Information pushing method and device, electronic equipment and storage medium

Abstract

This disclosure relates to an information push method, apparatus, electronic device, and storage medium. The method includes: acquiring first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold; based on the first operation information and second operation information of a second operation performed by a second user through the target application, determining target push information for pushing to the first user's user terminal from a predetermined set of information to be pushed, wherein the execution frequency of the second user for the second operation is greater than the preset frequency threshold; and pushing the target push information to the user terminal. This method alleviates the sparsity of user behavior and improves the accuracy of pushing information to users with sparse behavior.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to an information push method, apparatus, electronic device and storage medium. Background Technology

[0002] Information push is a technology that uses certain technical standards or protocols and methods to transmit information needed by users over the Internet in order to reduce information overload.

[0003] In existing technologies, different information (such as videos, news, images, and products) is typically pushed to users based on user behavior (e.g., searching, watching, liking, and posting comments). However, for users with sparse behavior, the accuracy of information push is often low. Summary of the Invention

[0004] In view of this, in order to solve some or all of the above-mentioned technical problems, the present disclosure provides an information push method, apparatus, electronic device and storage medium.

[0005] In a first aspect, embodiments of this disclosure provide an information push method, the method comprising:

[0006] Obtain first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold;

[0007] Based on the first operation information and the second operation information of the second operation performed by the second user through the target application, target push information for pushing to the user terminal of the first user is determined from a predetermined set of information to be pushed, wherein the execution frequency of the second user for the second operation is greater than the preset frequency threshold.

[0008] The target push information is pushed to the user terminal.

[0009] In one possible implementation, the step of determining target push information for pushing to the user terminal of the first user from a predetermined set of information to be pushed, based on the first operation information and the second operation information performed by the second user through the target application, includes:

[0010] Based on the first operation information and the second operation information of the second operation performed by the second user through the target application, a first operation feature of the first user is generated.

[0011] Based on the first operational feature, target push information for pushing to the first user's terminal is determined from a pre-determined set of information to be pushed.

[0012] In one possible implementation, generating the first operation feature of the first user based on the first operation information and the second operation information of the second operation performed by the second user through the target application includes:

[0013] The first operation information is input into a pre-trained first model to obtain the first operation feature of the first user. The first model represents the correspondence between the first operation information and the first operation feature. The first model is trained based on the second operation information of the second operation performed by the second user through the target application.

[0014] In one possible implementation, the first model is trained in the following manner:

[0015] The second operation information is input into a pre-trained second model to obtain the second operation feature of the second user, wherein the second model represents the correspondence between the second operation information and the second operation feature;

[0016] Based on the first operation information and the second operation features, the model parameters of the first initial model of the first model are adjusted to train the first initial model;

[0017] The first initial model that has been trained is identified as the first model.

[0018] In one possible implementation, the second model is trained in the following manner:

[0019] Obtain a training sample set, wherein the training samples in the training sample set include second operation information and corresponding user category information, and the user category information represents the user category of the user executing the operation indicated by the second operation information;

[0020] A machine learning algorithm is used to train a second initial model of the second model based on the second operation information and the user category information;

[0021] The trained second initial model is then identified as the second model.

[0022] In one possible implementation, the step of employing a machine learning algorithm to train a second initial model of the second model based on the second operation information and the user category information includes:

[0023] Obtain the second initial model of the second model;

[0024] The second operation information is input into the second initial model to obtain the first output data of the second initial model;

[0025] The first output data is input into the fully connected layer to obtain the actual output data of the fully connected layer;

[0026] Using machine learning algorithms, the model parameters of the second initial model are adjusted based on the user category information and the actual output data to train the second initial model.

[0027] In one possible implementation, adjusting the model parameters of the first initial model based on the first operation information and the second operation feature to train the first initial model includes:

[0028] Obtain the first initial model of the first model;

[0029] The first operation information is input into the first initial model to obtain the second output data of the first initial model;

[0030] The following steps are used to adjust the model parameters of the first initial model of the first model in order to train the first initial model:

[0031] When the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to the same user category, a first loss function is used to adjust the model parameters of the first initial model based on the second operation feature and the second output data. The function value of the first loss function is positively correlated with the target similarity, and the target similarity is the similarity between the second output data and the second operation feature.

[0032] When the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to different user categories, a second loss function is used to adjust the model parameters of the first initial model based on the second operation feature and the second output data. The function value of the second loss function is negatively correlated with the target similarity, which is the similarity between the second output data and the second operation feature.

[0033] In one possible implementation,

[0034] The first loss function is: L = d(H) ua H ub );

[0035] The second loss function is: L = max(0, md(H) ua H ub ));

[0036] Among them, H ua H represents the second output data.ub Characterizing the second operational feature, d(H) ua H ub The similarity of the target is represented by m, which is a preset threshold.

[0037] In one possible implementation, the initial model parameters of the second initial model are the same as those of the first initial model, and the model structure of the second initial model is the same as that of the first initial model.

[0038] Secondly, embodiments of this disclosure provide an information push device, the device comprising:

[0039] The acquisition unit is used to acquire first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold.

[0040] The determining unit is configured to determine, based on the first operation information and the second operation information of the second operation performed by the second user through the target application, target push information from a predetermined set of information to be pushed to the user terminal of the first user, wherein the execution frequency of the second operation by the second user is greater than the preset frequency threshold.

[0041] The push unit is used to push the target push information to the user terminal.

[0042] In one possible implementation, the step of determining target push information for pushing to the user terminal of the first user from a predetermined set of information to be pushed, based on the first operation information and the second operation information performed by the second user through the target application, includes:

[0043] Based on the first operation information and the second operation information of the second operation performed by the second user through the target application, a first operation feature of the first user is generated.

[0044] Based on the first operational feature, target push information for pushing to the first user's terminal is determined from a pre-determined set of information to be pushed.

[0045] In one possible implementation, generating the first operation feature of the first user based on the first operation information and the second operation information of the second operation performed by the second user through the target application includes:

[0046] The first operation information is input into a pre-trained first model to obtain the first operation feature of the first user. The first model represents the correspondence between the first operation information and the first operation feature. The first model is trained based on the second operation information of the second operation performed by the second user through the target application.

[0047] In one possible implementation, the first model is trained in the following manner:

[0048] The second operation information is input into a pre-trained second model to obtain the second operation feature of the second user, wherein the second model represents the correspondence between the second operation information and the second operation feature;

[0049] Based on the first operation information and the second operation features, the model parameters of the first initial model of the first model are adjusted to train the first initial model;

[0050] The first initial model that has been trained is identified as the first model.

[0051] In one possible implementation, the second model is trained in the following manner:

[0052] Obtain a training sample set, wherein the training samples in the training sample set include second operation information and corresponding user category information, and the user category information represents the user category of the user executing the operation indicated by the second operation information;

[0053] A machine learning algorithm is used to train a second initial model of the second model based on the second operation information and the user category information;

[0054] The trained second initial model is then identified as the second model.

[0055] In one possible implementation, the step of employing a machine learning algorithm to train a second initial model of the second model based on the second operation information and the user category information includes:

[0056] Obtain the second initial model of the second model;

[0057] The second operation information is input into the second initial model to obtain the first output data of the second initial model;

[0058] The first output data is input into the fully connected layer to obtain the actual output data of the fully connected layer;

[0059] Using machine learning algorithms, the model parameters of the second initial model are adjusted based on the user category information and the actual output data to train the second initial model.

[0060] In one possible implementation, adjusting the model parameters of the first initial model based on the first operation information and the second operation feature to train the first initial model includes:

[0061] Obtain the first initial model of the first model;

[0062] The first operation information is input into the first initial model to obtain the second output data of the first initial model;

[0063] The following steps are used to adjust the model parameters of the first initial model of the first model in order to train the first initial model:

[0064] When the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to the same user category, a first loss function is used to adjust the model parameters of the first initial model based on the second operation feature and the second output data. The function value of the first loss function is positively correlated with the target similarity, and the target similarity is the similarity between the second output data and the second operation feature.

[0065] When the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to different user categories, a second loss function is used to adjust the model parameters of the first initial model based on the second operation feature and the second output data. The function value of the second loss function is negatively correlated with the target similarity, which is the similarity between the second output data and the second operation feature.

[0066] In one possible implementation,

[0067] The first loss function is: L = d(H) ua H ub );

[0068] The second loss function is: L = max(0, md(H) ua H ub ));

[0069] Among them, H ua H represents the second output data. ub Characterizing the second operational feature, d(H) ua H ubThe similarity of the target is represented by m, which is a preset threshold.

[0070] In one possible implementation, the initial model parameters of the second initial model are the same as those of the first initial model, and the model structure of the second initial model is the same as that of the first initial model.

[0071] Thirdly, embodiments of this disclosure provide an electronic device, including:

[0072] Memory, used to store computer programs;

[0073] A processor is configured to execute a computer program stored in the aforementioned memory, and when the aforementioned computer program is executed, to implement the method of any embodiment of the information push method of the first aspect of this disclosure.

[0074] Fourthly, embodiments of this disclosure provide a computer-readable storage medium, which, when executed by a processor, implements the method of any embodiment of the information push method described in the first aspect above.

[0075] Fifthly, embodiments of this disclosure provide a computer program including computer-readable code that, when executed on a device, causes a processor in the device to execute instructions for implementing the steps of the method in any embodiment of the information push method as described in the first aspect above.

[0076] The information push method provided in this disclosure involves obtaining first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold. Then, based on the first operation information and second operation information of a second operation performed by a second user through the target application, target push information for pushing to the first user's user terminal is determined from a predetermined set of information to be pushed, wherein the execution frequency of the second user for the second operation is greater than the preset frequency threshold. Finally, the target push information is pushed to the user terminal. This method can determine push information for users with sparse behavior based on the operation information of users with intensive behavior, thereby alleviating the sparsity of user behavior and improving the accuracy of pushing information to users with sparse behavior. Attached Figure Description

[0077] Figure 1 A flowchart illustrating an information push method provided in an embodiment of this disclosure;

[0078] Figure 2 A flowchart illustrating another information push method provided in this embodiment of the disclosure;

[0079] Figure 3A A flowchart illustrating yet another information push method provided in this embodiment of the disclosure;

[0080] Figure 3B This is a schematic diagram of the structure of the second model involved in an information push method provided in an embodiment of this disclosure;

[0081] Figure 3C A schematic diagram illustrating the training of a first model involved in an information push method provided in an embodiment of this disclosure;

[0082] Figure 3D A flowchart illustrating another information push method provided in this embodiment of the present disclosure;

[0083] Figure 4 This is a schematic diagram of the structure of an information push device provided in an embodiment of the present disclosure;

[0084] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0085] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0086] Those skilled in the art will understand that the terms "first" and "second" in the embodiments of this disclosure are only used to distinguish different steps, devices or modules, and do not represent any specific technical meaning, nor do they indicate the logical order between them.

[0087] It should also be understood that in this embodiment, "multiple" can refer to two or more, and "at least one" can refer to one, two or more.

[0088] It should also be understood that any component, data or structure mentioned in the embodiments of this disclosure can generally be understood as one or more unless expressly defined or given to the contrary in the context.

[0089] Furthermore, the term "and / or" in this disclosure is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this disclosure generally indicates that the preceding and following related objects have an "or" relationship.

[0090] It should also be understood that the description of the various embodiments in this disclosure emphasizes the differences between the various embodiments, and the similarities or similarities can be referred to each other. For the sake of brevity, they will not be described in detail.

[0091] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use.

[0092] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0093] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0094] It should be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other. To facilitate understanding of the embodiments of this disclosure, the disclosure will be described in detail below with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are only some, not all, of the embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without creative effort are within the scope of protection of this disclosure.

[0095] Furthermore, it should be noted that the users described in this disclosure (including the first user, the second user, the user executing the operation indicated by the first operation information, the user executing the operation indicated by the second operation information, etc.) can be distinguished by user identifiers. For example, a user identifier can be a login account. In this scenario, if different people log in using the same account, they can be considered the same user; if the same person logs in using different accounts, they can be considered different users.

[0096] Figure 1 This is a flowchart illustrating an information push method provided in an embodiment of this disclosure. Figure 1 As shown, the method specifically includes:

[0097] 101. Obtain first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold.

[0098] In this embodiment, the first user can be a user whose frequency of the operation performed by the target application (i.e., the first operation mentioned above) is less than or equal to a preset frequency threshold.

[0099] The target application can be any application. For example, the target application could be shopping software, shopping website, video playback software, audio playback website, etc.

[0100] The first action can be any action performed by the first user through the target application. For example, the first action could include: searching, purchasing, adding to favorites, clicking, browsing, watching, sending comments, etc.

[0101] The first operation information can be various information related to the first operation mentioned above. As an example, the first operation information may include: purchase information, favorite information, click information, browsing information, viewing information, sending bullet comments information, operation object information, operation time information, etc.

[0102] The preset frequency threshold can be a predetermined frequency. For example, the preset frequency threshold could be once every three days, or 50 times every 90 days, etc.

[0103] Here, if the first user has used the search function 3 times and watched 20 videos in the past 90 days, then the execution frequency of the first operation can be: 23 times every 90 days.

[0104] 102. Based on the first operation information and the second operation information of the second operation performed by the second user through the target application, determine the target push information to be pushed to the user terminal of the first user from a predetermined set of information to be pushed, wherein the execution frequency of the second user for the second operation is greater than the preset frequency threshold.

[0105] In this embodiment, the second user can be a user whose frequency of performing an operation (i.e., the second operation mentioned above) through the target application (i.e., the execution frequency mentioned above) is greater than the preset frequency threshold mentioned above.

[0106] The second action can be any action performed by a second user through the target application. For example, this second action could include: searching, purchasing, adding to favorites, clicking, browsing, watching, sending comments, etc.

[0107] The second operation information can be various information related to the second operation mentioned above. As an example, the second operation information may include: purchase information, favorite information, click information, browsing information, viewing information, sending bullet comments information, operation object information, operation time information, etc.

[0108] Here, the difference between the second operation and the first operation lies in the user executing them. The second operation is executed by the second user, while the first operation is executed by the first user.

[0109] The set of information to be pushed may include one or more pieces of information to be pushed. These information may include videos, news, images, product information, etc. The set of information to be pushed may be a collection of all or part of the information available for push notification in the aforementioned target application.

[0110] The first user's user terminal can be any user terminal used by the first user. For example, the first user's user terminal can be any user terminal logged into with the first user's account.

[0111] The target push information can be one or more push information items from the set of information to be pushed to the aforementioned user terminals.

[0112] For example, for each piece of information to be pushed in the aforementioned set of information to be pushed, the similarity between the information to be pushed and the second operation information can be calculated. A preset number of pieces of information to be pushed can be selected in descending order of similarity to obtain a subset of information to be pushed. Then, for each piece of information to be pushed in the subset of information to be pushed, the similarity between the information to be pushed and the first operation information can be calculated. The piece of information to be pushed with the highest similarity to the first operation information is determined as the target push information to be pushed to the user terminal of the first user.

[0113] 103. Push the target push information to the user terminal.

[0114] Optionally, after receiving the target push information, the user terminal can perform semantic analysis on the target push information to obtain the semantic analysis results, and then perform corresponding operations based on the semantic analysis results. For example, if the semantic analysis result indicates "xx variety show has been updated", the user terminal can automatically perform a download operation (i.e., the corresponding operation mentioned above), or the user terminal can automatically display a prompt message to remind the user to download the variety show, and download the variety show if the user confirms the prompt message.

[0115] The information push method provided in this disclosure involves obtaining first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold. Then, based on the first operation information and second operation information of a second operation performed by a second user through the target application, target push information for pushing to the first user's user terminal is determined from a predetermined set of information to be pushed, wherein the execution frequency of the second user for the second operation is greater than the preset frequency threshold. Finally, the target push information is pushed to the user terminal. This method can determine push information for users with sparse behavior based on the operation information of users with intensive behavior, thereby alleviating the sparsity of user behavior and improving the accuracy of pushing information to users with sparse behavior.

[0116] Figure 2 This is a flowchart illustrating another information push method provided in an embodiment of this disclosure. Figure 2 As shown, the method specifically includes:

[0117] 201. Obtain first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold.

[0118] In this embodiment, step 201 and Figure 1 Step 101 in the corresponding embodiment is basically the same, and will not be repeated here.

[0119] 202. Based on the first operation information and the second operation information of the second operation performed by the second user through the target application, generate the first operation feature of the first user, wherein the execution frequency of the second user for the second operation is greater than the preset frequency threshold.

[0120] In this embodiment, the first operation feature can characterize the features of the first operation performed by the first user. In practice, the first operation feature can be represented by a vector.

[0121] For example, the first operation feature of the first user can be generated according to preset rules based on the first operation information and the second operation information of the second operation performed by the second user through the target application.

[0122] 203. Based on the first operation feature, determine the target push information to be pushed to the user terminal of the first user from the predetermined set of information to be pushed.

[0123] In this embodiment, for each piece of information to be pushed in the set of information to be pushed, the similarity between the information to be pushed and the first operation feature can be calculated, and the information to be pushed with the highest similarity among the calculated similarities can be determined as the target information to be pushed.

[0124] 204. Push the target push information to the user terminal.

[0125] In this embodiment, step 204 and Figure 1 Step 103 in the corresponding embodiment is basically the same, and will not be repeated here.

[0126] It should be noted that, in addition to the contents described above, this embodiment may also include... Figure 1 The technical features described in the corresponding embodiments, thereby achieving Figure 1 For details on the technical effects of the information push method shown, please refer to [link / reference]. Figure 1 The relevant descriptions are presented concisely and will not be elaborated upon here.

[0127] The information push method provided in this disclosure generates a first operation feature of a first user based on first operation information and second operation information, and then determines the target push information based on the first operation feature, thereby further improving the accuracy of information push.

[0128] Figure 3A This is a flowchart illustrating another information push method provided in this embodiment. This method can be applied to one or more electronic devices such as smartphones, laptops, desktop computers, portable computers, and servers. Furthermore, the executing entity of this method can be hardware or software. When the executing entity is hardware, it can be one or more of the aforementioned electronic devices. For example, a single electronic device can execute this method, or multiple electronic devices can cooperate with each other to execute this method. When the executing entity is software, this method can be implemented as multiple software programs or software modules, or as a single software program or software module. No specific limitations are made here.

[0129] Specifically, such as Figure 3A As shown, the method specifically includes:

[0130] 301. Obtain first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold.

[0131] In this embodiment, step 301 and Figure 1 Step 101 in the corresponding embodiment is basically the same, and will not be repeated here.

[0132] 302. Input the first operation information into a pre-trained first model to obtain the first operation feature of the first user, wherein the first model represents the correspondence between the first operation information and the first operation feature, the first model is trained based on the second operation information of the second operation performed by the second user through the target application, and the execution frequency of the second operation by the second user is greater than the preset frequency threshold.

[0133] In this embodiment, the first model may include structures such as convolutional neural networks.

[0134] For example, the first model can be trained in the following manner:

[0135] First, obtain the first initial model of the first model and the first training sample set. The first initial model can be the initial model of the first model. After training the first initial model and adjusting its model parameters (including the initial model parameters), the first model can be obtained. The training samples in the first training sample set can include input data and target output data. The input data can include first operation information and second operation information of the sample, and the target output data can include the first operation features of the sample.

[0136] Then, a machine learning algorithm is used to input the above-mentioned input data and the above-mentioned expected output data into the above-mentioned first initial model in order to train the first initial model and adjust the model parameters of the first initial model, thereby obtaining the first model.

[0137] 303. Based on the first operation feature, determine the target push information to be pushed to the user terminal of the first user from a predetermined set of information to be pushed.

[0138] In this embodiment, step 303 and Figure 2 Step 203 in the corresponding embodiment is basically the same, and will not be repeated here.

[0139] 304. Push the target push information to the user terminal.

[0140] In this embodiment, step 304 and Figure 1 Step 103 in the corresponding embodiment is basically the same, and will not be repeated here.

[0141] In some optional implementations of this embodiment, the first model is trained in the following manner:

[0142] The first step involves inputting the second operation information into a pre-trained second model to obtain the second operation features of the second user. The second model represents the correspondence between the second operation information and the second operation features.

[0143] The second step is to adjust the model parameters of the first initial model of the first model based on the first operation information and the second operation features, so as to train the first initial model.

[0144] Here, the model parameters of the first initial model can be adjusted by calculating the similarity between the first operation information and the second operation feature in order to train the first initial model.

[0145] The third step is to identify the first initial model that has been trained as the first model.

[0146] Here, the first initial model training can be determined to be complete if at least one of the following conditions is met: the similarity between the first operation information and the second operation feature is less than or equal to a preset similarity threshold, the training time exceeds a preset time, and the number of training iterations is greater than or equal to a preset number of iterations.

[0147] It is understood that in the above-mentioned optional implementation methods, the second user's second operation characteristics are obtained through the second model, and the first model is obtained based on the first operation information and the second operation characteristics. Therefore, using this first model for information push can further improve the accuracy of information push.

[0148] In some application scenarios of the above-mentioned optional implementation methods, the second model is trained in the following manner:

[0149] The first step is to obtain a training sample set. The training samples in this set include second operation information and corresponding user category information. The user category information represents the user category of the user executing the operation indicated by the second operation information.

[0150] The second step involves using a machine learning algorithm to train a second initial model of the second model based on the second operation information and the user category information.

[0151] The third step is to identify the trained second initial model as the second model.

[0152] Here, the second initial model training can be determined to be complete if at least one of the following conditions is met: the similarity between the actual output data and the user category information is less than or equal to a preset similarity threshold, the training time exceeds a preset time, and the number of training iterations is greater than or equal to a preset number of iterations.

[0153] The first output data can be obtained as follows: First, the second operation information is input into the second initial model to obtain the first output data. Then, the first output data is input into the fully connected layer to obtain the actual output data of the fully connected layer.

[0154] It is understandable that in the above application scenarios, a second model can be trained based on the second operation information and the corresponding user category information, thereby further improving the accuracy of information push.

[0155] In some of the application scenarios described above, the following method can be used to perform the second step:

[0156] First, obtain the second initial model of the second model.

[0157] The second initial model can be the initial model of the second model. After training the second initial model and adjusting its model parameters (including the initial model parameters), the second model can be obtained.

[0158] The initial model parameters of the second initial model may be the same as or different from those of the first initial model. The model structure of the second initial model may be the same as or different from that of the first initial model.

[0159] Then, the second operation information is input into the second initial model to obtain the first output data of the second initial model.

[0160] The first output data can be the data output by the second initial model after the second operation information is input into the second initial model.

[0161] Then, the first output data is input into the fully connected layer to obtain the actual output data of the fully connected layer.

[0162] The actual output data mentioned above can be the data output by the fully connected layer after the first output data is input into the fully connected layer.

[0163] Finally, a machine learning algorithm is used to adjust the model parameters of the second initial model based on the user category information and the actual output data in order to train the second initial model.

[0164] Specifically, the value of a preset loss function can be calculated based on user category information and the actual output data. Then, gradient descent and backpropagation are used to adjust the model parameters of the second initial model using the aforementioned function value, thereby training the second initial model.

[0165] It is understandable that, in the above situation, a second model can be trained based on the second operation information and user category information, thereby further improving the accuracy of information push.

[0166] In some application scenarios of the above-mentioned optional implementation methods, the model parameters of the first initial model of the first model can be adjusted based on the first operation information and the second operation features to train the first initial model:

[0167] First, obtain the first initial model of the first model.

[0168] Here, the first initial model can be the initial model of the first model. After training the first initial model and adjusting its model parameters (including the initial model parameters), the first model can be obtained.

[0169] Then, the first operation information is input into the first initial model to obtain the second output data of the first initial model.

[0170] The aforementioned second output data may be the data output by the first initial model after the first operation information is input into the first initial model.

[0171] Then, the model parameters of the first initial model are adjusted using the following steps (including the first step and the second step) to train the first initial model:

[0172] The first step involves adjusting the model parameters of the first initial model based on the second operation feature and the second operation feature, assuming the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to the same user category. The first loss function is positively correlated with the target similarity, which is the similarity between the second output data and the second operation feature.

[0173] The second step involves adjusting the model parameters of the first initial model when the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to different user categories. This is done using a second loss function based on the second operation feature and the second output data. The value of the second loss function is negatively correlated with the target similarity, which is the similarity between the second output data and the second operation feature.

[0174] It is understandable that in the above application scenarios, when the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to the same user category, a first loss function whose function value is positively correlated with the target similarity is used to adjust the model parameters of the first initial model, thereby enabling the first model to learn the commonalities among the operation features of users of the same user category; when the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to different user categories, a second loss function whose function value is negatively correlated with the target similarity is used to adjust the model parameters of the first initial model, thereby enabling the first model to learn the differences among the operation features of users of different user categories, thereby further improving the accuracy of information push.

[0175] In some of the above application scenarios, the first loss function is: L = d(H ua H ub The second loss function is: L = max(0, md(H)). ua H ub )).

[0176] Among them, H ua H represents the second output data. ub Characterizing the second operational feature, d(H) ua H ub The similarity of the target is represented by m, which is a preset threshold.

[0177] It is understandable that, in the above situation, using the first and second loss functions as described above to train the first model can further improve the accuracy of information push.

[0178] In some examples of the above situations, the initial model parameters of the second initial model are the same as those of the first initial model, and the model structure of the second initial model is the same as that of the first initial model.

[0179] It is understandable that in the above example, the same initial model can be used to train the first model and the second model respectively. Thus, the first model can be used to obtain the first user's first operation features more accurately, thereby further improving the accuracy of information push.

[0180] The following describes the embodiments of this disclosure by way of example. However, it should be noted that the embodiments of this disclosure may have the features described below, but the following description does not constitute a limitation on the scope of protection of the embodiments of this disclosure.

[0181] In this example, users whose number of actions (N) within 90 days is less than or equal to 50 are considered sparse users (represented by UB, i.e., the first user mentioned above), while others (i.e., users whose number of actions (N) within 90 days is greater than 50) are considered intensive users (represented by UA, i.e., the second user mentioned above). Here, the number of actions (N) represents the number of actions (i.e., the first or second operation mentioned above) within 90 days. These actions refer to the set of user behaviors such as searching and watching. For example, if a user used the search function 3 times and watched 20 videos in the target application over the past 90 days, then N for that user would be 23.

[0182] Here, the preset frequency threshold is 50 times every 90 days.

[0183] Specifically, refer to Figure 3B , Figure 3B This is a schematic diagram of the structure of the second model involved in an information push method provided in this embodiment. First, the user group (i.e., the users mentioned above) can be divided into T categories according to business needs, resulting in T user categories. The user agent (UA) behavior of each category (i.e., the second operation information mentioned above) is fed into the model (Model A, i.e., the second initial model of the second model mentioned above) for training. The training objective is the user group category to which this user belongs (i.e., the user category information mentioned above), and the training objective is a T-classification model (i.e., the second model mentioned above). Assume that the features of each UA user behavior (i.e., the first output data mentioned above) are F. ua ∈R s*d s represents the number of user actions, d represents the feature dimension of each action, and this feature is fed into a fully connected DNN (Deep Neural Network) to train a T-classification model. y' = Softmax(relu(F ua W1)O)∈R T y' is the model's prediction result, W1∈R d*wd These are trainable linear parameters, where wd is the number of hidden layers in the linear layer, and O∈R. wd*T It is the output layer, and the loss function is the cross-entropy function. y' is the predicted true class of the sample.

[0184] Then, refer to Figure 3C , Figure 3C This is a training diagram of the first model involved in an information push method provided in this embodiment. The same model (Model B, i.e., the aforementioned first initial model, the first initial model and the second initial model do not share parameters) is initialized. The input of this model is UB behavior, and the UB user behavior features (i.e., the aforementioned first operation features) are F. ub ∈R s*dSince s is less than N, therefore, for the eigenvector F ub Its feature elements after N are 0. ModelA and ModelB are combined into a dual-tower model and jointly trained. The input to ModelA is random UA behavior (i.e., the first operation information mentioned above), and the input to ModelB is random UB behavior (i.e., the second operation information mentioned above). Then, the feature H output from the hidden layer of the DNN is... ua =relu(F ua W1)∈R d*wd and H ub =relu(F ub W2)∈R d*wd , (W2∈R d*wd (These are the trainable linear parameters of ModelB), and training is performed using the pairwise loss function: If the selected UA and UB populations belong to the same population category, then the loss function used (i.e., the first loss function mentioned above) is: L = d(H ua H ub If the selected UA and UB populations belong to different population categories, the loss function used (i.e., the second loss function mentioned above) is: L = max(0, md(H) ua H ub )).

[0185] Where d is the distance metric function, which is the cosine distance, d(H ua H ub The similarity between targets is represented by m, where m is set as a threshold, and here m = 0.6.

[0186] Then, refer to Figure 3D To push information. Figure 3D This is a flowchart illustrating another information push method provided in this embodiment. Specifically, if the user is a User Agent (UA), the operation information (i.e., the second operation information mentioned above) F is input into Model A to obtain the user characteristics (i.e., the second operation characteristics mentioned above) H; otherwise, the operation information (i.e., the first operation characteristics mentioned above) F is input into Model B to obtain the user characteristics of sparse behavior (i.e., the first operation characteristics mentioned above) H. The user characteristic model in the diagram can be either Model A or Model B. Then, this feature H and the features of the information to be pushed in the set of information to be pushed (i.e., the query statement features mentioned in the diagram) are combined and placed into a ranking model for prediction to obtain a score.

[0187] The first step is to obtain the vector features v of the user's current search query. q ∈R d .

[0188] The second step is to obtain the user's vector features u∈R. ud .

[0189] The third step is to obtain the feature vector Q = [v] of the information to be pushed from the set of information to be pushed. q1 ,…,v qi ,…,v qn ], 1≤i≤n.

[0190] The fourth step is to multiply the user's vector features with the feature vector Q of the information to be pushed in the push information set to obtain the attention matrix A = uWQ. T , where W∈R ud×d , A∈R 1×n T represents the transpose.

[0191] The fifth step is to calculate the score of each message to be pushed using the attention matrix A.

[0192] Step 6: Return the m highest-scoring messages to be pushed as the recall results (i.e., the target push messages).

[0193] Therefore, in the above example, by combining the semantics of the information to be pushed, we can fully explore the semantic information of the information itself, find videos that sparse users may want to watch and wait for push information, improve the recall effect, and thus improve the effect of search recommendation.

[0194] It should be noted that, in addition to the contents described above, this embodiment may also include... Figure 1 and / or Figure 2 The technical features described in the corresponding embodiments, thereby achieving Figure 1 and / or Figure 2 For details on the technical effects of the information push method shown, please refer to [link / reference]. Figure 1 and / or Figure 2 The relevant descriptions are presented concisely and will not be elaborated upon here.

[0195] The information push method provided in this disclosure determines the target push information by using a first model trained based on second operation information, thereby further improving the accuracy of information push.

[0196] Figure 4 This is a schematic diagram of the structure of an information push device provided in an embodiment of this disclosure. Specifically, it includes:

[0197] The acquisition unit 401 is used to acquire first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold.

[0198] The determining unit 402 is configured to determine, based on the first operation information and the second operation information of the second operation performed by the second user through the target application, target push information from a predetermined set of information to be pushed, target push information for pushing to the user terminal of the first user, wherein the execution frequency of the second operation by the second user is greater than the preset frequency threshold.

[0199] The push unit 403 is used to push the target push information to the user terminal.

[0200] In one possible implementation, the step of determining target push information for pushing to the user terminal of the first user from a predetermined set of information to be pushed, based on the first operation information and the second operation information performed by the second user through the target application, includes:

[0201] Based on the first operation information and the second operation information of the second operation performed by the second user through the target application, a first operation feature of the first user is generated.

[0202] Based on the first operational feature, target push information for pushing to the first user's terminal is determined from a pre-determined set of information to be pushed.

[0203] In one possible implementation, generating the first operation feature of the first user based on the first operation information and the second operation information of the second operation performed by the second user through the target application includes:

[0204] The first operation information is input into a pre-trained first model to obtain the first operation feature of the first user. The first model represents the correspondence between the first operation information and the first operation feature. The first model is trained based on the second operation information of the second operation performed by the second user through the target application.

[0205] In one possible implementation, the first model is trained in the following manner:

[0206] The second operation information is input into a pre-trained second model to obtain the second operation feature of the second user, wherein the second model represents the correspondence between the second operation information and the second operation feature;

[0207] Based on the first operation information and the second operation features, the model parameters of the first initial model of the first model are adjusted to train the first initial model;

[0208] The first initial model that has been trained is identified as the first model.

[0209] In one possible implementation, the second model is trained in the following manner:

[0210] Obtain a training sample set, wherein the training samples in the training sample set include second operation information and corresponding user category information, and the user category information represents the user category of the user executing the operation indicated by the second operation information;

[0211] A machine learning algorithm is used to train a second initial model of the second model based on the second operation information and the user category information;

[0212] The trained second initial model is then identified as the second model.

[0213] In one possible implementation, the step of employing a machine learning algorithm to train a second initial model of the second model based on the second operation information and the user category information includes:

[0214] Obtain the second initial model of the second model;

[0215] The second operation information is input into the second initial model to obtain the first output data of the second initial model;

[0216] The first output data is input into the fully connected layer to obtain the actual output data of the fully connected layer;

[0217] Using machine learning algorithms, the model parameters of the second initial model are adjusted based on the user category information and the actual output data to train the second initial model.

[0218] In one possible implementation, adjusting the model parameters of the first initial model based on the first operation information and the second operation feature to train the first initial model includes:

[0219] Obtain the first initial model of the first model;

[0220] The first operation information is input into the first initial model to obtain the second output data of the first initial model;

[0221] The following steps are used to adjust the model parameters of the first initial model of the first model in order to train the first initial model:

[0222] When the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to the same user category, a first loss function is used to adjust the model parameters of the first initial model based on the second operation feature and the second output data. The function value of the first loss function is positively correlated with the target similarity, and the target similarity is the similarity between the second output data and the second operation feature.

[0223] When the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to different user categories, a second loss function is used to adjust the model parameters of the first initial model based on the second operation feature and the second output data. The function value of the second loss function is negatively correlated with the target similarity, which is the similarity between the second output data and the second operation feature.

[0224] In one possible implementation,

[0225] The first loss function is: L = d(H) ua H ub );

[0226] The second loss function is: L = max(0, md(H) ua H ub ));

[0227] Among them, H ua H represents the second output data. ub Characterizing the second operational feature, d(H) ua H ub The similarity of the target is represented by m, which is a preset threshold.

[0228] In one possible implementation, the initial model parameters of the second initial model are the same as those of the first initial model, and the model structure of the second initial model is the same as that of the first initial model.

[0229] The information push device provided in this embodiment can be as follows: Figure 4 The information push device shown can perform the following: Figure 1-3A All steps of the information push method in China, thereby achieving Figure 1-3A For details on the technical effects of the information push method shown, please refer to [link / reference]. Figure 1-3A The relevant descriptions are presented concisely and will not be elaborated upon here.

[0230] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. Figure 5The illustrated electronic device 500 includes at least one processor 501, a memory 502, at least one network interface 504, and other user interfaces 503. The various components in the electronic device 500 are coupled together via a bus system 505. It is understood that the bus system 505 is used to implement communication between these components. In addition to a data bus, the bus system 505 also includes a power bus, a control bus, and a status signal bus. However, for clarity, ... Figure 5 The general designated all buses as Bus System 505.

[0231] The user interface 503 may include a display, keyboard, or clicking device (e.g., mouse, trackball, touchpad, or touchscreen).

[0232] It is understood that the memory 502 in this embodiment of the present disclosure may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory may be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), and Direct Rambus RAM (DRRAM). The memory 502 described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0233] In some implementations, memory 502 stores elements, executable units or data structures, or subsets thereof, or extended sets thereof: operating system 5021 and application program 5022.

[0234] The operating system 5021 includes various system programs, such as the framework layer, core library layer, and driver layer, used to implement various basic business functions and handle hardware-based tasks. The application program 5022 includes various applications, such as a media player and a browser, used to implement various application functions. The program implementing the method of this embodiment can be included in the application program 5022.

[0235] In this embodiment, by calling the program or instructions stored in memory 502, specifically the program or instructions stored in application program 5022, processor 501 executes the method steps provided in each method embodiment, including, for example:

[0236] Obtain first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold;

[0237] Based on the first operation information and the second operation information of the second operation performed by the second user through the target application, target push information for pushing to the user terminal of the first user is determined from a predetermined set of information to be pushed, wherein the execution frequency of the second user for the second operation is greater than the preset frequency threshold.

[0238] The target push information is pushed to the user terminal.

[0239] The methods disclosed in the above embodiments of this disclosure can be applied to or implemented by processor 501. Processor 501 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware or by instructions in the form of software in processor 501. The processor 501 may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this disclosure. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this disclosure can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software units in the decoding processor. The software units may be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 502. Processor 501 reads the information in memory 502 and, in conjunction with its hardware, completes the steps of the above method.

[0240] It is understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described above, or combinations thereof.

[0241] For software implementation, the techniques described herein can be implemented by units that perform the functions described above. The software code can be stored in memory and executed by a processor. The memory can be implemented within the processor or external to the processor.

[0242] The electronic device provided in this embodiment may be as follows: Figure 5 The electronic device shown can perform the following: Figure 1-3A All steps of the information push method in China, thereby achieving Figure 1-3A For details on the technical effects of the information push method shown, please refer to [link / reference]. Figure 1-3A The relevant descriptions are presented concisely and will not be elaborated upon here.

[0243] This disclosure also provides a storage medium (computer-readable storage medium). This storage medium stores one or more programs. The storage medium may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as read-only memory, flash memory, hard disk, or solid-state drive; the memory may also include combinations of the above types of memory.

[0244] When one or more programs in the storage medium can be executed by one or more processors to implement the above-mentioned information push method executed on the electronic device side.

[0245] The processor described above is used to execute the information push program stored in the memory to implement the following steps of the information push method executed on the electronic device side:

[0246] Obtain first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold;

[0247] Based on the first operation information and the second operation information of the second operation performed by the second user through the target application, target push information for pushing to the user terminal of the first user is determined from a predetermined set of information to be pushed, wherein the execution frequency of the second user for the second operation is greater than the preset frequency threshold.

[0248] The target push information is pushed to the user terminal.

[0249] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0250] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented in hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0251] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this disclosure. It should be understood that the above description is only a specific embodiment of this disclosure and is not intended to limit the scope of protection of this disclosure. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. An information push method, characterized in that, The method includes: Obtain first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold. Based on the first operation information and the second operation information of the second operation performed by the second user through the target application, target push information for pushing to the user terminal of the first user is determined from a predetermined set of information to be pushed, including: based on the first operation information and the second operation information of the second operation performed by the second user through the target application, generating the first operation feature of the first user, including: inputting the first operation information into a pre-trained first model to obtain the first operation feature of the first user, wherein the first model represents the correspondence between the first operation information and the first operation feature, and the first model is trained based on the second operation information of the second operation performed by the second user through the target application; Based on the first operation feature, target push information for pushing to the first user's terminal is determined from a predetermined set of information to be pushed; wherein, the second user's execution frequency for the second operation is greater than the preset frequency threshold. The first model is trained as follows: the second operation information is input into a pre-trained second model to obtain the second operation feature of the second user, wherein the second model represents the correspondence between the second operation information and the second operation feature; based on the first operation information and the second operation feature, the model parameters of the first initial model of the first model are adjusted to train the first initial model; the trained first initial model is determined as the first model. The target push information is pushed to the user terminal.

2. The method according to claim 1, characterized in that, The second model was trained in the following manner: Obtain a training sample set, wherein the training samples in the training sample set include second operation information and corresponding user category information, and the user category information represents the user category of the user executing the operation indicated by the second operation information; A machine learning algorithm is used to train a second initial model of the second model based on the second operation information and the user category information; The trained second initial model is then identified as the second model.

3. The method according to claim 2, characterized in that, The step of employing a machine learning algorithm to train a second initial model of the second model based on the second operation information and the user category information includes: Obtain the second initial model of the second model; The second operation information is input into the second initial model to obtain the first output data of the second initial model; The first output data is input into the fully connected layer to obtain the actual output data of the fully connected layer; Using machine learning algorithms, the model parameters of the second initial model are adjusted based on the user category information and the actual output data to train the second initial model.

4. The method according to any one of claims 1-3, characterized in that, The step of adjusting the model parameters of the first initial model based on the first operation information and the second operation features to train the first initial model includes: Obtain the first initial model of the first model; The first operation information is input into the first initial model to obtain the second output data of the first initial model; The following steps are used to adjust the model parameters of the first initial model of the first model in order to train the first initial model: When the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to the same user category, a first loss function is used to adjust the model parameters of the first initial model based on the second operation feature and the second output data. The function value of the first loss function is positively correlated with the target similarity, and the target similarity is the similarity between the second output data and the second operation feature. When the user executing the operation indicated by the first operation information and the user executing the operation indicated by the second operation information belong to different user categories, a second loss function is used to adjust the model parameters of the first initial model based on the second operation feature and the second output data. The function value of the second loss function is negatively correlated with the target similarity, which is the similarity between the second output data and the second operation feature.

5. The method according to claim 4, characterized in that, The first loss function is: L = d(H) ua H ub ); The second loss function is: ; Among them, H ua H represents the second output data. ub Characterizing the second operational feature, d(H) ua H ub The similarity of the target is represented by m, which is a preset threshold.

6. The method according to claim 5, characterized in that, The initial model parameters of the second initial model are the same as those of the first initial model, and the model structure of the second initial model is the same as that of the first initial model.

7. An information push device, characterized in that, The device includes: The acquisition unit is used to acquire first operation information of a first operation performed by a first user through a target application, wherein the execution frequency of the first user for the first operation is less than or equal to a preset frequency threshold. The determining unit is configured to determine target push information for pushing to the user terminal of the first user from a predetermined set of information to be pushed, based on the first operation information and the second operation information of the second operation performed by the second user through the target application. Specifically, it is configured to generate a first operation feature of the first user based on the first operation information and the second operation information of the second operation performed by the second user through the target application. Specifically, it is configured to input the first operation information into a pre-trained first model to obtain the first operation feature of the first user, wherein the first model represents the correspondence between the first operation information and the first operation feature, and the first model is trained based on the second operation information of the second operation performed by the second user through the target application. Based on the first operation feature, target push information for pushing to the first user's terminal is determined from a predetermined set of information to be pushed; wherein, the second user's execution frequency for the second operation is greater than the preset frequency threshold. The determining unit is further configured to input the second operation information into a pre-trained second model to obtain the second operation feature of the second user, wherein the second model represents the correspondence between the second operation information and the second operation feature; adjust the model parameters of the first initial model of the first model based on the first operation information and the second operation feature to train the first initial model; and determine the trained first initial model as the first model. The push unit is used to push the target push information to the user terminal.

8. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing a computer program stored in the memory, wherein when the computer program is executed, it implements the method described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1-6.

Images