Target user determination method, information pushing method, device and electronic equipment

Abstract

The application relates to the technical field of artificial intelligence, and discloses a target user determination method, an information pushing method, a target user determination device, an information pushing device and electronic equipment. The target user determination method comprises the following steps: obtaining to-be-pushed information, and obtaining user features of a plurality of candidate users for the to-be-pushed information; based on the to-be-pushed information and the user features, a prediction model is used to determine a predicted preference degree of each candidate user for the to-be-pushed information; the prediction model is obtained by training a first initial model based on the to-be-pushed information and sample user features for the to-be-pushed information; the first initial model is obtained by training based on at least one historical pushing information and historical sample user features corresponding to each historical pushing information; and at least one target user is selected from the plurality of candidate users based on the predicted preference degree of each candidate user for the to-be-pushed information. The target user determination method provided by the application can more accurately determine the corresponding target user.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and more specifically, to a method for determining a target user, an information push method, an apparatus, and an electronic device. Background Technology

[0002] With the development of information technology, in order to promote products, advertise brands, or release information that the public needs to know, it is often necessary to push information to users, for example, through advertising. However, currently, the information to be pushed is not targeted at all members of the public; some is only targeted at a portion of the public.

[0003] Currently, when pushing push notifications, the model is usually trained directly based on the notification to be pushed. The resulting model is only applicable to the current notification and does not take into account the model's generalization ability. Therefore, the user positioning may be inaccurate for different notifications to be pushed. Summary of the Invention

[0004] The purpose of this application is to more accurately identify target users, and the following technical solution is proposed:

[0005] Firstly, a method for identifying target users is provided, including:

[0006] Obtain the information to be pushed, and obtain the user characteristics of multiple candidate users for the information to be pushed;

[0007] Based on the information to be pushed and user characteristics, a prediction model is used to determine the predicted preference of each candidate user for the information to be pushed; the predicted preference is related to the probability of the candidate user selecting the information to be pushed.

[0008] The prediction model is trained on the first initial model based on the information to be pushed and the sample user features of the information to be pushed; the first initial model is trained on at least one historical push information and the historical sample user features corresponding to each historical push information.

[0009] Based on the predicted preference of each candidate user for the information to be pushed, at least one target user is selected from multiple candidate users.

[0010] Secondly, an information push method is provided, including:

[0011] Obtain the information to be pushed, and obtain the user characteristics of multiple candidate users for the information to be pushed;

[0012] Based on the information to be pushed and user characteristics, a prediction model is used to determine the predicted preference of each candidate user for the information to be pushed; the predicted preference is related to the probability of the candidate user selecting the information to be pushed.

[0013] The prediction model is trained on the first initial model based on the information to be pushed and the sample user features of the information to be pushed; the first initial model is trained on at least one historical push information and the historical sample user features corresponding to each historical push information.

[0014] Based on the predicted preference of each candidate user for the information to be pushed, at least one target user is selected from multiple candidate users;

[0015] Push the information to be pushed to the target user.

[0016] Thirdly, a device for determining a target user is provided, comprising:

[0017] The first acquisition module is used to acquire the information to be pushed and to acquire the user characteristics of multiple candidate users for the information to be pushed.

[0018] The first prediction module is used to determine the predicted preference of each candidate user for the information to be pushed based on the information to be pushed and user characteristics through a prediction model; the predicted preference is related to the probability of the candidate user selecting the information to be pushed.

[0019] The prediction model is trained on the first initial model based on the information to be pushed and the sample user features of the information to be pushed; the first initial model is trained on at least one historical push information and the historical sample user features corresponding to each historical push information.

[0020] The first determining module is used to select at least one target user from multiple candidate users based on the predicted preference of each candidate user for the information to be pushed.

[0021] Fourthly, an information push device is provided, comprising:

[0022] The second acquisition module is used to acquire the information to be pushed and to acquire the user characteristics of multiple candidate users for the information to be pushed.

[0023] The second prediction module is used to determine the predicted preference of each candidate user for the information to be pushed based on the information to be pushed and user characteristics through a prediction model; the predicted preference is related to the probability of the candidate user selecting the information to be pushed.

[0024] The prediction model is trained on the first initial model based on the information to be pushed and the sample user features of the information to be pushed; the first initial model is trained on at least one historical push information and the historical sample user features corresponding to each historical push information.

[0025] The second determining module is used to select at least one target user from multiple candidate users based on the predicted preference of each candidate user for the information to be pushed.

[0026] The push module is used to push information to the target user.

[0027] Fifthly, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for determining a target user as described in the first aspect of this application.

[0028] In a sixth aspect, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the information push method shown in the second aspect of this application.

[0029] In a seventh aspect, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the method for determining a target user as described in the first aspect of this application.

[0030] Eighthly, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the information push method shown in the second aspect of this application.

[0031] The beneficial effects of the technical solution provided in this application are:

[0032] By training a first initial model based on at least one historical push notification and the historical sample user features corresponding to each historical push notification, the generalization ability of the first initial model can be improved, that is, it has the ability to predict different push notifications. Then, based on the push notification and the sample user features for the push notification, the first initial model is trained to obtain a prediction model. The prediction model can have a more accurate prediction ability for the push notification, so that while having generalization ability, the prediction model can more accurately identify the corresponding target user for the push notification.

[0033] Additional aspects and advantages of this application will be set forth in part in the description which follows, and will become apparent from the description or may be learned by practice of this application. Attached Figure Description

[0034] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0035] Figure 1 An application environment diagram for a method for determining a target user provided in an embodiment of this application;

[0036] Figure 2 A flowchart illustrating a method for determining a target user provided in an embodiment of this application;

[0037] Figure 3 A schematic diagram illustrating a scheme for determining a target user as provided in an embodiment of this application;

[0038] Figure 4 A schematic diagram illustrating a training prediction model provided in an embodiment of this application;

[0039] Figure 5 A schematic diagram illustrating the scheme for determining the parameters of the second model during each round of training, as provided in the embodiments of this application;

[0040] Figure 6 A schematic diagram illustrating a scheme for training a first initial model according to an embodiment of this application;

[0041] Figure 7 A flowchart illustrating an information push method provided in an embodiment of this application;

[0042] Figure 8 This is a schematic diagram of the structure of a target user determination device provided in an embodiment of this application;

[0043] Figure 9 This is a schematic diagram of the structure of an information push device provided in an embodiment of this application;

[0044] Figure 10 This is a schematic diagram of the structure of an electronic device for identifying a target user, provided as an embodiment of this application. Detailed Implementation

[0045] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0046] Those skilled in the art will understand that, unless explicitly stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in the specification of this application means the presence of features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0047] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0048] Artificial intelligence (AI) is the theory, methods, technology, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence. AI studies the design principles and implementation methods of various intelligent machines, enabling them to possess the functions of perception, reasoning, and decision-making.

[0049] Artificial intelligence (AI) is a comprehensive discipline encompassing a wide range of fields, including both hardware and software technologies. Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies primarily include computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0050] Machine Learning (ML) is a multidisciplinary field involving probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. It specifically studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to endow computers with intelligence; its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and instructional learning.

[0051] Blockchain is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and cryptographic algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include an underlying platform, a platform product service layer, and an application service layer.

[0052] The underlying blockchain platform can include processing modules such as user management, basic services, smart contracts, and operational monitoring. The user management module is responsible for managing the identity information of all blockchain participants, including maintaining public and private key generation (account management), key management, and maintaining the correspondence between user real identities and blockchain addresses (access management). Furthermore, under authorization, it monitors and audits transactions of certain real identities and provides risk control rule configuration (risk control audit). The basic services module is deployed on all blockchain node devices to verify the validity of business requests. After consensus is reached on valid requests, they are recorded in storage. For a new business request, the basic services first perform interface adaptation parsing and authentication (interface adaptation), and then encrypt the business information through a consensus algorithm (consensus management). After encryption, the data is transmitted completely and consistently to the shared ledger (network communication) and recorded and stored. The smart contract module is responsible for contract registration, issuance, triggering, and execution. Developers can define contract logic using a programming language and publish it to the blockchain (contract registration). According to the contract terms, the key or other events are invoked to trigger execution and complete the contract logic. It also provides functions for contract upgrades and cancellations. The operation monitoring module is mainly responsible for deployment, configuration modification, contract settings, cloud adaptation, and real-time status visualization output during product release, such as alarms, monitoring network conditions, and monitoring the health status of node devices.

[0053] The platform's product service layer provides the basic capabilities and implementation frameworks for typical applications. Developers can leverage these basic capabilities, along with the specific characteristics of their business needs, to implement blockchain-based business logic. The application service layer provides blockchain-based application services to business stakeholders.

[0054] The information to be pushed and the user characteristics of candidate users provided in this application embodiment can be pre-stored in the blockchain. When determining the target user, the server or terminal obtains the information to be pushed and the user characteristics from the blockchain.

[0055] The solutions provided in this application relate to the technology for determining target users in artificial intelligence, which will be specifically illustrated through the following embodiments.

[0056] In the mobile internet era, marketers always aim to present information such as products, content, or advertisements to potential customers through media channels. Marketing campaigns, in particular, aim to promote or sell a product, increasing its visibility and potential purchases. For example, in WeChat's "Discover" feature, promoted products might be videos or advertisements. Given a candidate set, also known as a seed set (where users are identified by experts as potentially interested in the product), look-alike techniques (candidate set expansion techniques) can find more potential users who might be interested in the product. Many companies have already adopted this technique in their marketing, such as Google, LinkedIn, and Ant Financial. A company might conduct hundreds of marketing campaigns daily to promote different products, and each specific campaign might only have a very small seed set, making it difficult to train a customized model from scratch for each campaign. We propose a meta-learning-based look-alike method that pre-trains a generalized model on data from existing marketing campaigns. For each new marketing campaign, a robust customized model can be easily obtained based on this generalized model.

[0057] Current look-alike methods train the model directly on the new information to be pushed. This approach does not consider the model's generalization ability. In fact, different information to be pushed, such as different marketing campaigns, vary greatly, making generalization ability extremely important.

[0058] The method, apparatus, electronic device, and computer-readable storage medium for determining the target user provided in this application are intended to solve the above-mentioned technical problems of the prior art.

[0059] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0060] The method for determining the target user provided in this application can be applied to, for example... Figure 1In the application environment shown, specifically, server 101 obtains the push-to-be-pushed information from terminal 102 and acquires user characteristics sent by multiple terminals 103; based on the push-to-be-pushed information and the user characteristics, server 101 determines the predicted preference degree of each candidate user for the push-to-be-pushed information through a prediction model; based on the predicted preference degree of each candidate user for the push-to-be-pushed information, server 101 selects at least one target user from multiple candidate users and pushes the push-to-be-pushed information to the target terminal 103 corresponding to the target user.

[0061] Those skilled in the art will understand that the term "terminal" as used herein can refer to mobile phones, tablets, computers, PDAs (Personal Digital Assistants), MIDs (Mobile Internet Devices), etc.; and "server" can be implemented using a standalone server or a server cluster composed of multiple servers.

[0062] Understandable Figure 1 This describes an example application scenario and does not limit the application scenario of the method for determining the target user in this application. In the above scenario, the server determines the target user. In other application scenarios, the terminal may determine the target user for the information to be pushed. In addition, in the above application scenario, terminal 102 is a computer and terminal 103 is a mobile phone. In other application scenarios, there is no limitation on the device type of terminal 102 and terminal 103. Terminal 102 and terminal 103 may be other types of devices, such as other types of terminals, or servers. For example, in other application scenarios, server 102 may send push information to server 101, and terminal 103 may be a tablet computer, etc.

[0063] This application provides one possible implementation method, such as... Figure 2 As shown, a method for determining target users is provided, which can be applied to... Figure 1 Taking the server in the example, the following steps may be included:

[0064] Step S201: Obtain the information to be pushed and obtain the user characteristics of multiple candidate users for the information to be pushed.

[0065] Among them, the information to be pushed is information intended to be pushed to target users among the candidate users, which may be information that the target users may be interested in; the information to be pushed can be in various forms, such as advertising information for marketing products, or popular science information that users may be interested in, etc.

[0066] User characteristics can include user attributes such as age, gender, and occupation, as well as user profiles. User characteristics can be used to determine a user's preference for different types of information to be pushed to them.

[0067] Specifically, for the information to be pushed, multiple candidate users can be obtained in a targeted manner. The tags of the information to be pushed can be determined first, and user groups that match the tags can be selected as candidate users. For example, the information to be pushed can have targeted age group tags, and user groups that match the age group can be obtained as candidate users.

[0068] Step S202: Based on the information to be pushed and user characteristics, determine the predicted preference degree of each candidate user for the information to be pushed through a prediction model.

[0069] Among them, the predicted preference is related to the probability of a candidate user selecting the information to be pushed.

[0070] Specifically, the predicted preference can be positively correlated with the probability of a user selecting the information to be pushed to them.

[0071] For example, the predicted preference can include "1" for interested or "0" for uninterested. If the predicted probability of a user selecting the information to be pushed is greater than or equal to 50%, the predicted preference is "1" for interested. If the predicted probability of a user selecting the information to be pushed is less than 50%, the predicted preference is "0" for uninterested.

[0072] Furthermore, the predicted preference can be directly proportional to the probability of a user selecting the information to be pushed to them; even further, the predicted preference can be the probability of a user selecting the information to be pushed to them.

[0073] The prediction model is trained on the first initial model based on the information to be pushed and the sample user features of the information to be pushed. The first initial model is trained on at least one historical push information and the historical sample user features corresponding to each historical push information. The training process of the prediction model will be described in detail below.

[0074] Step S203: Based on the predicted preference of each candidate user for the information to be pushed, at least one target user is selected from multiple candidate users.

[0075] Specifically, candidate users whose predicted preferences meet preset conditions can be set as target users.

[0076] In some implementations, the candidate users with the largest preset number of predicted preferences can be selected from all candidate users and set as target users. That is, the predicted preferences of all candidate users can be sorted in descending order, and the candidate users with the largest preset number of predicted preferences can be set as target users. For example, if there are 1,000 candidate users, the predicted preferences of the 1,000 candidate users can be determined, and the 100 candidate users with the largest 100 predicted preferences can be set as target users.

[0077] In some implementations, candidate users whose predicted preference is greater than a preset threshold can be set as target users. For example, if the predicted preference is proportional to the probability of a user selecting the information to be pushed, candidate users whose predicted preference is greater than 50% can be directly set as target users.

[0078] In other implementations, candidate users whose predicted preference is a specific preference can be set as target users. For example, if the predicted preference includes "1" for interest or "0" for disinterest, then candidate users whose predicted preference is "1" for interest can be set as target users.

[0079] like Figure 3 As shown, the user features of the information to be pushed and the candidate users corresponding to the information to be pushed are input into the prediction model to obtain the predicted preference of the candidate users for the information to be pushed. The target user is determined from the candidate users based on the predicted preference. The prediction model is obtained by training the first initial model based on the information to be pushed and the sample user features for the information to be pushed. The first initial model is obtained by training at least one round based on at least one historical push information and the historical sample user features corresponding to each historical push information.

[0080] In the above embodiments, by training a first initial model based on at least one historical push information and the historical sample user features corresponding to each historical push information, the generalization ability of the first initial model can be improved, that is, it has the ability to predict different push information. Then, based on the information to be pushed and the sample user features for the information to be pushed, the first initial model is trained to obtain a prediction model. The obtained prediction model can have a more accurate prediction ability for the information to be pushed, so that while having generalization ability, the prediction model can more accurately determine the corresponding target user for the information to be pushed.

[0081] The training process of the target user determination model will be further explained below with reference to the accompanying drawings and specific embodiments.

[0082] This application provides one possible implementation method, such as... Figure 4 As shown, the prediction model can be trained in the following way:

[0083] Step S401: Obtain at least one historical push information and obtain historical sample user characteristics for each historical push information.

[0084] Among them, the historical sample user characteristics can be set with corresponding sample user preferences for historical push information.

[0085] Specifically, historical sample user features can include positive sample user features and negative sample user features. Positive sample user features correspond to sample user preferences that indicate interest in historical push information, while negative sample user features correspond to sample user preferences that indicate disinterest in historical push information.

[0086] Step S402: Train the preset second initial model based on at least one historical push information and the historical sample user features of each historical push information to obtain the first initial model.

[0087] Specifically, a first initial model can be obtained by training a preset second initial model for at least one round based on at least one historical push information and the historical sample user characteristics of each historical push information.

[0088] Specifically, step S402 trains a preset second initial model based on at least one historical push information and the historical sample user features of each historical push information to obtain a first initial model, which may include:

[0089] (1) For each historical push information, a second initial model is trained once based on the historical push information and the corresponding historical sample user characteristics.

[0090] (2) The first initial model is obtained by training the preset second initial model at least once based on at least one historical push information and the historical sample user features of each historical push information.

[0091] Specifically, in each round of training, each historical push message and the corresponding historical sample user features are used to train the preset second initial model. For at least one historical push message, at least one round of training is performed, and the first initial model is obtained in the last round of training.

[0092] The following will elaborate on each round of training.

[0093] Specifically, based on historical push information and corresponding historical sample user characteristics, a second initial model is trained in one round, including:

[0094] a. Determine the first model parameters of the second initial model based on historical push information and corresponding historical sample user characteristics.

[0095] The historical sample user features may include the first historical sample user features corresponding to the first user set and the second historical sample user features corresponding to the second user set, wherein the sample users in the second user set are different from the sample users in the first user set.

[0096] In the specific implementation process, a historical user set can be obtained, which includes multiple different sample users. A first user set is extracted from the historical user set, and a second user set is extracted.

[0097] Specifically, both the first user set and the second user set include positive sample users and negative sample users, that is, both the first user set and the second user set include sample users who are interested in the corresponding historical push information and sample users who are not interested.

[0098] Specifically, determining the first model parameters of the second initial model based on historical push information and corresponding historical sample user characteristics may include:

[0099] The second initial model is trained based on historical push information and the characteristics of users in the first historical sample to obtain the parameters of the first model.

[0100] Specifically, the second initial model is trained based on historical push information and the characteristics of the first historical sample users to obtain the parameters of the first model, which may include:

[0101] ① Based on the second initial model, predict the first predicted preference of the first user set for historical push information;

[0102] ② Determine the first loss function based on the first true preference and the first predicted preference of the first user set;

[0103] ③ Train the second initial model based on the first loss function to obtain the parameters of the first model.

[0104] b. Update the parameters of the first model to obtain the second model parameters of the second initial model after this round of training.

[0105] Specifically, updating the parameters of the first model to obtain the second model parameters of the second initial model after this round of training can include:

[0106] b1. Determine the second loss function based on the user characteristics of the second historical sample;

[0107] b2. Update the parameters of the first model based on the second loss function to obtain the parameters of the second model.

[0108] The following will further elaborate on the training process of each round of the second initial model with reference to the embodiments.

[0109] Specifically, each round of training for the second initial model can include the following steps:

[0110] Obtain the first historical sample user features of the first user set, and obtain the second historical sample user features of the second user set; this can be done using D. a [c] This represents the first historical sample user characteristics for the current historical push information c, denoted by D. b [c] This represents the second historical sample user characteristics for the current historical push information c;

[0111] The second initial model predicts the first user set's first predicted preference for historical push information; that is, it inputs the user features of the first historical sample into the second initial model to obtain the first predicted preference.

[0112] Based on the first predicted preference The first true preference degree y corresponding to the first historical sample user characteristics determines the first loss function; specifically, it can be calculated using the following formula:

[0113]

[0114] Where θ represents the first model parameters of the first initial model; La(θ) represents the first loss function; D a [c] This represents the characteristics of the first historical sample users; y represents the first predicted preference; y represents the first true preference.

[0115] The first model parameter θ is determined by minimizing the first loss function La(θ). Specifically, the first model parameter θ can be calculated according to the following formula:

[0116]

[0117] Where, θ [c] α represents the first model parameters of the first initial model for the current historical push information c; α represents the preset first learning rate of the model.

[0118] The above uses the first historical sample user features D for the current historical push information c. a [c] The process of determining the first model parameters of the first initial model can be understood as the process of understanding user features (Understanding phase). The following will use the second historical sample user features D for the current historical push information c. b [c]Determining the second model parameters of the first initial model can also be understood as the process of finding a candidate set (Finding phase), which can specifically include the following steps:

[0119] The second loss function is determined based on the user characteristics of the second historical sample.

[0120] The parameters of the first model are updated based on the second loss function to obtain the parameters of the second model;

[0121] Specific update formulas may include:

[0122]

[0123] By taking the second derivative of formulas (2) and (3), we can obtain:

[0124]

[0125] Where Lb represents the second loss function; β represents the preset second learning rate of the model.

[0126] By combining formulas (3) and (4), the parameters of the first model are updated to obtain the parameters of the second model, i.e., the updated θ. [c] .

[0127] like Figure 5 As shown, in each round of training for the second initial model, the process of understanding user features is performed first: the second initial model is trained based on historical push information and the user features of the first historical sample to obtain the first model parameters; then the process of finding the user set is performed: the first model parameters are updated to obtain the second model parameters of the second initial model after this round of training.

[0128] like Figure 6 As shown, in each round of training for the second initial model, the second initial model is first trained based on historical push information and the characteristics of the first historical sample user to obtain the first model parameters; then the first model parameters are updated to obtain the second model parameters of the second initial model after that round of training; for example, in the first round of training, the first model parameter 1 is obtained based on historical push information 1 and the first historical sample user feature 1, and the first model parameter 1 is updated to obtain the second model parameter 1; the second model parameter 1 is used as the first model parameter 2 in the second round of training, and in each round of training, the second model parameter after that round of training is used as the initial model parameter of the second initial model in the next round of training, and the training is repeated to obtain new first model parameters and second model parameters, until the nth round of training, to obtain the second model parameter n, that is, to obtain the parameters of the first initial model.

[0129] Step S403: Obtain the characteristics of sample users for the information to be pushed.

[0130] Specifically, the sample user characteristics for the information to be pushed can also include positive sample characteristics and negative sample characteristics. Positive sample characteristics correspond to the sample user's interest in the information to be pushed, while negative sample characteristics correspond to the sample user's disinterest in the information to be pushed.

[0131] Step S404: Based on the information to be pushed and the sample user features of the information to be pushed, train the first initial model to obtain the prediction model.

[0132] Specifically, the first initial model is trained based on at least one historical push information and the historical sample user features corresponding to each historical push information. It has predictive ability for different push information and has strong generalization ability. Then, the first model is trained based on the information to be pushed and the sample user features corresponding to the information to be pushed. This makes the prediction model more accurate in predicting the information to be pushed, while also having generalization ability.

[0133] Specifically, the information to be pushed and the characteristics of the sample users corresponding to the information to be pushed can be input into the first initial model to obtain the current predicted preference output by the first initial model; based on the current actual preference of the sample users corresponding to the information to be pushed and the current predicted preference, the parameters of the first initial model are adjusted to obtain the prediction model.

[0134] The aforementioned method for determining target users involves training a first initial model based on at least one historical push notification and the historical sample user features corresponding to each historical push notification. This improves the generalization ability of the first initial model, meaning it has predictive capabilities for different push notifications. Then, the first initial model is trained based on the information to be pushed and the sample user features for that information to be pushed, resulting in a prediction model. This prediction model has a more accurate predictive ability for the information to be pushed, enabling it to more accurately identify the target user for the information to be pushed while maintaining generalization capabilities.

[0135] This application provides one possible implementation method, such as... Figure 7 As shown, an information push method is provided, which may include:

[0136] Step S701: Obtain the information to be pushed and obtain the user characteristics of multiple candidate users for the information to be pushed.

[0137] Step S702: Based on the information to be pushed and user characteristics, determine the predicted preference degree of each candidate user for the information to be pushed through a prediction model; the predicted preference degree is related to the probability of the candidate user selecting the information to be pushed.

[0138] The prediction model is trained on the first initial model based on the information to be pushed and the sample user features of the information to be pushed; the first initial model is trained on at least one historical push information and the historical sample user features corresponding to each historical push information.

[0139] Step S703: Based on the predicted preference of each candidate user for the information to be pushed, at least one target user is selected from multiple candidate users.

[0140] Step S704: Push the information to be pushed to the target user.

[0141] The specific process of the information push method provided in the above embodiments can be found in the above method for determining the target user, and will not be repeated here.

[0142] The aforementioned information push method can obtain a first initial model by training it based on at least one historical push information and the historical sample user features corresponding to each historical push information. This can improve the generalization ability of the first initial model, that is, it has the ability to predict different push information. Then, based on the information to be pushed and the sample user features for the information to be pushed, the first initial model is trained to obtain a prediction model. The obtained prediction model can have a more accurate prediction ability for the information to be pushed. This allows the prediction model to have generalization ability and to more accurately identify the corresponding target users for the information to be pushed, thereby making information push more accurate and more targeted.

[0143] This application provides one possible implementation method, such as... Figure 8 As shown, a target user determination device 80 is provided. The target user determination device 80 may include: a first acquisition module 801, a first prediction module 802, and a first determination module 803, wherein...

[0144] The first acquisition module 801 is used to acquire information to be pushed and to acquire user characteristics of multiple candidate users for the information to be pushed.

[0145] The first prediction module 802 is used to determine the predicted preference of each candidate user for the information to be pushed based on the information to be pushed and user characteristics through a prediction model; the predicted preference is related to the probability of the candidate user selecting the information to be pushed.

[0146] The prediction model is trained on the first initial model based on the information to be pushed and the sample user features of the information to be pushed; the first initial model is trained on at least one historical push information and the historical sample user features corresponding to each historical push information.

[0147] The first determining module 803 is used to select at least one target user from multiple candidate users based on the predicted preference of each candidate user for the information to be pushed.

[0148] This application provides a possible implementation, which also includes a training module for:

[0149] Obtain at least one historical push notification and obtain historical sample user characteristics for each historical push notification;

[0150] The first initial model is obtained by training a preset second initial model based on at least one historical push information and the historical sample user features of each historical push information.

[0151] Obtain the characteristics of sample users for the information to be pushed;

[0152] The first initial model is trained based on the information to be pushed and the sample user characteristics of the information to be pushed, to obtain the prediction model.

[0153] This application embodiment provides a possible implementation method in which the training module, when training a preset second initial model based on at least one historical push information and the historical sample user features of each historical push information to obtain a first initial model, specifically uses:

[0154] For each historical push notification, a second initial model is trained once based on the historical push notification and the corresponding historical sample user characteristics.

[0155] The first initial model is obtained by training the preset second initial model at least once based on at least one historical push information and the historical sample user features of each historical push information.

[0156] This application embodiment provides a possible implementation method in which the training module, when performing one round of training on a preset second initial model based on historical push information and corresponding historical sample user features, is specifically used for:

[0157] The first model parameters of the second initial model are determined based on historical push information and corresponding historical sample user characteristics.

[0158] The parameters of the first model are updated to obtain the second model parameters of the second initial model after this round of training.

[0159] This application provides a possible implementation method, wherein the historical sample user features include the first historical sample user features corresponding to the first user set;

[0160] When determining the first model parameters of the second initial model based on historical push information and corresponding historical sample user features, the training module is specifically used for:

[0161] The second initial model is trained based on historical push information and the characteristics of users in the first historical sample to obtain the parameters of the first model.

[0162] This application embodiment provides a possible implementation method in which the training module, when training the second initial model based on historical push information and the user characteristics of the first historical sample to obtain the first model parameters, is specifically used for:

[0163] Based on the second initial model, predict the first predicted preference of the first user set for historical push information;

[0164] Based on the first true preference and the first predicted preference of the first user set, determine the first loss function;

[0165] The second initial model is trained based on the first loss function to obtain the parameters of the first model.

[0166] This application provides a possible implementation method, wherein the historical sample user features also include the second historical sample user features corresponding to the second user set;

[0167] When the training module updates the parameters of the first model to obtain the second model parameters of the second initial model after this round of training, it is specifically used for:

[0168] The second loss function is determined based on the user characteristics of the second historical sample.

[0169] The parameters of the first model are updated based on the second loss function to obtain the parameters of the second model.

[0170] This application provides a possible implementation method in which the sample users in the second user set are different from the sample users in the first user set; both the first user set and the second user set include positive sample users and negative sample users.

[0171] The aforementioned target user determination device trains a first initial model based on at least one historical push information and the historical sample user features corresponding to each historical push information. This improves the generalization ability of the first initial model, meaning it has predictive ability for different push information. Then, based on the information to be pushed and the sample user features for the information to be pushed, the first initial model is trained to obtain a prediction model. The obtained prediction model has a more accurate predictive ability for the information to be pushed, enabling the prediction model to more accurately determine the corresponding target user for the information to be pushed while having generalization ability.

[0172] The image target user determination device of this disclosure embodiment can execute an image target user determination method provided in the embodiments of this disclosure. The implementation principle is similar. The actions performed by each module in the image target user determination device in each embodiment of this disclosure correspond to the steps in the image target user determination method in each embodiment of this disclosure. For detailed functional descriptions of each module of the image target user determination device, please refer to the descriptions of the corresponding image target user determination methods shown above, which will not be repeated here.

[0173] This application provides one possible implementation method, such as... Figure 9 As shown, an information push device 90 is provided, including a second acquisition module 901, a second prediction module 902, a second determination module 903, and a push module 904, wherein...

[0174] The second acquisition module 901 is used to acquire the information to be pushed and to acquire the user characteristics of multiple candidate users for the information to be pushed.

[0175] The second prediction module 902 is used to determine the predicted preference of each candidate user for the information to be pushed based on the information to be pushed and user characteristics through a prediction model; the predicted preference is related to the probability of the candidate user selecting the information to be pushed.

[0176] The prediction model is trained on the first initial model based on the information to be pushed and the sample user features of the information to be pushed; the first initial model is trained on at least one historical push information and the historical sample user features corresponding to each historical push information.

[0177] The second determining module 903 is used to select at least one target user from multiple candidate users based on the predicted preference of each candidate user for the information to be pushed.

[0178] The push module 904 is used to push the information to be pushed to the target user.

[0179] The aforementioned information push device can train a first initial model based on at least one historical push information and the historical sample user features corresponding to each historical push information. This can improve the generalization ability of the first initial model, meaning it has predictive ability for different push information. Then, based on the information to be pushed and the sample user features for the information to be pushed, the first initial model is trained to obtain a prediction model. The obtained prediction model has a more accurate predictive ability for the information to be pushed. This allows the prediction model to have generalization ability while more accurately identifying the target user for the information to be pushed, thus making the information push more accurate and targeted.

[0180] The image information push device of this disclosure can execute an image information push method provided in the embodiments of this disclosure. The implementation principle is similar. The actions performed by each module in the image information push device in each embodiment of this disclosure correspond to the steps in the image information push method in each embodiment of this disclosure. For detailed functional descriptions of each module of the image information push device, please refer to the descriptions of the corresponding image information push methods shown above, which will not be repeated here.

[0181] Based on the same principles as the methods shown in the embodiments of this disclosure, the embodiments of this disclosure also provide an electronic device, which may include, but is not limited to: a processor and a memory; the memory for storing computer operation instructions; and the processor for executing the target user determination method shown in the embodiments by invoking the computer operation instructions. Compared with the prior art, the target user determination method in this application can have a more accurate predictive ability for the information to be pushed, enabling the prediction model to have generalization ability while more accurately determining the corresponding target user for the information to be pushed.

[0182] Based on the same principles as the methods shown in the embodiments of this disclosure, the embodiments of this disclosure also provide an electronic device, which may include, but is not limited to: a processor and a memory; the memory for storing computer operation instructions; and the processor for executing the information push method shown in the embodiments by invoking the computer operation instructions. Compared with the prior art, the information push method in this application makes information push more accurate and more targeted.

[0183] In one alternative embodiment, an electronic device is provided, such as Figure 10 As shown, Figure 10 The illustrated electronic device 4000 includes a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, for example, via a bus 4002. Optionally, the electronic device 4000 may also include a transceiver 4004. It should be noted that in practical applications, the transceiver 4004 is not limited to one type, and the structure of this electronic device 4000 does not constitute a limitation on the embodiments of this application.

[0184] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 4001 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0185] Bus 4002 may include a pathway for transmitting information between the aforementioned components. Bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 4002 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 10 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0186] The memory 4003 may be ROM (Read Only Memory) or other types of static storage devices capable of storing static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0187] The memory 4003 stores application code that executes the scheme of this application, and its execution is controlled by the processor 4001. The processor 4001 executes the application code stored in the memory 4003 to implement the content shown in the foregoing method embodiments.

[0188] Among them, electronic devices include, but are not limited to: mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), and in-vehicle terminals (such as in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 10 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0189] This application provides a computer-readable storage medium storing a computer program that, when run on a computer, enables the computer to execute the corresponding content in the aforementioned method embodiments. Compared with the prior art, the target user determination method in this application can have a more accurate predictive ability for the information to be pushed, enabling the prediction model to have generalization ability while more accurately identifying the corresponding target user for the information to be pushed.

[0190] This application provides a computer-readable storage medium storing a computer program that, when run on a computer, enables the computer to execute the corresponding content described in the aforementioned method embodiments. Compared with the prior art, the information push method in this application makes information push more accurate and targeted.

[0191] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0192] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can 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 this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can 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.

[0193] 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.

[0194] The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the methods shown in the above embodiments.

[0195] This application provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the following actions:

[0196] Obtain the information to be pushed, and obtain the user characteristics of multiple candidate users for the information to be pushed;

[0197] Based on the information to be pushed and the user characteristics, a prediction model is used to determine the predicted preference degree of each candidate user for the information to be pushed; the predicted preference degree is related to the probability of the candidate user selecting the information to be pushed.

[0198] The prediction model is obtained by training a first initial model based on the information to be pushed and the sample user features of the information to be pushed; the first initial model is obtained by training at least one historical push information and the historical sample user features corresponding to each historical push information.

[0199] Based on the predicted preference of each candidate user for the information to be pushed, at least one target user is selected from multiple candidate users.

[0200] This application provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the following actions:

[0201] Obtain the information to be pushed, and obtain the user characteristics of multiple candidate users for the information to be pushed;

[0202] Based on the information to be pushed and the user characteristics, a prediction model is used to determine the predicted preference degree of each candidate user for the information to be pushed; the predicted preference degree is related to the probability of the candidate user selecting the information to be pushed.

[0203] The prediction model is obtained by training a first initial model based on the information to be pushed and the sample user features of the information to be pushed; the first initial model is obtained by training at least one historical push information and the historical sample user features corresponding to each historical push information.

[0204] Based on the predicted preference of each candidate user for the information to be pushed, at least one target user is selected from multiple candidate users.

[0205] The information to be pushed is sent to the target user.

[0206] Computer program code for performing the operations 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).

[0207] 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.

[0208] The modules described in the embodiments of this disclosure can be implemented in software or in hardware. The name of a module does not necessarily limit the module itself; for example, the first acquisition module can also be described as a "module for acquiring information to be pushed".

[0209] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope 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 concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

Claims

1. A method for determining target users, characterized in that, include: Obtain the information to be pushed, and obtain the user characteristics of multiple candidate users for the information to be pushed; Based on the information to be pushed and the user characteristics, a prediction model is used to determine the predicted preference degree of each candidate user for the information to be pushed; the predicted preference degree is related to the probability of the candidate user selecting the information to be pushed. Based on the predicted preference of each candidate user for the information to be pushed, at least one target user is selected from multiple candidate users. The prediction model is trained in the following manner: Obtain at least one historical push notification and acquire historical sample user features for each historical push notification. The historical sample user features include first historical sample user features corresponding to a first user set and second historical sample user features corresponding to a second user set. Each user set includes multiple sample users. Based on at least one of the historical push notifications and the historical sample user features of each historical push notification, perform at least one round of training operations on a preset second initial model. Use the second initial model after the last round of training operations as the first initial model. The second initial model is trained based on at least one historical push information and the corresponding first historical sample user features to obtain the first model parameters of the second initial model. The first model parameters are then updated based on the at least one historical push information and the corresponding second historical user features to obtain the second model parameters of the second initial model. Obtain sample user characteristics for the information to be pushed; The first initial model is trained based on the information to be pushed and the sample user features of the information to be pushed, to obtain the prediction model.

2. The method for determining the target user according to claim 1, characterized in that, The step of training the second initial model based on at least one historical push information and corresponding first historical sample user features to obtain the first model parameters of the second initial model includes: Based on the second initial model, predict the first predicted preference degree of the first user set for the historical push information; Based on the first true preference and the first predicted preference of the first user set, a first loss function is determined; The second initial model is trained based on the first loss function to obtain the parameters of the first model.

3. The method for determining the target user according to claim 2, characterized in that, The step of updating the first model parameters to obtain the second model parameters of the second initial model includes: The second loss function is determined based on the user characteristics of the second historical sample. The first model parameters are updated based on the second loss function to obtain the second model parameters.

4. The method for determining the target user according to claim 3, characterized in that, The sample users in the second user set are not the same as the sample users in the first user set; both the first user set and the second user set include positive sample users and negative sample users.

5. An information push method, characterized in that, include: Obtain the information to be pushed, and obtain the user characteristics of multiple candidate users for the information to be pushed; Based on the information to be pushed and the user characteristics, a prediction model is used to determine the predicted preference degree of each candidate user for the information to be pushed; the predicted preference degree is related to the probability of the candidate user selecting the information to be pushed. Based on the predicted preference of each candidate user for the information to be pushed, at least one target user is selected from multiple candidate users. The information to be pushed is sent to the target user; The prediction model was trained in the following manner: Obtain at least one historical push notification and acquire historical sample user features for each historical push notification. The historical sample user features include first historical sample user features corresponding to a first user set and second historical sample user features corresponding to a second user set. Each user set includes multiple sample users. Based on at least one of the historical push notifications and the historical sample user features of each historical push notification, perform at least one round of training operations on a preset second initial model. Use the second initial model after the last round of training operations as the first initial model. The second initial model is trained based on at least one historical push information and the corresponding first historical sample user features to obtain the first model parameters of the second initial model. The first model parameters are then updated based on the at least one historical push information and the corresponding second historical user features to obtain the second model parameters of the second initial model. Obtain sample user characteristics for the information to be pushed; The first initial model is trained based on the information to be pushed and the sample user features of the information to be pushed, to obtain the prediction model.

6. A device for determining a target user, characterized in that, include: The first acquisition module is used to acquire information to be pushed and to acquire user characteristics of multiple candidate users for the information to be pushed. The first prediction module is used to determine the predicted preference degree of each candidate user for the information to be pushed based on the information to be pushed and the user characteristics through a prediction model; the predicted preference degree is related to the probability of the candidate user selecting the information to be pushed. The first determining module is used to select at least one target user from multiple candidate users based on the predicted preference of each candidate user for the information to be pushed; The training module is used to acquire at least one historical push notification and to acquire historical sample user features for each historical push notification. The historical sample user features include first historical sample user features corresponding to a first user set and second historical sample user features corresponding to a second user set. Each user set includes multiple sample users. Based on at least one of the historical push notifications and the historical sample user features of each historical push notification, a preset second initial model is trained for at least one round of the following operations. The second initial model after the last round of training is used as the first initial model. The second initial model is trained based on at least one historical push information and the corresponding first historical sample user features to obtain the first model parameters of the second initial model. The first model parameters are then updated based on the at least one historical push information and the corresponding second historical user features to obtain the second model parameters of the second initial model. Obtain sample user characteristics for the information to be pushed; The first initial model is trained based on the information to be pushed and the sample user features of the information to be pushed, to obtain the prediction model.

7. The apparatus according to claim 6, characterized in that, When the training module trains the second initial model based on at least one historical push information and the corresponding first historical sample user features to obtain the first model parameters of the second initial model, it is specifically used for: Based on the second initial model, predict the first predicted preference degree of the first user set for the historical push information; Based on the first true preference and the first predicted preference of the first user set, a first loss function is determined; The second initial model is trained based on the first loss function to obtain the parameters of the first model.

8. The apparatus according to claim 7, characterized in that, When the training module updates the first model parameters to obtain the second model parameters of the second initial model, it is specifically used for: The second loss function is determined based on the user characteristics of the second historical sample. The first model parameters are updated based on the second loss function to obtain the second model parameters.

9. The apparatus according to claim 8, characterized in that, The sample users in the second user set are not the same as the sample users in the first user set; both the first user set and the second user set include positive sample users and negative sample users.

10. An information push device, characterized in that, include: The second acquisition module is used to acquire information to be pushed and to acquire user characteristics of multiple candidate users for the information to be pushed. The second prediction module is used to determine the predicted preference degree of each candidate user for the information to be pushed based on the information to be pushed and the user characteristics through a prediction model; the predicted preference degree is related to the probability of the candidate user selecting the information to be pushed. The second determining module is used to select at least one target user from multiple candidate users based on the predicted preference of each candidate user for the information to be pushed; The push module is used to push the information to be pushed to the target user; The prediction model was trained in the following manner: Obtain at least one historical push notification and acquire historical sample user features for each historical push notification. The historical sample user features include first historical sample user features corresponding to a first user set and second historical sample user features corresponding to a second user set. Each user set includes multiple sample users. Based on at least one of the historical push notifications and the historical sample user features of each historical push notification, perform at least one round of training operations on a preset second initial model. Use the second initial model after the last round of training operations as the first initial model. The second initial model is trained based on at least one historical push information and the corresponding first historical sample user features to obtain the first model parameters of the second initial model. The first model parameters are then updated based on the at least one historical push information and the corresponding second historical user features to obtain the second model parameters of the second initial model. Obtain sample user characteristics for the information to be pushed; The first initial model is trained based on the information to be pushed and the sample user features of the information to be pushed, to obtain the prediction model.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method for determining the target user as described in any one of claims 1-4.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the information push method of claim 5.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method for determining the target user as described in any one of claims 1-4.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the information push method of claim 5.

Images