Material sequence generation method, device, equipment, medium and program product

Abstract

This disclosure presents embodiments of a method, apparatus, device, medium, and program product for generating material sequences. One specific implementation of the method includes: in response to the existence of target creative materials in a set of creative materials to be recommended, determining a target set of creative materials to be recommended; for each target creative material to be recommended, selecting a predetermined number of target historical recommended creative materials from a set of historical recommended creative materials; generating an estimated click-through rate (CTR) for each target creative material to be recommended based on the CTR of each target historical recommended creative material; determining the CTR of each multiple recommended creative material; and sorting the creative materials to be recommended to obtain a sequence of creative materials to be recommended. This implementation is related to artificial intelligence; by determining the CTR corresponding to the target creative materials to be recommended, a more accurate material recommendation order can be generated for the set of creative materials to be recommended, improving the material recommendation effect.

Description

Technical Field

[0001] The embodiments disclosed herein relate to the field of computer technology, and specifically to methods, apparatus, devices, media, and program products for generating material sequences. Background Technology

[0002] Currently, content recommendation platforms often showcase the characteristics of various items by pushing content (such as creative images of items). For newly generated content, the common approach is to use a greedy algorithm to set specific thresholds for newly generated content, ensuring that the subsequent content recommendation platform gives it sufficient exposure.

[0003] However, the inventors discovered that when using the above method to push newly generated materials, the following technical problems often occur:

[0004] Giving newly generated content ample exposure without clearly identifying whether it is high-quality, high-potential creative material may result in poor actual content recommendation performance.

[0005] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0007] Some embodiments of this disclosure provide methods, apparatus, devices, media, and program products for generating material sequences to address the technical problems mentioned in the background section above.

[0008] In a first aspect, some embodiments of this disclosure provide a method for generating a material sequence, comprising: in response to the existence of a target creative material in an acquired set of creative materials to be recommended, determining a target set of creative materials to be recommended, wherein the target creative material to be recommended is a creative material that satisfies the determination conditions of the target material; for each target creative material in the aforementioned set of target creative materials to be recommended, filtering a predetermined number of target historical recommended creative materials from an acquired set of historical recommended creative materials to obtain a set of target historical recommended creative materials, wherein the material correlation between the target historical recommended creative materials and the aforementioned target creative material to be recommended is greater than a predetermined threshold; and based on the obtained... The click-through rates (CTRs) of each target historical recommended creative material in the target historical recommended creative material set are used to generate the estimated CTRs of each target recommended creative material in the target recommended creative material set. The CTRs of each multiple recommended creative material in the multiple recommended creative material set are determined, where the multiple recommended creative material set is the material set in the target recommended creative material set excluding the target recommended creative material set. Based on the estimated CTR set corresponding to the target recommended creative material set and the CTR set corresponding to the multiple recommended creative material set, the recommended creative materials in the target recommended creative material set are sorted to obtain a sequence of recommended creative materials.

[0009] Optionally, the aforementioned preset material conditions include: multiple preset material sub-conditions, the multiple preset material sub-conditions including: the time interval between the material generation time corresponding to the creative material to be recommended and the current time is less than a predetermined duration, and the exposure amount corresponding to the creative material to be recommended is less than a predetermined exposure value; and the aforementioned determination of the target set of creative materials to be recommended includes: determining the material generation time and exposure amount corresponding to each creative material to be recommended in the aforementioned set of creative materials to be recommended; and determining the creative materials to be recommended in the aforementioned set of creative materials to be recommended that satisfy the multiple preset material sub-conditions as the target creative materials to be recommended.

[0010] Optionally, the above-mentioned process of selecting a predetermined number of target historical recommended creative materials from the acquired historical recommended creative material set to obtain a target historical recommended creative material set includes: inputting each historical recommended creative material in the historical recommended creative material set into a pre-trained material encoding model to generate historical material encoding vectors, thereby obtaining a historical material encoding vector set; inputting the target creative material to be recommended into the material encoding model to generate a material encoding vector to be recommended; and, based on the historical material encoding vector set and the material encoding vector to be recommended, selecting historical recommended creative materials from the historical recommended creative material set whose vector distance satisfies a predetermined distance condition, as target historical recommended creative materials, thereby obtaining a target historical recommended creative material set, wherein the vector distance is the vector distance between the historical material encoding vector in the historical material encoding vector set and the material encoding vector to be recommended.

[0011] Optionally, the above-mentioned generation of the estimated click-through rate of each target recommended creative material in the target recommended creative material set based on the click-through rate of each target historical recommended creative material in the obtained target historical recommended creative material set includes: for each target recommended creative material in the target recommended creative material set, calculating the click-through rate of each target historical recommended creative material in the target historical recommended creative material set corresponding to the target recommended creative material, and obtaining a calculated value as the estimated click-through rate corresponding to the target recommended creative material.

[0012] Optionally, the above-mentioned material encoding model is trained through the following steps: obtaining a training sample set, wherein the training samples include: training creative materials and product term tags; using the above-mentioned training sample set, training the initial product term classification model to obtain the trained product term classification model, wherein the trained product term classification model includes: the material encoding model.

[0013] Optionally, the above method further includes: in response to the absence of a target creative material in the above set of creative materials to be recommended, inputting each creative material to be recommended in the above set of creative materials to be recommended into the material click-through rate generation model to output the material click-through rate and obtain a material click-through rate set; and sorting the various creative materials to be recommended included in the above set of creative materials to be recommended according to the material click-through rate set to obtain a sequence of creative materials to be recommended.

[0014] Optionally, the above method further includes sending at least one creative material from the above sequence of creative materials to be recommended, whose click-through rate meets the preset click-through rate condition, to the material recommendation terminal.

[0015] Secondly, some embodiments of this disclosure provide a material sequence generation apparatus, including: a first determining unit configured to determine a target creative material set in response to the existence of a target creative material in an acquired set of creative materials to be recommended, wherein the target creative material is a creative material that satisfies the determination conditions of the target material; a filtering unit configured to filter a predetermined number of target historical recommended creative materials from an acquired set of historical recommended creative materials for each target creative material in the target set of creative materials to be recommended, to obtain a target historical recommended creative material set, wherein the material correlation between the target historical recommended creative materials and the target creative material is greater than a predetermined threshold; and a generation unit configured to... Based on the click-through rates (CTRs) of each target historical recommended creative material in the obtained target historical recommended creative material set, an estimated CTR for each target recommended creative material in the aforementioned target recommended creative material set is generated; the second determining unit is configured to determine the CTR of each multiple recommended creative material in the multiple recommended creative material set, wherein the aforementioned multiple recommended creative material set is the material set in the aforementioned recommended creative material set excluding the aforementioned target recommended creative material set; the sorting unit is configured to sort each recommended creative material included in the recommended creative material set according to the estimated CTR set corresponding to the target recommended creative material set and the CTR set corresponding to the multiple recommended creative material set, to obtain a recommended creative material sequence.

[0016] Optionally, the aforementioned preset material conditions include: multiple preset material sub-conditions, the multiple preset material sub-conditions including: the time interval between the material generation time corresponding to the creative material to be recommended and the current time is less than a predetermined duration, and the exposure amount corresponding to the creative material to be recommended is less than a predetermined exposure value; and the first determining unit can be configured to: determine the material generation time and exposure amount corresponding to each creative material to be recommended in the aforementioned set of creative materials to be recommended; and determine the creative materials to be recommended in the aforementioned set of creative materials to be recommended that satisfy the multiple preset material sub-conditions as target creative materials to be recommended.

[0017] Optionally, the filtering unit can be configured to: input each historical recommended creative material in the aforementioned historical recommended creative material set into a pre-trained material encoding model to generate historical material encoding vectors, thus obtaining a historical material encoding vector set; input the aforementioned target creative material to be recommended into the aforementioned material encoding model to generate a material encoding vector to be recommended; and, based on the aforementioned historical material encoding vector set and the aforementioned material encoding vector to be recommended, filter historical recommended creative materials whose vector distance satisfies a predetermined distance condition from the aforementioned historical recommended creative material set, as target historical recommended creative materials, thus obtaining a target historical recommended creative material set, wherein the vector distance is the vector distance between the historical material encoding vector in the aforementioned historical material encoding vector set and the aforementioned material encoding vector to be recommended.

[0018] Optionally, the generation unit can be configured to: for each target creative material in the target creative material set to be recommended, calculate the click-through rate of each target historical creative material in the target historical creative material set corresponding to the target creative material to be recommended, and obtain a calculated value as the estimated click-through rate corresponding to the target creative material to be recommended.

[0019] Optionally, the above-mentioned material encoding model is trained through the following steps: obtaining a training sample set, wherein the training samples include: training creative materials and product term tags; using the above-mentioned training sample set, training the initial product term classification model to obtain the trained product term classification model, wherein the trained product term classification model includes: the material encoding model.

[0020] Optionally, the above apparatus further includes: in response to the absence of a target creative material in the set of creative materials to be recommended, inputting each creative material to be recommended in the set of creative materials to be recommended into the material click-through rate generation model to output the material click-through rate and obtain a set of material click-through rates; and sorting the various creative materials to be recommended included in the set of creative materials to be recommended according to the set of material click-through rates to obtain a sequence of creative materials to be recommended.

[0021] Optionally, the above-mentioned device further includes: sending at least one creative material from the above-mentioned creative material sequence that meets the preset click-through rate condition to the material recommendation terminal.

[0022] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, such that when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any implementation of the first aspect.

[0023] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method as described in any implementation of the first aspect.

[0024] Fifthly, some embodiments of this disclosure provide a computer program product, including a computer program that, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0025] The above embodiments of this disclosure have the following beneficial effects: By determining the click-through rate (CTR) corresponding to the target creative material to be recommended, the material sequence generation method of some embodiments of this disclosure can generate a more accurate material recommendation order for the set of creative materials to be recommended, thereby improving the material recommendation effect. Specifically, the reason for poor material recommendation effect is that it is impossible to determine whether the newly generated material is a high-quality, high-potential creative material. Directly giving the newly generated material sufficient exposure may lead to poor actual material recommendation effect. Based on this, the material sequence generation method of some embodiments of this disclosure firstly, in response to the existence of target creative materials to be recommended in the acquired set of creative materials to be recommended, determines the target set of creative materials to be recommended from the set of creative materials to be recommended, so as to generate the CTR corresponding to each target creative material to be recommended in the subsequent process. Then, for each target creative material to be recommended in the above target set of creative materials to be recommended, a predetermined number of target historical creative materials are selected from the acquired set of historical recommended creative materials to obtain the target historical recommended creative material set. Here, since the target creative materials to be recommended do not have a real CTR, it is impossible to determine whether the target creative materials to be recommended are high-quality, high-potential creative materials. Therefore, by selecting a predetermined number of target historical recommended creative materials from the historical recommended creative materials, based on their correlation with the target recommended creative materials, the click-through rate (CTR) of the target recommended creative materials can be estimated subsequently. Next, based on the CTR of each target historical recommended creative material in the obtained target historical recommended creative material set, the estimated CTR of each target recommended creative material in the aforementioned target recommended creative material set can be estimated relatively accurately. Furthermore, the CTR of each multiple recommended creative material in the multiple recommended creative material set is determined for use in the subsequent sorting of materials in the recommended creative material set. Finally, based on the estimated CTR set corresponding to the target recommended creative material set and the CTR set corresponding to the multiple recommended creative material set, the materials included in the recommended creative material set are sorted to obtain the recommended creative material sequence. Here, the estimated CTR of the target recommended creative material is used as the standard to measure the quality of the creative materials. The CTR of the multiple recommended creative materials is used to measure the quality of the multiple recommended creative materials. Therefore, by ranking the recommended creative materials based on the click-through rate of each creative material in the set to be recommended, the materials can be sorted according to their quality. Using a sequence of creative materials to recommend individual materials can significantly improve the effectiveness of material recommendation. Attached Figure Description

[0026] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0027] Figures 1-2 This is a schematic diagram illustrating an application scenario of a material sequence generation method according to some embodiments of the present disclosure;

[0028] Figure 3 This is a flowchart of some embodiments of the material sequence generation method according to this disclosure;

[0029] Figure 4 These are flowcharts of other embodiments of the material sequence generation method according to this disclosure;

[0030] Figure 5 These are schematic diagrams illustrating the structure of some embodiments of the material sequence generation apparatus according to this disclosure;

[0031] Figure 6 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0032] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0033] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0034] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0035] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0036] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0037] Before performing any of the creative materials (such as creative images and videos of objects) involved in this disclosure, the relevant organizations or individuals shall fulfill their obligations, including conducting a security impact assessment of the materials, informing the material creator, and obtaining prior authorization and consent from the material creator.

[0038] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0039] Figures 1-2 This is a schematic diagram of an application scenario of a material sequence generation method according to some embodiments of the present disclosure.

[0040] exist Figures 1-2In this application scenario, firstly, in response to the existence of target creative materials in the acquired set of creative materials to be recommended 101, the electronic device 101 can determine the target set of creative materials to be recommended. The target creative materials to be recommended are creative materials that meet the criteria for determining target materials. In this application scenario, the set of creative materials to be recommended 101 may include: creative materials to be recommended 1011, creative materials to be recommended 1012, creative materials to be recommended 1013, and creative materials to be recommended 1014. The target creative material to be recommended may be creative material 1013. The criteria for determining target materials may be that the content corresponding to the target creative material is a triangle. Then, for each target creative material in the above target set of creative materials to be recommended, the electronic device 101 can filter a predetermined number of target historical recommended creative materials from the acquired set of historical recommended creative materials 102 to obtain a target historical recommended creative material set. The correlation between the target historical recommended creative materials and the above target creative materials to be recommended is greater than a predetermined threshold. In this application scenario, the historical recommended creative material set 102 includes: historical recommended creative material 1021, historical recommended creative material 1022, historical recommended creative material 1023, and historical recommended creative material 1024. The target creative material to be recommended is creative material 1013, and the predetermined number can be two. The corresponding target historical recommended creative material set includes: historical recommended creative material 1022 and historical recommended creative material 1024. Next, the electronic device 101 can generate the estimated click-through rate (CTR) of each target creative material in the aforementioned target creative material set based on the CTR of each target historical recommended creative material in the obtained target historical recommended creative material set. In this application scenario, for the target creative material to be recommended, creative material 1013, the corresponding target historical recommended creative material set includes: historical recommended creative material 1022 and historical recommended creative material 1024. The CTR 106 corresponding to historical recommended creative material 1022 can be "0.3". The CTR 107 corresponding to historical recommended creative material 1024 can be "0.5". The average click-through rates (CTRs) 106 and 107 are calculated to obtain the average CTR, which is used as the estimated CTR 108 for the creative material to be recommended 1013. The estimated CTR 108 can be "0.4". Furthermore, the electronic device 101 can determine the CTR of each creative material in the multiple recommendation creative material set. The multiple recommendation creative material set is the set of creative materials in the creative material set to be recommended minus the target creative material set. In this application scenario, the multiple recommendation creative material set includes: creative material to be recommended 1011, creative material to be recommended 1012, and creative material to be recommended 1014. The CTR 103 of creative material to be recommended 1011 can be "0.45". The CTR 104 of creative material to be recommended 1012 can be "0.6".The click-through rate (CTR) 105 of the creative material to be recommended 1014 can be "0.1". Finally, the electronic device 101 can sort the creative materials included in the creative material set 101 according to the estimated CTR set corresponding to the target creative material set and the CTR set corresponding to the multiple recommended creative material sets, to obtain the creative material sequence 109. In this application scenario, the creative material set 101 is sorted according to the CTR of the creative materials to be recommended in ascending order to obtain the creative material sequence 109. The creative material sequence 109 can be [creative material to be recommended 1014, creative material to be recommended 1013, creative material to be recommended 1011, creative material to be recommended 1012].

[0041] It should be noted that the aforementioned electronic device 101 can be either hardware or software. When the electronic device is hardware, it can be implemented as a distributed cluster consisting of multiple servers or terminal devices, or as a single server or a single terminal device. When the electronic device is software, it can be installed in the hardware devices listed above. It can be implemented as, for example, multiple software programs or software modules used to provide distributed services, or as a single software program or software module. No specific limitations are made here.

[0042] It should be understood that Figures 1-2 The number of electronic devices shown is merely illustrative. Any number of electronic devices can be used depending on the implementation requirements.

[0043] Continue to refer to Figure 3 The diagram illustrates a flow 300 of some embodiments of a material sequence generation method according to the present disclosure. This material sequence generation method includes the following steps:

[0044] Step 301: In response to the existence of target creative materials in the acquired set of creative materials to be recommended, determine the target set of creative materials to be recommended.

[0045] In some embodiments, in response to the existence of target creative materials to be recommended in the acquired set of creative materials to be recommended, the execution entity of the above-mentioned material sequence generation method (e.g., Figure 1The electronic device 101 shown can determine a target set of creative materials to be recommended. The target creative materials to be recommended are creative materials that meet the criteria for determining the target materials. The criteria for determining the target materials can be conditions used to determine whether the creative materials to be recommended are new materials (i.e., newly generated materials). The target creative materials to be recommended are newly generated creative materials to be recommended. In practice, the criteria for determining the target materials vary depending on the criteria used to judge newly generated creative materials to be recommended. For example, if the criterion for judging newly generated creative materials to be recommended is the number of recommendations, the target material condition can be that the number of recommendations for the target creative materials to be recommended is less than a predetermined number. As another example, if the criterion for judging newly generated creative materials to be recommended is the duration of material placement, the target material condition can be that the duration of material placement for the target creative materials to be recommended is less than a predetermined duration. The aforementioned set of creative materials to be recommended can be at least one creative material to be recommended. In e-commerce scenarios, creative materials can be creative product materials designed for specific items. The aforementioned creative materials can be in the form of images or videos.

[0046] It should be noted that when the target creative material to be recommended appears, the material recommendation platform will be in a cold start period to determine how to recommend the target creative material.

[0047] As an example, if the recommendation criteria for the target creative material are that the number of recommendations for the target creative material is less than a predetermined number, the aforementioned implementing entity can determine whether the target creative material exists in the acquired set of creative materials to be recommended through the following steps:

[0048] The first step is to determine the number of times each creative material in the set of creative materials to be recommended is recommended, thus obtaining the set of material recommendation counts.

[0049] The second step is to determine that, in response to the determination that there are materials with a recommendation count of less than "1" in the above-mentioned material recommendation count set, to determine that there are target creative materials to be recommended in the creative material set to be recommended.

[0050] In some optional implementations of certain embodiments, the preset material conditions include: multiple preset material sub-conditions. These multiple preset material sub-conditions include: the time interval between the material generation time corresponding to the creative material to be recommended and the current time is less than a predetermined duration; and the exposure amount corresponding to the creative material to be recommended is less than a predetermined exposure value. Furthermore, determining the target set of creative materials to be recommended may include the following steps:

[0051] The first step is to determine the generation time and exposure volume of each creative material in the above set of creative materials to be recommended.

[0052] For example, the set of creative materials to be recommended includes: the first set of creative materials to be recommended, the second set of creative materials to be recommended, and the third set of creative materials to be recommended. The first set of creative materials to be recommended was created on August 1, 2021. The second set of creative materials to be recommended was created on September 1, 2021. The third set of creative materials to be recommended was created on April 15, 2022. The first set of creative materials to be recommended has been viewed 1,000 times. The second set of creative materials to be recommended has been viewed 500 times. The third set of creative materials to be recommended has been viewed 3,000 times.

[0053] The second step is to identify the creative materials that meet multiple preset material sub-conditions as the target creative materials to be recommended.

[0054] For example, the planned duration is 1 month, the planned exposure value is 2500, and the current time is May 1, 2022. Therefore, the time interval for the first recommended creative material is 8 months, the time interval for the second recommended creative material is 7 months, and the time interval for the third recommended creative material is half a month. Since the third recommended creative material has 3000 exposures, it meets multiple preset sub-conditions. Therefore, the third recommended creative material is selected as the target recommended creative material.

[0055] Step 302: For each target creative material in the target creative material set to be recommended, select a predetermined number of target historical recommended creative materials from the obtained historical recommended creative material set to obtain the target historical recommended creative material set.

[0056] In some embodiments, the executing entity may, for each target creative material in the target creative material set to be recommended, filter a predetermined number of target historical recommended creative materials from the acquired historical recommended creative material set to obtain a target historical recommended creative material set. The correlation between the target historical recommended creative materials and the target creative materials to be recommended is greater than a predetermined threshold. This correlation characterizes the similarity between the target historical recommended creative materials and the target creative materials to be recommended. For example, the similarity can be one of the following: content similarity, style similarity. Historical recommended creative materials can be creative materials that have been recommended multiple times in the past by the material recommendation end. The material recommendation end can be a terminal that recommends materials. In practice, the historical recommended creative material set can be multiple creative materials recommended by the material recommendation end within a predetermined historical time period. For example, the current time is May 1, 2022. The predetermined historical time period can be January 1, 2021 to March 1, 2021.

[0057] As an example, firstly, the aforementioned execution entity can input each historical recommended creative material from the historical recommended creative material set into a creative material style determination model to generate a historical creative material style. This creative material style determination model can be a model for determining material styles. For example, it could be a multi-layer convolutional neural network (CNN) model. Then, the target creative material to be recommended is input into the aforementioned creative material style determination model to generate a style for the target creative material. Finally, historical recommended creative materials with the same style as the target creative material are selected from the historical recommended creative material set and used as the target historical recommended creative materials, thus obtaining the target historical recommended creative material set.

[0058] In some optional implementations of certain embodiments, the process of filtering a predetermined number of target historical recommended creative materials from the acquired historical recommended creative material set to obtain the target historical recommended creative material set may include the following steps:

[0059] The first step is to input each historical recommended creative material from the aforementioned historical recommended creative material set into a pre-trained material encoding model to generate historical material encoding vectors, thus obtaining a set of historical material encoding vectors.

[0060] The material encoding model can be a model that encodes materials using vectors. Historical material encoding vectors can represent the material feature information of historically recommended creative materials.

[0061] Here, the aforementioned execution entity uses the historical material encoding vector as the index and the corresponding historical recommended creative material as the value, and stores them in the faiss index library.

[0062] Optionally, the aforementioned material encoding model includes: a residual network model and multiple serial fully connected layers; and the above-mentioned inputting each historical recommended creative material from the aforementioned historical recommended creative material set into the pre-trained material encoding model to generate a historical material encoding vector may include the following steps:

[0063] Step 1: Input the above-mentioned historical recommended creative materials into the above residual network model to obtain the model output results.

[0064] Step 2: Input the output of the above model into the above multiple serial fully connected layers to obtain the above historical material encoding vector.

[0065] The second step is to input the target creative materials to be recommended into the material coding model to generate the material coding vector to be recommended.

[0066] Among them, the encoding vector of the material to be recommended can represent the material feature information of the target creative material to be recommended.

[0067] Here, the aforementioned execution entity uses the encoding vector of the material to be recommended as the index and the corresponding target creative material to be recommended as the value, and stores them in the faiss index library.

[0068] The third step involves selecting historical recommended creative materials from the historical recommended creative material set based on the aforementioned historical material encoding vector set and the aforementioned material encoding vector to be recommended, and using these as target historical recommended creative materials to obtain the target historical recommended creative material set.

[0069] The vector distance is the distance between the historical material coding vectors in the aforementioned historical material coding vector set and the coding vector of the material to be recommended. The predetermined distance condition can be that the vector distance between the coding vectors of the historical recommended creative material and the material to be recommended is greater than a predetermined distance value. For example, the predetermined distance value could be 0.5.

[0070] As an example, firstly, the executing entity can determine the cosine distance between each historical material encoding vector in the historical material encoding vector set and the aforementioned material encoding vector to be recommended, as the vector distance. Then, it selects historical recommended creative materials from the aforementioned historical recommended creative material set whose vector distance meets the predetermined distance condition, as the target historical recommended creative materials, thus obtaining the target historical recommended creative material set.

[0071] Optionally, the above-mentioned material encoding model is trained through the following steps:

[0072] The first step is to obtain the training sample set.

[0073] The training samples include: training creative materials and product tag names.

[0074] The second step is to use the above training sample set to train the initial product word classification model, and obtain the trained product word classification model.

[0075] The trained item / product term classification model includes a material encoding model. This model can be used to determine the item / product terms corresponding to the creative material. The item / product term classification model may include a material encoding model and multiple cascaded fully connected layers.

[0076] Step 303: Based on the click-through rates of each target historical recommended creative material in the obtained target historical recommended creative material set, generate the estimated click-through rates of each target recommended creative material in the above target recommended creative material set.

[0077] In some embodiments, the aforementioned executing entity can generate the estimated click-through rate (CTR) of each target recommended creative material in the aforementioned target recommended creative material set based on the CTR of each target historical recommended creative material in the obtained target historical recommended creative material set. Each target historical recommended creative material has a corresponding CTR.

[0078] As an example, firstly, the executing entity can determine the maximum and minimum click-through rates (CTRs) of each target historical recommended creative material within the target historical recommended creative material set. Then, the maximum and minimum CTRs are removed from the CTR set corresponding to the target historical recommended creative material set, resulting in a CTR set after removal. Finally, the average CTR of each item in the removed CTR set is determined as the CTR corresponding to the existence of the target historical recommended creative material.

[0079] Step 304: Determine the click-through rate of each creative material in the multiple recommended creative material set.

[0080] In some embodiments, the executing entity may determine the click-through rate of each multiple-recommendation creative material in the multiple-recommendation creative material set. The multiple-recommendation creative material set is the set of creative materials in the set to be recommended excluding the target creative material set.

[0081] As an example, the aforementioned entity can query the click-through rate of each multiple recommended creative material in the multiple recommended creative material set in the database.

[0082] Step 305: Based on the estimated click-through rate set corresponding to the target set of creative materials to be recommended and the click-through rate set corresponding to the multiple sets of recommended creative materials, sort the creative materials included in the set of creative materials to be recommended to obtain the sequence of creative materials to be recommended.

[0083] In some embodiments, the aforementioned executing entity may sort the various creative materials included in the creative material set to be recommended based on the estimated click-through rate set corresponding to the target creative material set to be recommended and the click-through rate set corresponding to the multiple recommended creative material sets, thereby obtaining a sequence of creative materials to be recommended.

[0084] As an example, the aforementioned executing entity can sort the various creative materials included in the aforementioned creative material set to be recommended in ascending order of click-through rate, based on the estimated click-through rate set corresponding to the aforementioned target creative material set to be recommended and the click-through rate set corresponding to the aforementioned multiple recommended creative material sets, to obtain the sequence of creative materials to be recommended.

[0085] In some optional implementations of certain embodiments, after step 305, the following steps are further included:

[0086] The first step is to respond to the fact that there is no target creative material in the above set of creative materials to be recommended. Then, each creative material in the set of creative materials to be recommended is input into the material click-through rate generation model to output the material click-through rate and obtain the material click-through rate set.

[0087] The aforementioned click-through rate (CTR) generation model can be based on a historical set of creative materials corresponding to the items to be recommended, generating a model for the CTR at a target time. For example, the CTR generation model could be a recurrent neural network model.

[0088] The second step is to sort the creative materials to be recommended in the above set of click-through rates to obtain a sequence of creative materials to be recommended.

[0089] As an example, the aforementioned implementing entity can sort the creative materials to be recommended in the set of creative materials to be recommended according to the order of click-through rate from smallest to largest, thus obtaining a sequence of creative materials to be recommended.

[0090] Optionally, the steps also include:

[0091] At least one creative material from the aforementioned sequence of recommended creative materials that meets the preset click-through rate (CTR) condition is sent to the creative material recommendation terminal. The creative material recommendation terminal can be the terminal that recommends creative materials (i.e., delivers creative materials). The preset CTR condition can be at least one creative material from the sequence of recommended creative materials whose CTR ranks within the top predetermined number.

[0092] The above embodiments of this disclosure have the following beneficial effects: By determining the click-through rate (CTR) corresponding to the target creative material to be recommended, the material sequence generation method of some embodiments of this disclosure can generate a more accurate material recommendation order for the set of creative materials to be recommended, thereby improving the material recommendation effect. Specifically, the reason for poor material recommendation effect is that it is impossible to determine whether the newly generated material is a high-quality, high-potential creative material. Directly giving the newly generated material sufficient exposure may lead to poor actual material recommendation effect. Based on this, the material sequence generation method of some embodiments of this disclosure firstly, in response to the existence of target creative materials to be recommended in the acquired set of creative materials to be recommended, determines the target set of creative materials to be recommended from the set of creative materials to be recommended, so as to generate the CTR corresponding to each target creative material to be recommended in the subsequent process. Then, for each target creative material to be recommended in the above target set of creative materials to be recommended, a predetermined number of target historical creative materials are selected from the acquired set of historical recommended creative materials to obtain the target historical recommended creative material set. Here, since the target creative materials to be recommended do not have a real CTR, it is impossible to determine whether the target creative materials to be recommended are high-quality, high-potential creative materials. Therefore, by selecting a predetermined number of target historical recommended creative materials from the historical recommended creative materials, based on their correlation with the target recommended creative materials, the click-through rate (CTR) of the target recommended creative materials can be estimated subsequently. Next, based on the CTR of each target historical recommended creative material in the obtained target historical recommended creative material set, the estimated CTR of each target recommended creative material in the aforementioned target recommended creative material set can be estimated relatively accurately. Furthermore, the CTR of each multiple recommended creative material in the multiple recommended creative material set is determined for use in the subsequent sorting of materials in the recommended creative material set. Finally, based on the estimated CTR set corresponding to the target recommended creative material set and the CTR set corresponding to the multiple recommended creative material set, the materials included in the recommended creative material set are sorted to obtain the recommended creative material sequence. Here, the estimated CTR of the target recommended creative material is used as the standard to measure the quality of the creative materials. The CTR of the multiple recommended creative materials is used to measure the quality of the multiple recommended creative materials. Therefore, by ranking the recommended creative materials based on the click-through rate of each creative material in the set to be recommended, the materials can be sorted according to their quality. Using a sequence of creative materials to recommend individual materials can significantly improve the effectiveness of material recommendation.

[0093] Further reference Figure 4 The diagram illustrates a flow 400 of another embodiment of the material sequence generation method according to the present disclosure. This material sequence generation method includes the following steps:

[0094] Step 401: In response to the existence of target creative materials in the acquired set of creative materials to be recommended, determine the target set of creative materials to be recommended.

[0095] Step 402: For each target creative material in the target creative material set to be recommended, select a predetermined number of target historical recommended creative materials from the obtained historical recommended creative material set to obtain the target historical recommended creative material set.

[0096] In some embodiments, the specific implementation of steps 401-402 and the resulting technical effects can be found in [reference needed]. Figure 3 Steps 301-302 in the corresponding embodiments will not be repeated here.

[0097] Step 403: For each target creative material in the target creative material set to be recommended, calculate the click-through rate of each target historical creative material in the target historical creative material set corresponding to the target creative material to be recommended, and obtain the calculated value as the estimated click-through rate corresponding to the target creative material to be recommended.

[0098] In some embodiments, the executing entity (e.g. Figure 1 The electronic device 101 shown can calculate the click-through rate of each target historical recommended creative material in the target historical recommended creative material set corresponding to each target target recommended creative material in the target target recommended creative material set, and obtain the calculated value as the estimated click-through rate corresponding to the target target recommended creative material.

[0099] For example, for a target creative to be recommended, the corresponding set of historical recommended creatives includes: historical recommended creatives for the first target, historical recommended creatives for the second target, and historical recommended creatives for the third target. The click-through rate (CTR) for the first historical recommended creative is 2%. The CTR for the second historical recommended creative is 4%. The CTR for the third historical recommended creative is 6%. Averaging the CTRs of the first, second, and third historical recommended creatives yields the average CTR. The average CTR is 4%. Therefore, the CTR for the target creative to be recommended is 4%.

[0100] Step 404: Determine the click-through rate of each creative material in the multiple recommended creative material set.

[0101] Step 405: Based on the estimated click-through rate set corresponding to the target set of creative materials to be recommended and the click-through rate set corresponding to the multiple sets of recommended creative materials, sort the creative materials included in the set of creative materials to be recommended to obtain the sequence of creative materials to be recommended.

[0102] In some embodiments, the specific implementation of steps 404-405 and the resulting technical effects can be found in [reference needed]. Figure 3 Steps 304-305 in the corresponding embodiments will not be repeated here.

[0103] from Figure 4 It can be seen from this that, with Figure 3 Compared to the description of some corresponding embodiments, Figure 4 In some corresponding embodiments, the process 400 of the material sequence generation method calculates the click-through rate (CTR) of each historical recommended creative material in the target historical recommended creative material set corresponding to each target creative material to be recommended. This can more accurately represent the CTR of each target creative material to be recommended, facilitating the subsequent generation of a more accurate sequence of recommended creative materials.

[0104] Further reference Figure 5 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a material sequence generation apparatus, which are similar to... Figure 3 Corresponding to the method embodiments shown, the device can be specifically applied to various electronic devices.

[0105] like Figure 5As shown, a material sequence generation device 500 includes: a first determining unit 501, a filtering unit 502, a generation unit 503, a second determining unit 504, and a sorting unit 505. The first determining unit 501 is configured to determine a target set of creative materials to be recommended in response to the existence of target creative materials in the acquired set of creative materials to be recommended, wherein the target creative materials to be recommended are creative materials that meet the criteria for determining target materials. The filtering unit 502 is configured to, for each target creative material in the target set of creative materials to be recommended, filter a predetermined number of target historical recommended creative materials from the acquired set of historical recommended creative materials to obtain a target historical recommended creative material set, wherein the material correlation between the target historical recommended creative materials and the target creative materials to be recommended is greater than a predetermined threshold. The generation unit 503 is configured to, based on the obtained target historical recommended creative materials... The click-through rates (CTRs) of each target historical recommended creative material in the creative material set are used to generate the estimated CTRs of each target recommended creative material in the target recommended creative material set. The second determining unit 504 is configured to determine the CTRs of each multiple recommended creative material in the multiple recommended creative material set, wherein the multiple recommended creative material set is the material set in the recommended creative material set excluding the target recommended creative material set. The sorting unit 505 is configured to sort each recommended creative material included in the recommended creative material set according to the estimated CTR set corresponding to the target recommended creative material set and the CTR set corresponding to the multiple recommended creative material set, to obtain a recommended creative material sequence.

[0106] In some optional implementations of certain embodiments, the aforementioned preset material conditions include: multiple preset material sub-conditions, the multiple preset material sub-conditions including: the time interval between the material generation time corresponding to the creative material to be recommended and the current time is less than a predetermined duration, and the exposure amount corresponding to the creative material to be recommended is less than a predetermined exposure value; and the first determining unit 501 in the aforementioned device 500 may be further configured to: determine the material generation time and exposure amount corresponding to each creative material to be recommended in the aforementioned set of creative materials to be recommended; and determine the creative materials to be recommended in the aforementioned set of creative materials to be recommended that satisfy the multiple preset material sub-conditions as target creative materials to be recommended.

[0107] In some optional implementations of certain embodiments, the filtering unit 502 in the above-mentioned device 500 may be further configured to: input each historical recommended creative material in the above-mentioned historical recommended creative material set into a pre-trained material encoding model to generate historical material encoding vectors, thereby obtaining a historical material encoding vector set; input the above-mentioned target creative material to be recommended into the above-mentioned material encoding model to generate a material encoding vector to be recommended; and, based on the above-mentioned historical material encoding vector set and the above-mentioned material encoding vector to be recommended, filter out historical recommended creative materials whose vector distance meets a predetermined distance condition from the above-mentioned historical recommended creative material set as target historical recommended creative materials, thereby obtaining a target historical recommended creative material set, wherein the vector distance is the vector distance between the historical material encoding vector in the above-mentioned historical material encoding vector set and the material encoding vector to be recommended.

[0108] In some optional implementations of some embodiments, the generation unit 503 in the above-mentioned device 500 may be further configured to: for each target creative material in the target creative material set to be recommended, calculate the click-through rate of each target historical creative material in the target historical creative material set corresponding to the target creative material to be recommended, and obtain a calculated value as the estimated click-through rate corresponding to the target creative material to be recommended.

[0109] In some optional implementations of certain embodiments, the above-mentioned material encoding model is trained through the following steps: obtaining a training sample set, wherein the training samples include: training creative materials and product term tags; using the above-mentioned training sample set, training the initial product term classification model to obtain the trained product term classification model, wherein the trained product term classification model includes: the material encoding model.

[0110] In some optional implementations of certain embodiments, the apparatus 500 further includes an input unit and a material sorting unit (not shown in the figure). The input unit can be configured to: in response to the absence of a target creative material in the set of creative materials to be recommended, input each creative material in the set of creative materials to be recommended into a material click-through rate (CTR) generation model to output the material CTR, thereby obtaining a material CTR set. The material sorting unit can be configured to: sort the various creative materials included in the set of creative materials to be recommended according to the material CTR set, thereby obtaining a sequence of creative materials to be recommended.

[0111] In some optional implementations of certain embodiments, the apparatus 500 further includes a sending unit (not shown in the figure). The sending unit can be configured to send at least one creative material from the sequence of creative materials to be recommended, whose click-through rate meets a preset click-through rate condition, to the material recommendation terminal.

[0112] It is understandable that the units described in the device 500 are related to the reference. Figure 3 The steps in the described method correspond accordingly. Therefore, the operations, features, and beneficial effects described above for the method also apply to the device 500 and the units contained therein, and will not be repeated here.

[0113] The following is for reference. Figure 6 It illustrates electronic devices suitable for implementing some embodiments of this disclosure (e.g., Figure 1 A schematic diagram of the structure of electronic device 101)600 in the middle. Figure 6 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0114] like Figure 6 As shown, electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processor, etc.) 601, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 602 or a program loaded from storage device 608 into random access memory (RAM) 603. RAM 603 also stores various programs and data required for the operation of electronic device 600. Processing device 601, ROM 602, and RAM 603 are interconnected via bus 604. Input / output (I / O) interface 605 is also connected to bus 604.

[0115] Typically, the following devices can be connected to I / O interface 605: input devices 606 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 607 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 608 including, for example, magnetic tapes, hard disks, etc.; and communication devices 609. Communication device 609 allows electronic device 600 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 An electronic device 600 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 6 Each box shown can represent a device or multiple devices as needed.

[0116] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 609, or installed from a storage device 608, or installed from a ROM 602. When the computer program is executed by the processing device 601, it performs the functions defined above in the methods of some embodiments of this disclosure.

[0117] It should be noted that, in some embodiments of this disclosure, the computer-readable medium described above may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0118] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed networks.

[0119] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: in response to the existence of target creative materials in the acquired set of creative materials to be recommended, determine a set of target creative materials to be recommended, wherein the target creative materials to be recommended are creative materials that satisfy the criteria for determining target materials; for each target creative material in the set of target creative materials to be recommended, filter a predetermined number of target historical recommended creative materials from the acquired set of historical recommended creative materials to obtain a set of target historical recommended creative materials, wherein the material correlation between the target historical recommended creative materials and the aforementioned target creative materials is greater than [value missing]. A predetermined threshold is set; based on the click-through rate (CTR) of each target historical recommended creative material in the obtained target historical recommended creative material set, the estimated CTR of each target recommended creative material in the above-mentioned target recommended creative material set is generated; the CTR of each multiple recommended creative material in the multiple recommended creative material set is determined, wherein the above-mentioned multiple recommended creative material set is the material set in the above-mentioned recommended creative material set excluding the above-mentioned target recommended creative material set; based on the estimated CTR set corresponding to the target recommended creative material set and the CTR set corresponding to the multiple recommended creative material set, the various recommended creative materials included in the recommended creative material set are sorted to obtain the recommended creative material sequence.

[0120] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0121] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0122] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including a first determining unit, a filtering unit, a generating unit, a second determining unit, and a sorting unit. The names of these units do not necessarily limit the specific unit itself; for example, the first determining unit may also be described as "a unit that determines the target set of creative materials to be recommended in response to the existence of a target set of creative materials to be recommended in the acquired set of creative materials to be recommended."

[0123] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0124] Some embodiments of this disclosure also provide a computer program product, including a computer program that, when executed by a processor, implements any of the above-described material sequence generation methods.

[0125] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A method for generating a material sequence, comprising: In response to the existence of target creative materials in the acquired set of creative materials to be recommended, a target set of creative materials to be recommended is determined, wherein the target creative materials to be recommended are creative materials that meet the conditions for determining the target materials; For each target creative material in the target creative material set to be recommended, a predetermined number of target historical recommended creative materials are selected from the acquired historical recommended creative material set to obtain a target historical recommended creative material set, wherein the material correlation between the target historical recommended creative materials and the target creative material to be recommended is greater than a predetermined threshold. Based on the click-through rate of each target historical recommended creative material in the obtained target historical recommended creative material set, the estimated click-through rate of each target recommended creative material in the target recommended creative material set is generated; Determine the click-through rate of each creative material in the multiple recommendation creative material set, wherein the multiple recommendation creative material set is the material set in the creative material set to be recommended excluding the target creative material set; Based on the estimated click-through rate set corresponding to the target set of creative materials to be recommended and the click-through rate set corresponding to the multiple sets of recommended creative materials, the creative materials to be recommended in the set of creative materials to be recommended are sorted to obtain a sequence of creative materials to be recommended.

2. The method of claim 1, wherein, The target material determination conditions include: multiple preset material sub-conditions, which include: the time interval between the material generation time and the current time corresponding to the creative material to be recommended is less than a predetermined duration; the exposure of the creative material to be recommended is less than a predetermined exposure value; and... The identified set of creative materials to be recommended includes: Determine the generation time and exposure volume of each creative material in the set of creative materials to be recommended; The creative materials to be recommended that meet multiple preset material sub-conditions are identified as target creative materials to be recommended.

3. The method of claim 1, wherein, The step of filtering a predetermined number of target historical recommended creative materials from the acquired historical recommended creative material set to obtain the target historical recommended creative material set includes: Each historical recommended creative material in the historical recommended creative material set is input into a pre-trained material encoding model to generate a historical material encoding vector, thus obtaining a historical material encoding vector set; The target creative material to be recommended is input into the material encoding model to generate the material encoding vector to be recommended; Based on the historical material encoding vector set and the material encoding vector to be recommended, historical recommended creative materials whose vector distance meets a predetermined distance condition are selected from the historical recommended creative material set and used as target historical recommended creative materials to obtain the target historical recommended creative material set. Here, the vector distance is the vector distance between the historical material encoding vector in the historical material encoding vector set and the material encoding vector to be recommended.

4. The method of claim 1, wherein, The step of generating the estimated click-through rate of each target creative material in the target creative material set based on the click-through rate of each target historical recommended creative material in the obtained target historical recommended creative material set includes: For each target creative material in the target creative material set to be recommended, the click-through rate of each target historical creative material in the target historical recommended creative material set corresponding to the target creative material to be recommended is calculated and processed to obtain a calculated value, which is used as the estimated click-through rate corresponding to the target creative material to be recommended.

5. The method of claim 3, wherein, The material encoding model is trained through the following steps: Obtain a training sample set, which includes: training creative materials and product term tags; Using the training sample set, the initial item product word classification model is trained to obtain the trained item product word classification model, wherein the trained item product word classification model includes: a material encoding model.

6. The method of claim 1, wherein, The method further includes: In response to the absence of a target creative material in the set of creative materials to be recommended, each creative material to be recommended in the set of creative materials to be recommended is input into the material click-through rate generation model to output the material click-through rate and obtain the material click-through rate set; Based on the click-through rate set of the materials, the creative materials to be recommended in the set of creative materials to be recommended are sorted to obtain the sequence of creative materials to be recommended.

7. The method of claim 1 or 6, wherein, The method further includes: Send at least one creative material from the sequence of creative materials to be recommended that meets the preset click-through rate condition to the material recommendation terminal.

8. A material sequence generation apparatus, comprising: The first determining unit is configured to determine the target set of creative materials to be recommended in response to the existence of target creative materials in the acquired set of creative materials to be recommended, wherein the target creative materials to be recommended are creative materials that meet the determination conditions of target materials; The filtering unit is configured to, for each target creative material in the target creative material set to be recommended, filter a predetermined number of target historical recommended creative materials from the acquired historical recommended creative material set to obtain a target historical recommended creative material set, wherein the material correlation between the target historical recommended creative materials and the target creative material to be recommended is greater than a predetermined threshold. The generation unit is configured to generate the estimated click-through rate of each target creative material in the target creative material set based on the click-through rate of each target historical recommended creative material in the obtained target historical recommended creative material set; The second determining unit is configured to determine the click-through rate of each multiple recommended creative material in the multiple recommended creative material set, wherein the multiple recommended creative material set is the material set in the creative material set to be recommended excluding the target creative material set; The sorting unit is configured to sort the various creative materials included in the creative material set to be recommended according to the estimated click-through rate set corresponding to the target creative material set and the click-through rate set corresponding to the multiple recommended creative material sets, so as to obtain a sequence of creative materials to be recommended.

9. An electronic device, comprising: One or more processors; Storage device, on which one or more programs are stored, When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-7.

10. A computer readable medium having stored thereon a computer program, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1-7.

11. A computer program product comprising a computer program that, when executed by a processor, implements the method as described in any one of claims 1-7.

Images