product to be identified into the target parameter recognition model to output the second parameter; then the electronic device can determine the parameters in the first target quality recognition model corresponding to the Korean product to be identified based on the parameters in the second target quality recognition model and the second parameter. In the embodiment of the present disclosure, the electronic device can determine the output result (i.e., the second parameter) corresponding to the product to be identified in the trained target parameter recognition model by using the product identification to be identified, and can determine the parameters in the first target quality recognition model corresponding to the product to be identified in combination with the parameters in the trained second target quality recognition model, and can determine the first target quality recognition model corresponding to the product to be identified with higher accuracy, and then based on the first target quality recognition model, can obtain or output a more accurate quality score for the advertising material, i.e., the first quality score.

[0133] Combine Figure 6 ,like Figure 7 As shown, in one implementation of the embodiment of the present disclosure, the above-mentioned target parameter recognition model includes an embedding layer and a fully connected layer. The above-mentioned input of the product identification of the product to be identified into the target parameter recognition model and output of the second parameter may specifically include S1161-S1162.

[0134] S1161. The electronic device inputs the product identification of the product to be identified into the embedding layer to obtain the identification feature of the product to be identified.

[0135] It should be understood that the embedding layer included in the target parameter recognition model is used to input the identification features of each product in the at least one product (including the product to be recognized).

[0136] S1162. The electronic device inputs the identification feature of the product to be identified into the fully connected layer and outputs a second parameter.

[0137] It can be understood that the fully connected layer included in the target parameter recognition model is used to output the output result corresponding to each product in the above-mentioned at least one product (including the product to be recognized), that is, the second parameter.

[0138] In an optional implementation, the embedding layer and the fully connected layer also include corresponding parameters. Updating the parameters in the initial parameter recognition model may specifically include updating the parameters in the embedding layer and the parameters in the fully connected layer.

[0139] The technical solution provided by the above embodiment can at least bring the following beneficial effects: As can be seen from S1161-S1162, the electronic device can input the product identification of the product to be identified into the embedding layer to obtain the identification feature of the product to be identified; and input the identification feature of the product to be identified into the fully connected layer to output the second parameter. In the embodiment of the present disclosure, the electronic device can output the identification feature of each product in at least one product based on the embedding layer, and output the corresponding output result of each product in the target parameter recognition model based on the fully connected layer. The output result corresponding to the product in the target parameter recognition model can be reasonably and effectively determined, thereby improving the effectiveness of determining the parameters in the first target quality recognition model corresponding to the product to be identified.

[0140] Combine Figure 6 ,like Figure 8 As shown, in one implementation of the embodiment of the present disclosure, the parameters in the second target quality recognition model include a first weight and a first bias, the second parameters include a second weight and a second bias, and the parameters in the first target quality recognition model corresponding to the product to be identified include a third weight and a third bias. Determining the parameters in the first target quality recognition model corresponding to the product to be identified based on the parameters in the second target quality recognition model and the second parameters may specifically include S1171-S1172.

[0141] S1171. The electronic device determines a third weight according to the first weight and the second weight.

[0142] In an optional implementation, the electronic device may determine the sum of the first weight and the second weight as the third weight.

[0143] S1172: The electronic device determines a third bias according to the first bias and the second bias.

[0144] In an optional implementation, the electronic device may determine the sum of the first bias and the second bias as the third bias.

[0145] At this point, the electronic device can determine the parameters of the first target quality recognition model corresponding to the product to be identified, that is, can obtain the first target quality recognition model corresponding to the product to be identified with higher prediction accuracy.

[0146] The technical solution provided by the above embodiment can at least bring the following beneficial effects: It can be seen from S1171-S1172 that the electronic device can determine the third weight based on the first weight and the second weight; and determine the third bias based on the first bias and the second bias. In the embodiment of the present disclosure, the electronic device can determine the weight parameter (i.e., the third weight) included in the first target quality recognition model corresponding to the product to be identified based on the weight parameter (i.e., the first weight) included in the second target quality recognition model and the weight (i.e., the second weight) output by the target parameter recognition model; and can also determine the bias parameter (i.e., the third bias) included in the first target quality recognition model corresponding to the product to be identified based on the bias parameter (i.e., the first bias) included in the second target quality recognition model and the bias (i.e., the second bias) output by the target parameter recognition model. The parameters in the first target quality recognition model corresponding to the product to be identified can be reasonably and effectively determined, that is, the first target quality recognition model corresponding to the product to be identified with higher prediction accuracy can be reasonably and effectively obtained. In addition, the first quality score of each advertising material can be accurately and effectively determined.

[0147] like Figure 9 As shown, the model method provided by the embodiment of the present disclosure may include S201-S203.

[0148] S201: The electronic device obtains parameters in a second target quality identification model.

[0149] In combination with the description of the above embodiment, the second target quality identification model is used to output a second quality score of a certain advertising material (for example, a second advertising material). The second quality score can be used to characterize the quality of the second advertising material among all advertising materials (which can also be understood as the overall advertising market).

[0150] S202: The electronic device determines an output result corresponding to the product to be identified in the target parameter identification model.

[0151] In combination with the description of the above embodiment, it should be understood that the product to be identified is the product corresponding to the above first advertising material (and the second advertising material).

[0152] S203. The electronic device determines the parameters in the first target quality recognition model corresponding to the product to be identified based on the parameters in the second target quality recognition model and the output result corresponding to the product to be identified in the target parameter recognition model to obtain the first target quality recognition model.

[0153] In one implementation of the disclosed embodiment, the parameters in the second target quality recognition model may include a first weight and a first bias, the output result corresponding to the product to be recognized in the target parameter recognition model may include a second weight and a second bias, and the parameters in the first target quality recognition model corresponding to the product to be recognized may include a third weight and a third bias. The electronic device may determine the third weight based on the first weight and the second weight, and determine the third bias based on the first bias and the second bias.

[0154] The technical solution provided by the above embodiment can at least bring the following beneficial effects: As can be seen from S201-S203, the electronic device can obtain the parameters in the second target quality recognition model and determine the output result corresponding to the product to be identified in the target parameter recognition model; then the electronic device determines the parameters in the first target quality recognition model corresponding to the product to be identified based on the parameters in the second target quality recognition model and the output result, so as to obtain the first target quality recognition model. In the embodiment of the present disclosure, the electronic device can determine the output result corresponding to the product to be identified in the trained target parameter recognition model, and combined with the parameters in the trained second target quality recognition model, it can accurately and effectively determine the parameters in the first target quality recognition model corresponding to the product to be identified, and can determine the first target quality recognition model corresponding to the product to be identified with higher accuracy. Furthermore, based on the first target quality recognition model, the electronic device can obtain or output a more accurate quality score for the advertising material to improve the accuracy of the advertising material recommendation.

[0155] Combine Figure 9 ,like Figure 10 As shown, the model training method provided by the embodiment of the present disclosure also includes S204-S207.

[0156] S204: The electronic device determines multimodal features of a plurality of second advertising materials corresponding to the product identification of the product to be identified.

[0157] S205: The electronic device inputs the multimodal features of the second advertising material into a second initial quality recognition model to obtain a second quality score of the second advertising material.

[0158] S206: The electronic device determines a first loss.

[0159] The first loss is used to represent the degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material.

[0160] S207: The electronic device updates parameters in the second initial quality identification model according to the first loss to obtain a second target quality identification model.

[0161] It should be noted that the explanations in S204-S207 are the same as or similar to the descriptions in the above S104-S107, and will not be repeated here.

[0162] The technical solution provided by the above embodiment can at least bring about the following beneficial effects: As can be seen from S204-S207, the electronic device can determine the multimodal features of multiple second advertising materials corresponding to the product identification of the product to be identified, and input the multimodal features of the second advertising material into the second initial quality recognition model to obtain the second quality score of the second advertising material, and determine the first loss; then the electronic device can update the parameters in the second initial quality recognition model according to the first loss to obtain the second target quality recognition model. In the embodiment of the present disclosure, since the first loss can characterize the degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material, the electronic device can guide the update of the parameters in the second initial quality recognition model based on the inconsistency to obtain a second target quality recognition model with a higher accuracy in predicting the second quality score, which can improve the training efficiency of the second target quality recognition model, and thereby improve the efficiency of obtaining the first target quality recognition model corresponding to the product to be identified.

[0163] Combine Figure 9 ,like Figure 11 As shown, the model training method provided by the embodiment of the present disclosure also includes S208-S211.

[0164] S208: The electronic device determines multimodal features of a plurality of second advertising materials corresponding to the product identification of the product to be identified.

[0165] S209: The electronic device inputs the multimodal features of the second advertising material into a first initial quality recognition model corresponding to the product to be recognized, to obtain a quality prediction score for the second advertising material.

[0166] S210: The electronic device determines a second loss.

[0167] The second loss is used to represent the degree of inconsistency between the first quality score of the second advertising material and the usage rate of the second advertising material.

[0168] S211. The electronic device updates the parameters in the initial parameter identification model according to the second loss to obtain a target parameter identification model.

[0169] It should be noted that the explanations in S208-S211 are the same as or similar to the descriptions in the above S111-S114, and will not be repeated here.

[0170] The technical solution provided by the above embodiment can at least bring the following beneficial effects: As can be seen from S208-S211, the electronic device can determine the multimodal features of multiple second advertising materials corresponding to the product identification of the product to be identified, and input the multimodal features of the second advertising materials into the first initial quality recognition model corresponding to the product to be identified to obtain the quality prediction score of the second advertising material; then the electronic device can determine the second loss, and update the parameters in the initial parameter recognition model according to the second loss to obtain the target parameter recognition model. In the embodiment of the present disclosure, since the second loss can characterize the degree of inconsistency between the quality prediction score of the second advertising material and the usage rate of the second advertising material, the electronic device can guide the update of the parameters in the initial parameter recognition model based on the inconsistency degree to obtain a target parameter recognition model with a more accurate corresponding output result. It can improve the training efficiency of the target parameter recognition model, and thereby improve the efficiency of obtaining the first target quality recognition model corresponding to the product to be identified.

[0171] It can be understood that in actual implementation, the electronic device described in the embodiment of the present disclosure may include one or more hardware structures and / or software modules for implementing the aforementioned corresponding advertising material recommendation method, and these execution hardware structures and / or software modules may constitute an electronic device. Those skilled in the art should easily realize that, in combination with the algorithm steps of each example described in the embodiments disclosed herein, the present disclosure can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this disclosure.

[0172] Based on this understanding, the present disclosure also provides an advertising material recommendation device. Figure 12 FIG. 1 shows a schematic diagram of the structure of the advertising material recommendation device provided by the embodiment of the present disclosure. Figure 12 As shown, the advertising material recommendation device 10 may include: a determination module 101 and a processing module 102 .

[0173] The determination module 101 is configured to determine multimodal features of a plurality of first advertising materials.

[0174] The processing module 102 is configured to input the multimodal features of the first advertising material into a first target quality recognition model corresponding to the product to be identified to obtain a first quality score of the first advertising material. The product to be identified is the product corresponding to the first advertising material. The parameters in the first target quality recognition model are determined based on the parameters in the second target quality recognition model and the output results corresponding to the product to be identified in the target parameter recognition model.

[0175] The determination module 101 is further configured to determine an advertisement material to be recommended from the plurality of first advertisement materials according to the first quality score of each first advertisement material in the plurality of first advertisement materials.

[0176] Optionally, the determination module 101 is specifically configured to determine an image feature of each first advertising material in the plurality of first advertising materials, a text feature of each first advertising material, and an identification feature of each first advertising material.

[0177] The processing module 102 is further configured to input the image features of each first advertising material, the text features of each first advertising material, and the identification features of each first advertising material into the second target quality recognition model to perform feature recognition to obtain the multimodal features of each first advertising material.

[0178] Optionally, the determination module 101 is further configured to determine multimodal features of the plurality of second advertising materials.

[0179] The processing module 102 is further configured to input the multimodal features of the second advertising material into a second initial quality recognition model to obtain a second quality score of the second advertising material.

[0180] The determination module 101 is further configured to determine a first loss, where the first loss is used to represent a degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material.

[0181] The processing module 102 is further configured to update the parameters in the second initial quality identification model according to the first loss to obtain the second target quality identification model.

[0182] Optionally, the advertising material recommendation device 20 further includes an acquisition module 103 .

[0183] The acquisition module 103 is configured to acquire the product identification of the product to be identified.

[0184] The processing module 102 is further configured to input the product identification of the product to be identified into the initial parameter recognition model and output a first parameter.

[0185] The determination module 101 is further configured to determine the parameters in the first initial quality recognition model corresponding to the product to be recognized based on the parameters in the second target quality recognition model and the first parameters.

[0186] The determination module 101 is further configured to determine multimodal features of a plurality of second advertising materials corresponding to the product identifier.

[0187] The processing module 102 is further configured to input the multimodal features of the second advertising material into the first initial quality recognition model corresponding to the product to be identified to obtain a quality prediction score of the second advertising material, where the product identifier corresponding to the second advertising material is the same as the product identifier corresponding to the first advertising material.

[0188] The determination module 101 is further configured to determine a second loss, where the second loss is used to represent a degree of inconsistency between the quality prediction score of the second advertising material and the usage rate of the second advertising material.

[0189] The processing module 102 is further configured to update the parameters in the initial parameter identification model according to the second loss to obtain the target parameter identification model.

[0190] Optionally, the acquisition module 103 is configured to acquire a product identification of the product to be identified.

[0191] The processing module 102 is further configured to input the product identification of the product to be identified into the target parameter recognition model and output a second parameter.

[0192] The determination module 101 is further configured to determine the parameters in the first target quality recognition model corresponding to the product to be recognized based on the parameters in the second target quality recognition model and the second parameter.

[0193] Optionally, the target parameter recognition model includes an embedding layer and a fully connected layer.

[0194] The processing module 102 is specifically configured to input the product identification into the embedding layer to obtain the identification feature of the product to be identified.

[0195] The processing module 102 is further configured to input the identification feature into the fully connected layer and output the second parameter.

[0196] Optionally, the parameters in the above-mentioned second target quality recognition model include a first weight and a first bias, the second parameters include a second weight and a second bias, and the parameters in the first target quality recognition model corresponding to the product to be identified include a third weight and a third bias.

[0197] The determination module 101 is specifically configured to determine the third weight according to the first weight and the second weight.

[0198] The determination module 101 is further configured to determine the third bias according to the first bias and the second bias.

[0199] Optionally, the acquisition module 103 is configured to obtain historical recommendation data, which includes the number of recommendations of each second advertising material in the multiple second advertising materials and the number of times each second advertising material is used, and the number of times each second advertising material is used is used to represent the number of times the target behavior occurs for each second advertising material.

[0200] The determination module 101 is specifically configured to determine the usage rate of each second advertising material according to the number of recommendations of each second advertising material and the number of uses of each second advertising material.

[0201] The determination module 101 is further configured to determine the first loss according to the usage rate of each second advertising material and the second quality score of each second advertising material.

[0202] As described above, the embodiment of the present disclosure can divide the advertising material recommendation device into functional modules according to the above method example. Among them, the above-mentioned integrated modules can be implemented in the form of hardware or in the form of software functional modules. In addition, it should be noted that the division of modules in the embodiment of the present disclosure is schematic and is only a logical functional division. There may be other division methods in actual implementation. For example, each functional module can be divided according to each function, or two or more functions can be integrated into one processing module.

[0203] Regarding the advertising material recommendation device in the above embodiment, the specific manner in which each module performs operations and the beneficial effects thereof have been described in detail in the aforementioned method embodiment and will not be repeated here.

[0204] Figure 13 This is a structural diagram of another advertising material recommendation device provided by the present disclosure. Figure 13 The advertising material recommendation apparatus 20 may include at least one processor 201 and a memory 203 for storing processor-executable instructions. The processor 201 is configured to execute the instructions in the memory 203 to implement the advertising material recommendation method in the above embodiment.

[0205] In addition, the advertising material recommendation apparatus 20 may further include a communication bus 202 and at least one communication interface 204 .

[0206] The processor 201 may be a central processing unit (CPU), a microprocessor unit, an ASIC, or one or more integrated circuits for controlling the execution of the program of the disclosed solution.

[0207] The communication bus 202 may include a pathway for transmitting information between the aforementioned components.

[0208] The communication interface 204 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), etc.

[0209] The memory 203 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory may exist independently and be connected to the processing unit via a bus. The memory may also be integrated with the processing unit.

[0210] The memory 203 is used to store instructions for executing the solution of the present disclosure, and the execution is controlled by the processor 201. The processor 201 is used to execute the instructions stored in the memory 203, thereby realizing the functions of the method of the present disclosure.

[0211] In a specific implementation, as an embodiment, the processor 201 may include one or more CPUs, such as Figure 13 CPU0 and CPU1 in.

[0212] In a specific implementation, as an embodiment, the advertising material recommendation device 20 may include multiple processors, such as Figure 13 201 and processor 207 in FIG. Each of these processors may be a single-CPU processor or a multi-CPU processor. A processor herein may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0213] In a specific implementation, as an embodiment, the advertising material recommendation device 20 may further include an output device 205 and an input device 206. The output device 205 communicates with the processor 201 and can display information in a variety of ways. For example, the output device 205 can be a liquid crystal display (LCD), a light emitting diode (LED) display device, a cathode ray tube (CRT) display device, or a projector. The input device 206 communicates with the processor 201 and can receive user input in a variety of ways. For example, the input device 206 can be a mouse, a keyboard, a touch screen device, or a sensor device.

[0214] Figure 14 This is a structural example diagram of a model training device provided by the present disclosure. Figure 14 As shown, the model training device 30 may include: an acquisition module 301 and a determination module 302.

[0215] The acquisition module 301 is configured to acquire parameters in the second target quality identification model.

[0216] The determination module 302 is configured to determine the output result corresponding to the product to be identified in the target parameter identification model.

[0217] The determination module 302 is further configured to determine the parameters in the first target quality recognition model corresponding to the product to be identified based on the parameters in the second target quality recognition model and the output result, so as to obtain the first target quality recognition model.

[0218] Optionally, the model training device 30 further includes a processing module 303 .

[0219] The determination module 302 is further configured to determine multimodal features of a plurality of second advertising materials corresponding to the product identification of the product to be identified.

[0220] The processing module 303 is configured to input the multimodal features of the second advertising material into a second initial quality recognition model to obtain a second quality score of the second advertising material.

[0221] The determination module 302 is further configured to determine a first loss, where the first loss is used to represent a degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material.

[0222] The processing module 303 is further configured to update the parameters in the second initial quality identification model according to the first loss to obtain the second target quality identification model.

[0223] Optionally, the determination module 302 is further configured to determine multimodal features of a plurality of second advertising materials corresponding to the product identification of the product to be identified.

[0224] The processing module 303 is configured to input the multimodal features of the second advertising material into the first initial quality recognition model corresponding to the product to be identified, and obtain a quality prediction score of the second advertising material.

[0225] The determination module 302 is further configured to determine a second loss, where the second loss is used to represent a degree of inconsistency between the first quality score of the second advertising material and the usage rate of the second advertising material.

[0226] The processing module 303 is further configured to update the parameters in the initial parameter identification model according to the second loss to obtain the target parameter identification model.

[0227] Figure 15 This is a schematic diagram of the structure of another model training device provided by the present disclosure. Figure 15 The model training device 40 may include at least one processor 401 and a memory 403 for storing processor-executable instructions. The processor 401 is configured to execute instructions in the memory 403 to implement the model training method in the above embodiment.

[0228] In addition, the model training device 40 may further include a communication bus 402 and at least one communication interface 404 .

[0229] The processor 401 may be a CPU, a microprocessor unit, an ASIC, or one or more integrated circuits for controlling the execution of the program of the disclosed solution.

[0230] The communication bus 402 may include a pathway for transmitting information between the aforementioned components.

[0231] The communication interface 404 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, RAN, WLAN, etc.

[0232] The memory 403 may be a ROM or other type of static storage device capable of storing static information and instructions, a RAM or other type of dynamic storage device capable of storing information and instructions, an EEPROM, a CD-ROM or other optical disk storage, an optical disk storage (including a compact disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. The memory may be independent and connected to the processing unit via a bus. The memory may also be integrated with the processing unit.

[0233] The memory 403 is used to store instructions for executing the solution of the present disclosure, and the execution is controlled by the processor 401. The processor 401 is used to execute the instructions stored in the memory 403, thereby realizing the functions of the method of the present disclosure.

[0234] In a specific implementation, as an embodiment, the processor 401 may include one or more CPUs, such as Figure 15 CPU0 and CPU1 in.

[0235] In a specific implementation, as an embodiment, the model training device 40 may include multiple processors, such as Figure 15 4 and 5. The processors 401 and 407 are shown in FIG. Each of these processors may be a single-CPU processor or a multi-CPU processor. A processor herein may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0236] In a specific implementation, as an embodiment, the model training device 40 may further include an output device 405 and an input device 406. The output device 405 communicates with the processor 401 and can display information in a variety of ways. For example, the output device 405 can be an LCD, LED display device, CRT display device, or projector. The input device 406 communicates with the processor 401 and can receive user input in a variety of ways. For example, the input device 406 can be a mouse, keyboard, touch screen device, or sensor device.

[0237] Those skilled in the art will understand that Figure 13 The structure shown in the figure does not constitute a limitation on the advertising material recommendation device 20, and Figure 15 The structure shown in the figure does not constitute a limitation on the model training device 40. It may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component arrangement.

[0238] In addition, the present disclosure also provides a computer-readable storage medium, including instructions, which, when executed by an electronic device, enable the electronic device to execute the advertising material recommendation method provided in the above embodiment, or execute the model training method provided in the above embodiment.

[0239] In addition, the present disclosure also provides a computer program product, including instructions, which, when executed by an electronic device, enable the electronic device to execute the advertising material recommendation method provided in the above embodiment, or execute the model training method provided in the above embodiment.

[0240] Other embodiments of the present disclosure will readily occur to those skilled in the art after considering the specification and practicing the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the present disclosure being indicated by the claims.

Claims

1. A method for recommending advertising materials, characterized in that: include: determining a multimodal feature of the plurality of first advertising creatives; Inputting the multimodal features of the first advertising creative into a first target quality recognition model corresponding to the product to be identified to obtain a first quality score for the first advertising creative, wherein the product to be identified is the product corresponding to the first advertising creative, and parameters in the first target quality recognition model are determined based on parameters in a second target quality recognition model and an output result corresponding to the product to be identified in the target parameter recognition model; the second target quality recognition model is used to determine a second quality score for the first advertising creative; The second quality score is used to represent the quality of the first advertising material among all advertising materials; The target parameter recognition model is used to determine the relationship between the multiple first advertising materials corresponding to the product to be identified; An advertising creative to be recommended is determined from the plurality of first advertising creatives according to the first quality score of each first advertising creative in the plurality of first advertising creatives.

2. The advertising material recommendation method according to claim 1, characterized in that: The determining of the multimodal features of the plurality of first advertising materials includes: Determining an image feature of each first advertising creative among the plurality of first advertising creatives, a text feature of each first advertising creative, and an identification feature of each first advertising creative; The image features of each first advertising material, the text features of each first advertising material, and the identification features of each first advertising material are input into the second target quality recognition model to perform feature recognition to obtain multimodal features of each first advertising material.

3. The advertising material recommendation method according to claim 1, characterized in that: The method further comprises: determining multimodal characteristics of the plurality of second advertising creatives; inputting the multimodal features of the second advertising material into a second initial quality recognition model to obtain a second quality score of the second advertising material; determining a first loss, the first loss being used to represent a degree of inconsistency between a second quality score of the second advertising creative and a usage rate of the second advertising creative; According to the first loss, parameters in the second initial quality identification model are updated to obtain the second target quality identification model.

4. The advertising material recommendation method according to claim 1, characterized in that: The method further comprises: Obtaining a product identification of the product to be identified; Inputting the product identification of the product to be identified into the initial parameter recognition model and outputting a first parameter; Determining parameters in a first initial quality identification model corresponding to the product to be identified based on the parameters in the second target quality identification model and the first parameters; Determining multimodal features of a plurality of second advertising materials corresponding to the product identifier; inputting the multimodal features of the second advertising material into the first initial quality recognition model corresponding to the product to be identified, to obtain a quality prediction score for the second advertising material, wherein the product identifier corresponding to the second advertising material is the same as the product identifier corresponding to the first advertising material; determining a second loss, the second loss being used to represent a degree of inconsistency between the quality prediction score of the second advertising creative and the usage rate of the second advertising creative; According to the second loss, the parameters in the initial parameter identification model are updated to obtain the target parameter identification model.

5. The advertising material recommendation method according to claim 1, characterized in that: The method further comprises: Obtaining a product identification of the product to be identified; Inputting the product identification of the product to be identified into the target parameter recognition model and outputting a second parameter; The parameters in the first target quality recognition model corresponding to the product to be recognized are determined according to the parameters in the second target quality recognition model and the second parameters.

6. The advertising material recommendation method according to claim 5, characterized in that: The target parameter recognition model includes an embedding layer and a fully connected layer. Inputting the product identification of the product to be recognized into the target parameter recognition model and outputting the second parameter includes: Inputting the product identification into the embedding layer to obtain the identification feature of the product to be identified; The identification feature is input into the fully connected layer, and the second parameter is output.

7. The advertising material recommendation method according to claim 5, characterized in that: The parameters in the second target quality recognition model include a first weight and a first bias, the second parameters include a second weight and a second bias, and the parameters in the first target quality recognition model corresponding to the product to be identified include a third weight and a third bias. Determining the parameters in the first target quality recognition model corresponding to the product to be identified based on the parameters in the second target quality recognition model and the second parameters includes: determining the third weight according to the first weight and the second weight; The third bias is determined according to the first bias and the second bias.

8. The advertising material recommendation method according to claim 3, characterized in that: The determining of the first loss includes: Acquire historical recommendation data, the historical recommendation data including the number of times each second advertising creative in the plurality of second advertising creatives is recommended and the number of times each second advertising creative is used, the number of times each second advertising creative is used being used to represent the number of times a target behavior occurs with each second advertising creative; determining a usage rate of each second advertising creative according to the number of times each second advertising creative is recommended and the number of times each second advertising creative is used; The first loss is determined according to the usage rate of each second advertising creative and the second quality score of each second advertising creative.

9. A model training method, characterized in that: include: obtaining parameters in a second target quality identification model; The second target quality identification model is used to determine a second quality score of the first advertising material corresponding to the product to be identified; The second quality score is used to represent the quality of the first advertising material among all advertising materials; Determine the output result corresponding to the product to be identified in the target parameter identification model; The target parameter recognition model is used to determine the relationship between the multiple first advertising materials corresponding to the product to be identified; According to the parameters in the second target quality recognition model and the output result, the parameters in the first target quality recognition model corresponding to the product to be recognized are determined to obtain the first target quality recognition model.

10. The model training method according to claim 9, characterized in that: The method further comprises: Determining multimodal features of a plurality of second advertising materials corresponding to the product identification of the product to be identified; inputting the multimodal features of the second advertising material into a second initial quality recognition model to obtain a second quality score of the second advertising material; determining a first loss, the first loss being used to represent a degree of inconsistency between a second quality score of the second advertising creative and a usage rate of the second advertising creative; According to the first loss, parameters in the second initial quality identification model are updated to obtain the second target quality identification model.

11. The model training method according to claim 9 or 10, characterized in that: The method further comprises: Determining multimodal features of a plurality of second advertising materials corresponding to the product identification of the product to be identified; Inputting the multimodal features of the second advertising material into the first initial quality recognition model corresponding to the product to be identified, to obtain a quality prediction score of the second advertising material; determining a second loss, the second loss being used to represent a degree of inconsistency between the first quality score of the second advertising creative and a usage rate of the second advertising creative; According to the second loss, the parameters in the initial parameter identification model are updated to obtain the target parameter identification model.

12. An advertising material recommendation device, characterized in that: include: Identify modules and process modules; The determining module is configured to determine multimodal features of a plurality of first advertising materials; The processing module is configured to input the multimodal features of the first advertising material into a first target quality recognition model corresponding to the product to be identified, to obtain a first quality score for the first advertising material, wherein the product to be identified is the product corresponding to the first advertising material, wherein parameters in the first target quality recognition model are determined based on parameters in a second target quality recognition model and an output result corresponding to the product to be identified in the target parameter recognition model; and wherein the second target quality recognition model is used to determine a second quality score for the first advertising material; The second quality score is used to represent the quality of the first advertising material among all advertising materials; the target parameter recognition model is used to determine the relationship between the multiple first advertising materials corresponding to the product to be identified; The determination module is further configured to determine an advertisement material to be recommended from the plurality of first advertisement materials according to the first quality score of each first advertisement material in the plurality of first advertisement materials.

13. The advertising material recommendation device according to claim 12, wherein: The determining module is specifically configured to determine an image feature of each first advertising material among the plurality of first advertising materials, a text feature of each first advertising material, and an identification feature of each first advertising material; The processing module is further configured to input the image features of each first advertising material, the text features of each first advertising material, and the identification features of each first advertising material into the second target quality recognition model to perform feature recognition to obtain the multimodal features of each first advertising material.

14. The advertising material recommendation device according to claim 12, wherein: The determining module is further configured to determine multimodal features of the plurality of second advertising materials; The processing module is further configured to input the multimodal features of the second advertising material into a second initial quality recognition model to obtain a second quality score of the second advertising material; The determining module is further configured to determine a first loss, where the first loss is used to represent a degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material; The processing module is further configured to update parameters in the second initial quality identification model according to the first loss to obtain the second target quality identification model.

15. The advertising material recommendation device according to claim 12, wherein: The advertising material recommendation device further includes an acquisition module; The acquisition module is configured to acquire the product identification of the product to be identified; The processing module is further configured to input the product identification of the product to be identified into an initial parameter recognition model and output a first parameter; The determining module is further configured to determine parameters in a first initial quality identification model corresponding to the product to be identified based on the parameters in the second target quality identification model and the first parameters; The determining module is further configured to determine multimodal features of a plurality of second advertising materials corresponding to the product identifier; The processing module is further configured to input the multimodal features of the second advertising material into a first initial quality recognition model corresponding to the product to be identified, to obtain a quality prediction score for the second advertising material, wherein the product identifier corresponding to the second advertising material is the same as the product identifier corresponding to the first advertising material; The determining module is further configured to determine a second loss, where the second loss is used to represent a degree of inconsistency between the quality prediction score of the second advertising material and the usage rate of the second advertising material; The processing module is further configured to update the parameters in the initial parameter identification model according to the second loss to obtain the target parameter identification model.

16. The advertising material recommendation device according to claim 12, wherein: The advertising material recommendation device further includes an acquisition module; The acquisition module is configured to acquire the product identification of the product to be identified; The processing module is further configured to input the product identification of the product to be identified into the target parameter recognition model and output a second parameter; The determination module is further configured to determine the parameters in the first target quality recognition model corresponding to the product to be identified based on the parameters in the second target quality recognition model and the second parameters.

17. The advertising material recommendation device according to claim 16, wherein: The target parameter recognition model includes an embedding layer and a fully connected layer; The processing module is specifically configured to input the product identification into the embedding layer to obtain the identification feature of the product to be identified; The processing module is further configured to input the identification feature into the fully connected layer and output the second parameter.

18. The advertising material recommendation device according to claim 16, wherein: The parameters in the second target quality recognition model include a first weight and a first bias, the second parameters include a second weight and a second bias, and the parameters in the first target quality recognition model corresponding to the product to be identified include a third weight and a third bias; The determining module is specifically configured to determine the third weight according to the first weight and the second weight; The determining module is further configured to determine the third bias according to the first bias and the second bias.

19. The advertising material recommendation device according to claim 14, wherein: The advertising material recommendation device further includes an acquisition module; The acquisition module is configured to acquire historical recommendation data, the historical recommendation data including the number of times each second advertising material in the plurality of second advertising materials is recommended and the number of times each second advertising material is used, wherein the number of times each second advertising material is used is used to represent the number of times a target behavior occurs with each second advertising material; The determining module is specifically configured to determine the usage rate of each second advertising material according to the number of recommendations of each second advertising material and the number of uses of each second advertising material; The determining module is further configured to determine the first loss according to the usage rate of each second advertising material and the second quality score of each second advertising material.

20. A model training device, characterized in that: include: Get module and determine module; The acquisition module is configured to acquire parameters in the second target quality identification model; The second target quality identification model is used to determine a second quality score of the first advertising material corresponding to the product to be identified; The second quality score is used to represent the quality of the first advertising material among all advertising materials; The determination module is configured to determine the output result corresponding to the product to be identified in the target parameter identification model; The target parameter recognition model is used to determine the relationship between the multiple first advertising materials corresponding to the product to be identified; The determination module is further configured to determine the parameters in the first target quality recognition model corresponding to the product to be identified based on the parameters in the second target quality recognition model and the output result, so as to obtain the first target quality recognition model.

21. The model training device according to claim 20, characterized in that: The model training device also includes a processing module; The determining module is further configured to determine multimodal features of a plurality of second advertising materials corresponding to the product identification of the product to be identified; The processing module is configured to input the multimodal features of the second advertising material into a second initial quality recognition model to obtain a second quality score of the second advertising material; The determining module is further configured to determine a first loss, where the first loss is used to represent a degree of inconsistency between the second quality score of the second advertising material and the usage rate of the second advertising material; The processing module is further configured to update parameters in the second initial quality identification model according to the first loss to obtain the second target quality identification model.

22. The model training device according to claim 20 or 21, characterized in that: The model training device also includes a processing module; The determining module is further configured to determine multimodal features of a plurality of second advertising materials corresponding to the product identification of the product to be identified; The processing module is configured to input the multimodal features of the second advertising material into the first initial quality recognition model corresponding to the product to be identified, to obtain a quality prediction score of the second advertising material; The determining module is further configured to determine a second loss, where the second loss is used to represent a degree of inconsistency between the first quality score of the second advertising material and the usage rate of the second advertising material; The processing module is further configured to update the parameters in the initial parameter identification model according to the second loss to obtain the target parameter identification model.

23. An electronic device, characterized in that: The electronic device comprises: processor; a memory configured to store instructions executable by the processor; The processor is configured to execute the instructions to implement the advertising material recommendation method according to any one of claims 1 to 8, or to implement the model training method according to any one of claims 9 to 11.

24. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions in the computer-readable storage medium are executed by an electronic device, the electronic device is enabled to execute the advertising material recommendation method according to any one of claims 1 to 8, or execute the model training method according to any one of claims 9 to 11.

25. A computer program product, characterized in that The computer program product includes computer instructions, and when the computer instructions are executed on an electronic device, the electronic device executes the advertising material recommendation method according to any one of claims 1 to 8, or executes the model training method according to any one of claims 9 to 11.

Images