Model training methods, text push methods, devices, computer equipment and media

Abstract

This application provides a model training method, a text push method, an apparatus, a computer device, and a medium, belonging to the field of artificial intelligence technology. The method includes: acquiring sample data, including target text content, the target text category to which the target text content belongs, and the historical click-through rate of multimedia resources with the target text content as the title; processing the target text content based on a target model to obtain a first result and a second result, where the first result represents the probability that the target text content belongs to each text category, and the second result represents the predicted click-through rate of the target text content; and training the target model based on the first result and the second result, using the target text category and historical click-through rate as supervision information. The above technical solution enables the trained model to more accurately determine the text category to which the text content belongs and the corresponding click-through rate.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a model training method, a text push method, an apparatus, a computer device, and a medium. Background Technology

[0002] With the rapid development of internet technology, users can now access various multimedia resources of interest online, such as documents and videos. Taking videos as an example, users typically determine their interest in a video based on its title before viewing it. Therefore, how to push multimedia resources with more appealing titles to users is a problem that needs to be solved.

[0003] For content to be pushed, the CTR (Click-Through-Rate) of the multimedia resource is usually estimated based on the historical exposure and click statistics of the multimedia resources that have been pushed.

[0004] The above approach uses a single set of estimated parameters for content that has never been pushed before, which makes it impossible to accurately predict the CTR of that content. Summary of the Invention

[0005] This application provides a model training method, a text push method, an apparatus, a computer device, and a medium, enabling the trained target model to more accurately determine the text category of the text content and the click-through rate of the text content within that text category. The technical solution is as follows:

[0006] On the one hand, a model training method is provided, the method comprising:

[0007] Obtain sample data, which includes target text content, the target text category to which the target text content belongs, and the historical click pass rate of multimedia resources with the target text content as the title;

[0008] The target text content is processed based on the target model to obtain a first result and a second result. The first result is used to represent the probability that the target text content belongs to each text category, and the second result is used to represent the predicted click pass rate of the target text content.

[0009] The target model is trained based on the first result and the second result, using the target text category and the historical click pass rate as supervision information.

[0010] On the other hand, a text push method is provided, the method comprising:

[0011] Get the text to be processed;

[0012] The text to be processed is processed based on a text classification model to obtain the recommended text category and the target click pass rate of the text to be processed.

[0013] In response to the target click pass rate being within a preset click pass rate range, the text to be processed is pushed according to the recommended text category.

[0014] On the other hand, a model training apparatus is provided, the apparatus comprising:

[0015] The sample acquisition module is used to acquire sample data, which includes target text content, the target text category to which the target text content belongs, and the historical click pass rate of multimedia resources with the target text content as the title.

[0016] The result acquisition module is used to process the target text content based on the target model and obtain a first result and a second result. The first result is used to represent the probability that the target text content belongs to each text category, and the second result is used to represent the predicted click pass rate of the target text content.

[0017] The training module is used to train the target model based on the first result and the second result, using the target text category and the historical click pass rate as supervision information.

[0018] In one optional implementation, the result acquisition module is used to encode the target text content based on the first network in the target model to obtain a target text vector; classify the target text vector based on the second network in the target model to obtain the first result; and predict the click pass rate corresponding to the target text vector based on the third network in the target model to obtain the second result.

[0019] In one alternative implementation, the training module includes:

[0020] The first loss determination unit is used to determine the classification loss based on the first result and the target text category;

[0021] The second loss determination unit is used to determine the predicted loss based on the second result and the historical click pass rate;

[0022] A training unit is used to train the target model based on the classification loss and the prediction loss.

[0023] In one optional implementation, the training unit is configured to: determine the target loss based on the classification loss, the first hyperparameter corresponding to the classification loss, the prediction loss, and the second hyperparameter corresponding to the prediction loss;

[0024] The target model is trained based on the target loss.

[0025] On the other hand, a text push device is provided, the device comprising:

[0026] The text acquisition module is used to acquire the text to be processed.

[0027] The text processing module is used to process the text to be processed based on a text classification model, and to obtain the recommended text category and the target click pass rate of the text to be processed.

[0028] The text push module is used to push the text to be processed according to the recommended text category in response to the target click pass rate being within a preset click pass rate range.

[0029] On the other hand, a computer device is provided, the computer device including a processor and a memory, the memory being used to store at least one computer program, the at least one computer program being loaded and executed by the processor to implement the operations in the model training method of the present application embodiments, or to implement the operations in the text push method of the present application embodiments.

[0030] On the other hand, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer-readable storage medium, the at least one computer program being loaded and executed by a processor to implement the operations in the model training method as described in the embodiments of this application, or to implement the operations in the text push method as described in the embodiments of this application.

[0031] On the other hand, a computer program product or computer program is provided, which includes computer program code stored in a computer-readable storage medium. A processor of a computer device reads the computer program code from the computer-readable storage medium and executes the computer program code, causing the computer device to perform the model training method or text push method provided in the various optional implementations described above.

[0032] The beneficial effects of the technical solutions provided in this application are:

[0033] This application provides a model training method that improves the semantic understanding ability of the target model by obtaining the probability of the target text content belonging to each text category and the predicted click pass rate of the text content. It effectively establishes the implicit relationship between text content and text category, enabling the trained model to more accurately determine the text category to which the text content belongs and the corresponding click pass rate. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a schematic diagram of the implementation environment of a model training method provided in an embodiment of this application;

[0036] Figure 2 This is a flowchart of a model training method provided according to an embodiment of this application;

[0037] Figure 3 This is a flowchart of another model training method provided according to an embodiment of this application;

[0038] Figure 4 This is a flowchart illustrating another model training method provided according to an embodiment of this application;

[0039] Figure 5 This is a flowchart of a text push method provided according to an embodiment of this application;

[0040] Figure 6 This is a block diagram of a model training apparatus according to an embodiment of this application;

[0041] Figure 7 This is a device block diagram of a text push device according to an embodiment of this application;

[0042] Figure 8 This is a structural block diagram of a terminal provided according to an embodiment of this application;

[0043] Figure 9 This is a schematic diagram of the structure of a server according to an embodiment of this application. Detailed Implementation

[0044] 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.

[0045] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0046] In video search and recommendation, video titles are a key component of video content. Parsing the title text using natural language processing and other fundamental algorithms to enhance the understanding of semantic information is a core function of any video search system. Currently, title attractiveness can be estimated based on its semantics, which is a method of predicting click-through rate. However, the same video title can have different click-through rates in different text categories. For example, a video title including a game streamer's name will obviously have different click-through rates in the text category of "games" versus "TV series." Therefore, understanding the semantics of video titles is one of the main challenges and key points in the above approach. Failure to accurately understand the semantics of video titles will result in inconsistent performance of the same words across different topic categories.

[0047] The following is a brief introduction to the technologies that may be used in the embodiments of this application.

[0048] 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 learn-by-doing.

[0049] Big data refers to data sets that cannot be captured, managed, and processed within a certain timeframe using conventional software tools. It represents massive, rapidly growing, and diverse information assets that require new processing models to achieve stronger decision-making, insightful discovery, and process optimization capabilities. With the advent of the cloud era, big data has attracted increasing attention. Big data requires specialized technologies to effectively process large amounts of data within a tolerable timeframe. Technologies suitable for big data include massively parallel processing databases, data mining, distributed file systems, distributed databases, cloud computing platforms, the Internet, and scalable storage systems. This application's embodiments are based on the ability of big data to obtain user characteristics and content characteristics of the content to be recommended.

[0050] CTR (Click-Through-Rate) is a commonly used term in internet advertising. It refers to the click-through rate of online ads (image ads / text ads / keyword ads / ranking ads / video ads, etc.), which is the actual number of clicks on the ad (strictly speaking, the number of times the ad reaches the target page) divided by the number of ad impressions.

[0051] The model training method provided in this application can be applied to a computer device. Optionally, the computer device is a terminal or a server. The following section first uses a server as an example to describe the implementation environment of the model training method provided in this application. Figure 1 This is a schematic diagram illustrating the implementation environment of a model training method provided in an embodiment of this application. See also... Figure 1 The implementation environment includes terminal 101 and server 102.

[0052] Terminal 101 and server 102 can be connected directly or indirectly via wired or wireless communication, and this application does not impose any restrictions on this.

[0053] Optionally, the terminal 101 may be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these.

[0054] Optionally, server 102 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Server 102 is used for model training. Optionally, server 102 undertakes the primary task, and terminal 101 undertakes the secondary task; or, server 102 undertakes the secondary task, and terminal 101 undertakes the primary task; or, server 102 and terminal 101 collaborate on computation using a distributed computing architecture.

[0055] In this implementation environment, the model training method trains a target model for text classification and click-through rate prediction.

[0056] For example, the server retrieves historical push notifications and video title texts as sample texts. Each sample text is labeled with at least one of the following: its text category and click-through rate, serving as supervisory data. Each sample text is then input into the target model for training, resulting in the trained target model.

[0057] Optionally, server 102 comprises an access server, a model training server, and a database server. The access server provides access services to terminal 101. The model training server provides model training services. There can be one or more model training servers. When there are multiple model training servers, at least two model training servers provide different services, and / or at least two model training servers provide the same service, such as providing the same service in a load-balanced manner; this embodiment does not limit this. Optionally, the model training server is equipped with a model training model, thereby enabling model training based on this model training model.

[0058] Terminal 101 can refer to one of multiple terminals; this embodiment uses terminal 101 as an example. Those skilled in the art will understand that the number of terminals can be more or less. For example, there may be only one terminal, or there may be dozens or hundreds, or even more. This application does not limit the number of terminals or the type of device.

[0059] Figure 2 This is a flowchart of a model training method provided according to an embodiment of this application, such as... Figure 2 As shown, this embodiment of the application uses the model training method applied to a computer device as an example for illustration. The method includes the following steps:

[0060] 201. Obtain sample data, which includes the target text content, the target text category to which the target text content belongs, and the historical click pass rate of multimedia resources with the target text content as the title.

[0061] In this application embodiment, the target content is the title of multimedia data such as text or video that has been pushed to the user. Based on this target content, the user can understand the main content of the corresponding text or video. Whether the target content is attractive to the user directly affects whether the user will watch the text or video. The target text category includes one of the following: games, square dancing, digital products, movies and TV dramas, and comedy. The historical click-through rate is the click-through rate of multimedia resources with this target text content as the title within a period of time after being pushed and before the current time.

[0062] 202. Based on the target model, process the target text content to obtain a first result and a second result. The first result is used to represent the probability that the target text content belongs to each text category, and the second result is used to represent the predicted click pass rate of the target text content.

[0063] In this embodiment of the application, the computer device can classify the target text content and predict the click pass rate based on the target model, thereby obtaining a first result and a second result.

[0064] 203. Using the target text category and historical click pass rate as supervision information, the target model is trained based on the first result and the second result.

[0065] In this embodiment of the application, the computer device can use the target text category and historical click pass rate as supervision information, and adjust the parameters in the target model based on the difference between the first result and the target text category, and the difference between the second result and the historical click pass rate, until the target model reaches the training termination condition. This embodiment of the application does not limit this.

[0066] This application provides a model training method that improves the semantic understanding ability of the target model by obtaining the probability of the target text content belonging to each text category and the predicted click pass rate of the text content. It effectively establishes the implicit relationship between text content and text category, enabling the trained model to more accurately determine the text category to which the text content belongs and the corresponding click pass rate.

[0067] Figure 3 This is a flowchart of another model training method provided according to an embodiment of this application, such as... Figure 3 As shown, this embodiment of the application uses an example of one iteration of the model training method when applied to a computer device. The method includes the following steps:

[0068] 301. Obtain sample data, which includes the target text content, the target text category to which the target text content belongs, and the historical click pass rate of multimedia resources with the target text content as the title.

[0069] In this embodiment of the application, the computer device can obtain multimedia data such as text or video that has been pushed to the user, and then obtain at least one of the following: the title corresponding to the multimedia data, the text category to which the title belongs, and the historical click pass rate of the multimedia data, thereby obtaining a sample data.

[0070] For example, see Table 1, which exemplarily shows four sample data points obtained based on pushed videos. The titles of these four sample data points belong to the three text categories of games, square dancing, and digital products, respectively.

[0071] Table 1

[0072]

[0073] 302. Based on the first network in the target model, the target text content is encoded to obtain the target text vector.

[0074] In this embodiment, the target model is the model corresponding to the current iteration process, and the first network is the encoding network in the target model. The computer device can encode the target text content for any sample based on the encoding network in the target model. Optionally, the encoding network can be any of CNN (Convolutional Neural Networks), LSTM (Long Short-Term Memory), or LSTM+Attention. This embodiment does not limit the encoding method.

[0075] In an alternative implementation, the computer device can also use LSTM+Attention as the sentence encoding network and DNN (Deep Neural Networks) as the word encoding network.

[0076] 303. Based on the second network in the target model, classify the target text vector to obtain a first result, which is used to represent the probability that the target text content belongs to each text category.

[0077] In this embodiment, the computer device first inputs the target text vector into a fully connected layer in the second network, mapping the target text vector to a classification vector. The number of dimensions of the classification vector is equal to the total number of text categories. Then, the computer device processes the classification vector based on the softmax function in the first network to obtain the probability that the target text content belongs to each text category.

[0078] 304. Based on the third network in the target model, the click pass rate corresponding to the target text vector is predicted to obtain a second result, which is used to represent the predicted click pass rate of the target text content.

[0079] In this embodiment, the computer device first inputs the target text vector into a fully connected layer in the third network, mapping the target text vector to a prediction vector, wherein the prediction vector has a dimension of 1. Then, the computer device processes the prediction vector based on the sigmoid function in the third network to obtain the predicted click-through rate of the target text content.

[0080] 305. Using the target text category and historical click pass rate as supervision information, the target model is trained based on the first result and the second result.

[0081] In this embodiment of the application, the computer device can train the model in a supervised learning manner, using the target text category and historical click pass rate as supervision information, and train the target model based on the difference between the first result and the target text category, and the difference between the second result and the historical click pass rate.

[0082] In one optional implementation, the computer device determines a classification loss based on a first result and the target text category; determines a prediction loss based on a second result and the historical click pass rate; and then trains the target model based on the classification loss and the prediction loss. The computer device can use any one of the following loss functions to obtain the classification loss: negative log-likelihood loss function, cross-entropy loss function, and exponential loss function. Similarly, the computer device can use any one of the following loss functions to obtain the prediction loss: mean squared error loss function, cross-entropy loss function, and exponential loss function.

[0083] In one alternative implementation, the computer device harmonizes the classification loss and prediction loss using hyperparameters, so that the classification loss and prediction loss have different degrees of influence on the target model. Accordingly, the computer device determines the target loss based on the classification loss, the first hyperparameter corresponding to the classification loss, the prediction loss, and the second hyperparameter corresponding to the prediction loss; then, the target model is trained based on the target loss.

[0084] The target loss is calculated using the following formula (1).

[0085] Loss = λ1loss cls +λ2loss regression (1);

[0086] Where Loss represents the target loss, λ1 represents the first hyperparameter, and loss cls Let λ² represent the classification loss, and λ² represent the second hyperparameter. regression This indicates the predicted loss.

[0087] It should be noted that computer devices can also use other hyperparameters to calculate target loss, and this application embodiment does not limit this.

[0088] Optionally, the computer device can determine whether the target model has reached the training termination condition based on the aforementioned target loss. If the training termination condition is reached, the computer device determines the target model as a completed training model. If the training termination condition is not reached, the computer device adjusts the model parameters of the target model based on the target loss and performs the next iteration. The training termination condition includes target model convergence, reaching the required number of training iterations, etc.

[0089] To make the model training methods described in steps 301 to 305 above clearer, see [link to relevant documentation]. Figure 4 As shown, Figure 4 This is a flowchart illustrating another model training method provided according to an embodiment of this application. For example... Figure 4 As shown, taking the target text content in the sample data as "This Lu Ban is hopeless, the economy is suppressed, it can't recover at all, here's your phone to play with!" as an example, the computer device inputs this target text content into the first network to obtain the target text vector. Then, it processes the data using the softmax function in the second network to obtain the classification loss. cls And the prediction loss is obtained by processing the sigmoid function in the third network. regression Then, based on the first hyperparameter λ1, the classification loss is calculated. cls The second hyperparameter λ2 and the prediction loss regression The target loss is determined by this method. Finally, the target model is trained based on this target loss.

[0090] This application provides a model training method that improves the semantic understanding ability of the target model by obtaining the probability of the target text content belonging to each text category and the predicted click-through rate of the text content. It effectively establishes the implicit relationship between text content and text category, enabling the trained model to more accurately determine the text category to which the text content belongs and the corresponding click-through rate. Furthermore, by jointly training the second and third networks, sample text can be used to alleviate the problem of insufficient samples in the second network.

[0091] Figure 5 This is a flowchart of a text push method provided according to an embodiment of this application, such as... Figure 5 As shown, this application embodiment uses the example of a text push method applicable to a computer device for illustration. The method includes the following steps:

[0092] 501. Obtain the text to be processed.

[0093] In this embodiment of the application, for any content to be pushed, such as multimedia content such as videos, news, and novels to be pushed, the title text of the multimedia content is obtained as the text to be processed.

[0094] 502. Based on the text classification model, process the text to be processed to obtain the recommended text category and the target click pass rate of the text to be processed.

[0095] In this embodiment, the computer device can input the text to be processed into a text classification model, which is a model trained based on the aforementioned model training method. Then, the computer device obtains the recommended text category and target click-through rate output by the text classification model.

[0096] 503. In response to the target click-through rate being within the preset click-through rate range, the text to be processed is pushed according to the recommended text category.

[0097] In this embodiment of the application, after the computer device obtains the target click pass rate, it determines whether the target click pass rate is within a preset click pass rate range. If the target click pass rate is not within the preset click pass rate range, it means that the text to be processed cannot attract users; if the target click pass rate is within the preset click pass rate range, it means that the text to be processed can attract users. At this time, the computer device pushes the text to be processed according to the recommended text category so that the multimedia content corresponding to the text to be processed can obtain a higher click pass rate.

[0098] Optionally, the text classification model can also output the probability of each text category to which the text to be processed belongs and the click pass rate corresponding to each text category, so that the technicians responsible for multimedia resource push can determine the recommended text category of the text to be processed based on the probability of each text category and the click pass rate corresponding to each text category.

[0099] The text push method provided in this application embodiment can classify the title text of the multimedia content to be pushed, determine the recommended text category of the title text, and determine the possible click-through rate when the title text is recommended under the recommended text category, thereby obtaining a higher click-through rate for the multimedia content to be pushed.

[0100] Figure 6 This is a block diagram of a model training apparatus according to an embodiment of this application. The apparatus is used to perform the steps in the above-described model training method. (See also...) Figure 6 The device includes: a sample acquisition module 601, a result acquisition module 602, and a training module 603.

[0101] The sample acquisition module 601 is used to acquire sample data, which includes target text content, the target text category to which the target text content belongs, and the historical click pass rate of multimedia resources with the target text content as the title;

[0102] The result acquisition module 602 is used to process the target text content based on the target model and obtain a first result and a second result. The first result is used to represent the probability that the target text content belongs to each text category, and the second result is used to represent the predicted click pass rate of the target text content.

[0103] Training module 603 is used to train the target model based on the first result and the second result, with the target text category and the historical click pass rate as supervision information.

[0104] In one alternative implementation, the result acquisition module 602 is used to encode the target text content based on the first network in the target model to obtain a target text vector; classify the target text vector based on the second network in the target model to obtain the first result; and predict the click pass rate corresponding to the target text vector based on the third network in the target model to obtain the second result.

[0105] In one alternative implementation, the training module 603 includes:

[0106] The first loss determination unit is used to determine the classification loss based on the first result and the target text category;

[0107] The second loss determination unit is used to determine the predicted loss based on the second result and the historical click pass rate;

[0108] The training unit is used to train the target model based on the classification loss and the prediction loss.

[0109] In one alternative implementation, the training unit is used to: determine the target loss based on the classification loss, the first hyperparameter corresponding to the classification loss, the prediction loss, and the second hyperparameter corresponding to the prediction loss;

[0110] Based on this target loss, the target model is trained.

[0111] This application provides a model training device that improves the semantic understanding ability of the target model by acquiring the probability of the target text content belonging to each text category and the predicted click-through rate of the text content. It effectively establishes the implicit relationship between text content and text category, enabling the trained model to more accurately determine the text category to which the text content belongs and the corresponding click-through rate. Furthermore, by jointly training the second and third networks, sample text can be used to alleviate the problem of insufficient samples in the second network.

[0112] It should be noted that the model training device provided in the above embodiments is only illustrated by the division of the above functional modules during model training. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the model training device and the model training method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0113] Figure 7 This is a block diagram of a text push device according to an embodiment of this application. The device is used to perform the steps in the above-described text push method. (See also...) Figure 7 The device includes: a text acquisition module 701, a text processing module 702, and a text push module 703.

[0114] The text acquisition module 701 is used to acquire the text to be processed.

[0115] The text processing module 702 is used to process the text to be processed based on the text classification model, and to obtain the recommended text category and the target click pass rate of the text to be processed.

[0116] The text push module 703 is used to push the text to be processed according to the recommended text category in response to the target click pass rate being within the preset click pass rate range.

[0117] The text push device provided in this application embodiment can classify the title text of the multimedia content to be pushed, determine the recommended text category of the title text, and also determine the possible click pass rate when the title text is recommended under the recommended text category, thereby obtaining a higher click pass rate for the multimedia content to be pushed.

[0118] It should be noted that the text push device provided in the above embodiments is only illustrated by the division of the above functional modules when performing text classification. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the text push device and the text push method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0119] Figure 8This is a structural block diagram of a terminal 800 according to an embodiment of this application. The terminal 800 can be a portable mobile terminal, such as a smartphone, tablet computer, MP3 player (Moving Picture Experts Group Audio Layer III), MP4 player (Moving Picture Experts Group Audio Layer IV), laptop computer, or desktop computer. The terminal 800 may also be referred to as a user device, portable terminal, laptop terminal, desktop terminal, or other names.

[0120] Typically, terminal 800 includes a processor 801 and a memory 802.

[0121] Processor 801 may include one or more processing cores, such as a quad-core processor or an octa-core processor. Processor 801 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). Processor 801 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 801 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, processor 801 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0122] The memory 802 may include one or more computer-readable storage media, which may be non-transitory. The memory 802 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments, the non-transitory computer-readable storage media in the memory 802 are used to store at least one computer program, which is executed by the processor 801 to implement the model training method or text push method provided in the method embodiments of this application.

[0123] In some embodiments, the terminal 800 may also optionally include a peripheral device interface 803 and at least one peripheral device. The processor 801, memory 802, and peripheral device interface 803 can be connected via a bus or signal line. Each peripheral device can be connected to the peripheral device interface 803 via a bus, signal line, or circuit board. Specifically, the peripheral device includes at least one of the following: a radio frequency circuit 804, a display screen 805, a camera assembly 806, an audio circuit 807, a positioning assembly 808, and a power supply 809.

[0124] Peripheral device interface 803 can be used to connect at least one I / O (Input / Output) related peripheral device to processor 801 and memory 802. In some embodiments, processor 801, memory 802 and peripheral device interface 803 are integrated on the same chip or circuit board; in some other embodiments, any one or two of processor 801, memory 802 and peripheral device interface 803 can be implemented on separate chips or circuit boards, which is not limited in this embodiment.

[0125] The radio frequency (RF) circuit 804 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The RF circuit 804 communicates with communication networks and other communication devices via electromagnetic signals. The RF circuit 804 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals back into electrical signals. Optionally, the RF circuit 804 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, etc. The RF circuit 804 can communicate with other terminals through at least one wireless communication protocol. This wireless communication protocol includes, but is not limited to: the World Wide Web, metropolitan area networks, intranets, various generations of mobile communication networks (2G, 3G, 4G, and 5G), wireless local area networks, and / or WiFi (Wireless Fidelity) networks. In some embodiments, the RF circuit 804 may also include circuitry related to NFC (Near Field Communication), which is not limited in this application.

[0126] Display screen 805 is used to display a UI (User Interface). This UI may include graphics, text, icons, videos, and any combination thereof. When display screen 805 is a touch display screen, it also has the ability to collect touch signals on or above its surface. These touch signals can be input as control signals to processor 801 for processing. In this case, display screen 805 can also be used to provide virtual buttons and / or a virtual keyboard, also known as soft buttons and / or a soft keyboard. In some embodiments, there may be one display screen 805, disposed on the front panel of terminal 800; in other embodiments, there may be at least two display screens, disposed on different surfaces of terminal 800 or in a folded design; in other embodiments, display screen 805 may be a flexible display screen, disposed on a curved or folded surface of terminal 800. Furthermore, display screen 805 may be configured as a non-rectangular irregular shape, i.e., a non-rectangular screen. Display screen 805 may be made of materials such as LCD (Liquid Crystal Display) or OLED (Organic Light-Emitting Diode).

[0127] The camera assembly 806 is used to acquire images or videos. Optionally, the camera assembly 806 includes a front-facing camera and a rear-facing camera. Typically, the front-facing camera is located on the front panel of the terminal, and the rear-facing camera is located on the back of the terminal. In some embodiments, there are at least two rear-facing cameras, which are any one of a main camera, a depth-sensing camera, a wide-angle camera, and a telephoto camera, to achieve background blurring by fusion of the main camera and the depth-sensing camera, panoramic shooting by fusion of the main camera and the wide-angle camera, VR (Virtual Reality) shooting, or other fusion shooting functions. In some embodiments, the camera assembly 806 may also include a flash. The flash can be a single-color temperature flash or a dual-color temperature flash. A dual-color temperature flash refers to a combination of a warm-light flash and a cool-light flash, which can be used for light compensation at different color temperatures.

[0128] The audio circuit 807 may include a microphone and a speaker. The microphone is used to collect sound waves from the user and the environment, converting the sound waves into electrical signals that are input to the processor 801 for processing, or input to the radio frequency circuit 804 to achieve voice communication. For stereo sound acquisition or noise reduction purposes, multiple microphones may be used, each located at a different part of the terminal 800. The microphone may also be an array microphone or an omnidirectional microphone. The speaker is used to convert the electrical signals from the processor 801 or the radio frequency circuit 804 into sound waves. The speaker may be a conventional diaphragm speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, it can convert electrical signals not only into audible sound waves but also into inaudible sound waves for purposes such as distance measurement. In some embodiments, the audio circuit 807 may also include a headphone jack.

[0129] The positioning component 808 is used to determine the current geographical location of the terminal 800 in order to enable navigation or LBS (Location Based Service). The positioning component 808 can be a positioning component based on the US GPS (Global Positioning System), China's BeiDou system, or Russia's Galileo system.

[0130] Power supply 809 is used to supply power to the various components in terminal 800. Power supply 809 can be AC ​​power, DC power, a disposable battery, or a rechargeable battery. When power supply 809 includes a rechargeable battery, the rechargeable battery can be a wired rechargeable battery or a wireless rechargeable battery. A wired rechargeable battery is a battery that is charged via a wired line, and a wireless rechargeable battery is a battery that is charged via a wireless coil. The rechargeable battery can also be used to support fast charging technology.

[0131] In some embodiments, the terminal 800 further includes one or more sensors 810. The one or more sensors 810 include, but are not limited to: an accelerometer 811, a gyroscope 812, a pressure sensor 813, a fingerprint sensor 814, an optical sensor 815, and a proximity sensor 816.

[0132] Accelerometer 811 can detect the magnitude of acceleration on the three coordinate axes of a coordinate system established by terminal 800. For example, accelerometer 811 can be used to detect the components of gravitational acceleration on the three coordinate axes. Processor 801 can control display screen 805 to display the user interface in either a landscape or portrait view based on the gravitational acceleration signal acquired by accelerometer 811. Accelerometer 811 can also be used for games or for acquiring user motion data.

[0133] The gyroscope sensor 812 can detect the orientation and rotation angle of the terminal 800. The gyroscope sensor 812, in conjunction with the accelerometer sensor 811, can collect 3D motion data from the user on the terminal 800. Based on the data collected by the gyroscope sensor 812, the processor 801 can perform the following functions: motion sensing (e.g., changing the UI based on the user's tilt), image stabilization during shooting, game control, and inertial navigation.

[0134] The pressure sensor 813 can be disposed on the side bezel of the terminal 800 and / or on the lower layer of the display screen 805. When the pressure sensor 813 is disposed on the side bezel of the terminal 800, it can detect the user's grip signal on the terminal 800, and the processor 801 can perform left / right hand recognition or quick operation based on the grip signal collected by the pressure sensor 813. When the pressure sensor 813 is disposed on the lower layer of the display screen 805, the processor 801 can control the operable controls on the UI interface based on the user's pressure operation on the display screen 805. The operable controls include at least one of button controls, scroll bar controls, icon controls, and menu controls.

[0135] The fingerprint sensor 814 is used to collect the user's fingerprint. The processor 801 identifies the user's identity based on the fingerprint collected by the fingerprint sensor 814, or the fingerprint sensor 814 identifies the user's identity based on the collected fingerprint. When the user's identity is identified as trusted, the processor 801 authorizes the user to perform relevant sensitive operations, including unlocking the screen, viewing encrypted information, downloading software, making payments, and changing settings. The fingerprint sensor 814 can be located on the front, back, or side of the terminal 800. When the terminal 800 has a physical button or manufacturer logo, the fingerprint sensor 814 can be integrated with the physical button or manufacturer logo.

[0136] An optical sensor 815 is used to collect ambient light intensity. In one embodiment, the processor 801 can control the display brightness of the display screen 805 based on the ambient light intensity collected by the optical sensor 815. Specifically, when the ambient light intensity is high, the display brightness of the display screen 805 is increased; when the ambient light intensity is low, the display brightness of the display screen 805 is decreased. In another embodiment, the processor 801 can also dynamically adjust the shooting parameters of the camera assembly 806 based on the ambient light intensity collected by the optical sensor 815.

[0137] The proximity sensor 816, also known as a distance sensor, is typically located on the front panel of the terminal 800. The proximity sensor 816 is used to detect the distance between the user and the front of the terminal 800. In one embodiment, when the proximity sensor 816 detects that the distance between the user and the front of the terminal 800 is gradually decreasing, the processor 801 controls the display screen 805 to switch from a screen-on state to a screen-off state; when the proximity sensor 816 detects that the distance between the user and the front of the terminal 800 is gradually increasing, the processor 801 controls the display screen 805 to switch from a screen-off state to a screen-on state.

[0138] Those skilled in the art will understand that Figure 8 The structure shown does not constitute a limitation on terminal 800 and may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0139] Figure 9 This is a schematic diagram of a server structure according to an embodiment of this application. The server 900 can vary considerably due to different configurations or performance. It may include one or more Central Processing Units (CPUs) 901 and one or more memories 902. The memory 902 stores at least one computer program, which is loaded and executed by the processor 901 to implement the model training method or text push method provided in the above-described method embodiments. Of course, the server may also have wired or wireless network interfaces, a keyboard, and input / output interfaces for input and output. The server may also include other components for implementing device functions, which will not be elaborated here.

[0140] This application also provides a computer-readable storage medium applied to a computer device. The computer-readable storage medium stores at least one computer program, which is loaded and executed by a processor to implement the operations performed by the computer device in the model training method or text push method of the above embodiments.

[0141] This application also provides a computer program product or computer program, which includes computer program code stored in a computer-readable storage medium. The terminal's processor reads the computer program code from the computer-readable storage medium and executes the computer program code, causing the computer device to perform the model training method or text push method provided in the various optional implementations described above.

[0142] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0143] The above description is merely an optional embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A model training method, characterized in that, The method includes: Obtain sample data, which includes target text content, the target text category to which the target text content belongs, and the historical click pass rate of multimedia resources with the target text content as the title; The target text content is processed based on the target model to obtain a first result and a second result. The first result is used to represent the probability that the target text content belongs to each text category, and the second result is used to represent the predicted click pass rate of the target text content. Based on the first result and the target text category, a classification loss is determined; based on the second result and the historical click pass rate, a prediction loss is determined; based on the classification loss, the first hyperparameter corresponding to the classification loss, the prediction loss, and the second hyperparameter corresponding to the prediction loss, a target loss is determined; based on the target loss, the target model is trained.

2. The method according to claim 1, characterized in that, The target text content is processed based on the target model to obtain a first result and a second result, including: Based on the first network in the target model, the target text content is encoded to obtain the target text vector; Based on the second network in the target model, the target text vector is classified to obtain the first result; Based on the third network in the target model, the click pass rate corresponding to the target text vector is predicted to obtain the second result.

3. A text push method, characterized in that, The method includes: Get the text to be processed; The text to be processed is processed based on a text classification model to obtain the recommended text category and the target click pass rate of the text to be processed. The text classification model is a target model trained based on any one of claims 1-2. In response to the target click pass rate being within a preset click pass rate range, the text to be processed is pushed according to the recommended text category.

4. A model training device, characterized in that, The device includes: The sample acquisition module is used to acquire sample data, which includes target text content, the target text category to which the target text content belongs, and the historical click pass rate of multimedia resources with the target text content as the title. The result acquisition module is used to process the target text content based on the target model and obtain a first result and a second result. The first result is used to represent the probability that the target text content belongs to each text category, and the second result is used to represent the predicted click pass rate of the target text content. The training module includes: a first loss determination unit, configured to determine a classification loss based on the first result and the target text category; a second loss determination unit, configured to determine a prediction loss based on the second result and the historical click pass rate; a training unit, configured to determine a target loss based on the classification loss, a first hyperparameter corresponding to the classification loss, the prediction loss, and a second hyperparameter corresponding to the prediction loss; and to train the target model based on the target loss.

5. The apparatus according to claim 4, characterized in that, The result acquisition module is used to encode the target text content based on the first network in the target model to obtain a target text vector; and to classify the target text vector based on the second network in the target model to obtain the first result. Based on the third network in the target model, the click pass rate corresponding to the target text vector is predicted to obtain the second result.

6. A text push device, characterized in that, The device includes: The text acquisition module is used to acquire the text to be processed. The text processing module is used to process the text to be processed based on a text classification model, and to obtain the recommended text category and the target click pass rate of the text to be processed. The text classification model is a target model trained based on any one of claims 1-2. The text push module is used to push the text to be processed according to the recommended text category in response to the target click pass rate being within a preset click pass rate range.

7. A computer device, characterized in that, The computer device includes a processor and a memory, the memory being used to store at least one computer program, the at least one computer program being loaded by the processor and executed as the model training method according to any one of claims 1 to 2, or as the text push method according to claim 3.

8. A storage medium, characterized in that, The storage medium is used to store at least one computer program, which is used to execute the model training method according to any one of claims 1 to 2, or to execute the text push method according to claim 3.

Images