Information display method, apparatus, device, medium, and program product

Abstract

An information display method, device, equipment, medium and program product, relating to the technical field of computer processing. It includes: obtaining first information and second information, the first information reflects the first interaction feature and the first device feature, and the second information reflects the first content feature of the candidate media content; based on the analysis and processing of the first information and the second information by the first processing module in the first machine learning model, the first feature information is obtained; based on the processing of the first feature information by a plurality of second processing modules in the first machine learning model, the third information is obtained, each second processing module corresponds to a candidate display task, and the second processing module is used for outputting the display strategy of the candidate display task, and the candidate media content is displayed based on the third information. The problem that the personalized display requirements of the media content cannot be met is solved, the effect of intelligently determining the media content to be displayed in the intelligent device and the display strategy of the associated elements of the media content are realized.

Description

Technical Field

[0001] This invention relates to the field of computer processing technology, and in particular to an information display method, apparatus, equipment, medium, and program product. Background Technology

[0002] With the rapid development of networking, the scenarios for playing media content on smart devices are becoming increasingly widespread, making the display strategies for media content-related elements particularly important. Summary of the Invention

[0003] This invention provides an information display method, device, equipment, medium, and program product that achieves the effect of displaying media content and related elements in a way that is adaptable and personalized to user needs.

[0004] In one scenario, this paper provides an information display method, which includes: Acquire first information and second information, wherein the first information is used to reflect the first interaction characteristics and the first device characteristics of the first object, and the second information is used to reflect the first content characteristics of the candidate media content; Based on the analysis and processing of the first information and the second information by the first processing module in the first machine learning model, the first feature information is obtained; The first feature information is processed by multiple second processing modules in the first machine learning model to obtain third information. Each second processing module corresponds to a candidate display task. The second processing module is used to output the display strategy of the candidate display task. The candidate display task is related to the presentation of the candidate media content and related elements. The related elements are related to the candidate media content. In response to displaying the candidate media content, the candidate media content is displayed based on the third information.

[0005] In one instance, this document also provides an information display device, which includes: The first acquisition module is used to acquire first information and second information. The first information is used to reflect the first interaction feature and the first device feature of the first object, and the second information is used to reflect the first content feature of the candidate media content. The feature acquisition module is used to analyze and process the first information and the second information based on the first processing module in the first machine learning model to obtain first feature information; The second acquisition module is used to process the first feature information based on multiple second processing modules in the first machine learning model to obtain third information. Each second processing module corresponds to a candidate display task. The second processing module is used to output the display strategy of the candidate display task. The candidate display task is related to the presentation of the candidate media content and related elements. The related elements are related to the candidate media content. The information display module is used to display the candidate media content based on the third information in response to the display of the candidate media content.

[0006] In one instance, this document also provides an electronic device comprising: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the information display method as described herein.

[0007] In one instance, this document also provides a storage medium containing computer-executable instructions that, when executed by a computer processor, are used to perform the information display method as described herein.

[0008] In another scenario, this document also provides a computer program product, including a computer program that, when executed by a processor, implements the information display method as described in any of the present documents.

[0009] In this paper, during the process of a first object browsing media content, the interaction behavior between the first object and the media content played by the first device can be reflected based on the first interaction features, the first device features, and the first content features. This interaction behavior is incorporated as a component of the environmental state, enabling the first machine learning model to perceive real-time user behavior feedback and session temporal changes. Combined with device and content features, it infers a display strategy suitable for candidate media content. Furthermore, the first machine model evaluates the display strategies for at least one candidate display task to obtain a suitable display strategy (i.e., third information) for displaying candidate media content. Based on this third information, the operating parameters of the media playback device are adjusted so that the adjusted parameters allow the media playback device to display the media content. This solves the problem that existing technologies cannot meet the personalized display needs of media content. By combining the device state of the first device, the media content features, and the first object's interaction behavior with the media content, it achieves the effect of intelligently determining the media content to be displayed on the smart device and the display strategy suitable for the associated elements of that media content. Attached Figure Description

[0010] The above and other features, advantages, and aspects of the embodiments described herein will become more apparent when taken in conjunction with the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0011] Figure 1 This is a schematic diagram of the structure of an information display system under one scenario. Figure 2 This is a flowchart illustrating an information display method in one scenario. Figure 3 A flowchart illustrating a specific example of an information display method in another scenario; Figure 4 A flowchart illustrating an alternative example of an information display method in another scenario; Figure 5 This is a flowchart illustrating another method of information display. Figure 6 This is a schematic diagram illustrating the training process for a second machine learning model in another scenario. Figure 7 This is a schematic diagram of an information display device in one scenario. Figure 8 This is a schematic diagram of the structure of an electronic device used in one scenario. Detailed Implementation

[0012] The embodiments will now be described in more detail with reference to the accompanying drawings. While some embodiments are shown in the drawings, it should be understood that the technical solutions can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the technical solutions herein. It should be understood that the illustrated drawings and embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the technical solutions.

[0013] It should be understood that the steps described in the method implementation may be performed in different orders and / or in parallel. Furthermore, the method implementation may include additional steps and / or omit the steps shown. The scope of this document is not limited in this respect.

[0014] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one situation" means "at least one situation"; the term "another situation" means "at least one additional situation"; the term "some situations" means "at least some situations". Definitions of other terms will be given in the following description.

[0015] It should be noted that the concepts of "first" and "second" mentioned are only used to distinguish different devices, modules or units, and are not used to limit the order of the functions performed by these devices, modules or units or their interdependencies.

[0016] It should be noted that the terms "one" and "more" used in this document are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0017] The names of messages or information exchanged between multiple devices in this document are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0018] It is understood that before using the technical solutions disclosed in the various embodiments of this document, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this document in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0019] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the software or hardware, such as electronic devices, applications, servers, or storage media, that perform the operations described herein, based on the prompt message.

[0020] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0021] It is understood that the above notification and user authorization process is merely illustrative and does not limit the implementation method described in this article. Other methods that comply with relevant laws and regulations may also be applied to the implementation method described in this article.

[0022] It is understood that the data involved in the technical solutions in this article (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0023] Before introducing the information display method, we can first illustrate the application scenario. This scenario can be applied to intelligently matching display strategies to media content played on smart devices, and then personalized displaying the media content based on the appropriate display strategy.

[0024] In some cases, the provided solution can be applied to Figure 1 The information display system shown may include a client 101 and a server 102. The client 101 may include, but is not limited to, browsers, applications, web applications such as Hypertext Markup Language (HTML) applications, lightweight applications (also known as mini-programs, a type of lightweight application), or cloud applications. The client 101 may be deployed on an electronic device and relies on the operation of that device or certain applications on the device to implement its functions. The electronic device may be, for example, a device with a display screen that supports information browsing, such as a smartphone, tablet, personal computer, or other client terminal. For ease of understanding, Figure 1 The client is primarily represented as a device. Other applications, such as media content publishing applications and session applications, can also be configured on the electronic device. Server 102 can be one or more servers providing various services. That is, it can be implemented as a distributed server cluster composed of multiple servers, or as a single server; furthermore, it can be a server in a distributed system, a cloud server, or an intelligent cloud computing server or intelligent cloud host deployed with machine learning models, etc.

[0025] The information display method described herein allows interaction between client 101 and server 102, such as receiving or sending messages. For example, in this paper, server 102 can receive first and second information sent by client 101 based on an information carrier. The first information reflects the first interaction characteristics and first device characteristics of the first object, and the second information reflects the first content characteristics of the candidate media content. Based on the first processing module in the first machine learning model, the first and second information are analyzed and processed to obtain first feature information. Based on multiple second processing modules in the first machine learning model, the first feature information is processed to obtain third information. Each second processing module corresponds to a candidate display task. The second processing module is used to output the display strategy of the candidate display task. The candidate display task is related to the presentation of candidate media content and associated elements. The associated elements are related to the candidate media content. In response to displaying candidate media content, the candidate media content is displayed based on the third information. The system acquires first and second information. The first information reflects the first interaction characteristics and first device characteristics of the first object, and the second information reflects the first content characteristics of the candidate media content. Based on the first processing module in the first machine learning model, the first and second information are analyzed and processed to obtain first feature information. Based on multiple second processing modules in the first machine learning model, the first feature information is processed to obtain third information. Each second processing module corresponds to a candidate display task, and the second processing module is used to output the display strategy of the candidate display task. The candidate display task is related to the presentation of candidate media content and associated elements, and the associated elements are related to the candidate media content. In response to displaying candidate media content, the candidate media content is displayed based on the third information. This information is then sent to client 101 for display on the display interface.

[0026] It should be noted that the information display method can be executed by client 101, or by client 101 and server 102, with different functional parts of the information display device deployed on client 101 and server 102 respectively; specifically, the information display module is deployed on client 101, while the first acquisition module, feature acquisition module, and second acquisition module are deployed on server 102. Client 101 and server 102 achieve data interaction and functional collaboration through network communication. It should be understood that... Figure 1 The number of clients and servers shown is for illustrative purposes only. Any number of clients and servers can be configured to meet specific implementation requirements.

[0027] In one scenario, client 101 or server 102 can implement the provided detailed information processing method by running a computer program. For example, the computer executable instructions can be microprogram-level commands, machine instructions, or software instructions. The computer program can be a native program or software module in the operating system; it can be a native application, i.e., a program that needs to be installed in the operating system to run, such as a media content processing application or an instant messaging application; or it can be a small program that can be embedded in any application, i.e., a program that only needs to be downloaded to a browser environment to run. In short, the aforementioned computer executable instructions can be any form of instruction, and the aforementioned computer program can be any form of application, module, or plugin.

[0028] Figure 2 This is a flowchart illustrating an information display method for one scenario. This method is applicable to scenarios where media content played on smart devices is intelligently matched with display strategies, and the media content is personalized based on the adapted display strategies. This information display method can be executed by an information display device, which can be implemented in software and / or hardware, optionally through a computer program. Figure 2 As shown, this information display method may specifically include: S210. Obtain first information and second information, wherein the first information is used to reflect the first interaction characteristics and the first device characteristics of the first object, and the second information is used to reflect the first content characteristics of the candidate media content.

[0029] The first object can be understood as a user browsing media content on a first device. During the browsing process, the first object can interact with the media content by triggering the content. The first device can be a smart device that supports media content playback; for example, it can be a mobile communication device or a computer client, or any device capable of network interaction. A media playback module can be deployed on the first device to display media content; for example, the media playback module can be a media content playback application. This media playback module can support displaying locally pre-stored media content, real-time loaded media content, and completed media content.

[0030] The first information can be understood as a carrier of the original information set associated with the first object. The first information reflects two different dimensions of information associated with the first object: firstly, the first interaction characteristics between the first object and the media content; and secondly, the first device characteristics of the first device associated with the first object. The first information includes, but is not limited to, the basic information of the first object, as well as the original interaction data, device operation data, and session context data of the current round during the first object's use of the first device. For example, taking the first object browsing media content as an example, the first information may include the basic information associated with the first object collected when the first object registers the media playback module. The first information may also include the first interaction characteristics between the first object and the displayed media content during the browsing process, as well as session context data during the browsing process, such as the number of times the first object browses the media content in the current session round, the number of times there is buffering or frame dropping within the session, the cumulative data consumption within the session, and the duration of the current session round. The first information may also include the basic device information of the first device, such as the device model, memory location, network type, and current battery level.

[0031] The second information is used to characterize the content features of different candidate media content. The second information includes, but is not limited to, media material tags, content semantic statistics, and content quality parameters, so that the first content features can be parsed based on the second information to complete the content dimension characterization of the candidate media content.

[0032] The first interaction feature is used to reflect the behavioral interaction between the first object and the media content. For example, the first interaction feature can be the first object's interaction features with the media content currently being played on the first device, such as playback duration, completion status, likes, pauses, skips, and playback.

[0033] The first device characteristic can be understood as the device status information of the first device, such as the device type, device model, internal storage location, and current battery level. The first device characteristic is used to reflect at least the device status and / or the environmental operating status of the first device.

[0034] Candidate media content can be understood as media content predicted based on the browsing feedback of the first object's historical media content, which meets the first object's media content preview requirements. Candidate media content can be determined based on the first object's interaction characteristics with at least one historical media content. For example, historical interaction characteristics may include the first object's behavior characteristics such as playback duration, completion status, likes, pauses, skips, and playback of historical media content. For instance, to improve the first object's user experience while browsing media content, the browsing feedback results of the first object's media content are usually evaluated based on the content style and content type of the first object's historical media content, and then candidate media content associated with the first object is determined based on the historical media content.

[0035] The first content feature can be understood as the feature items that characterize the inherent attributes, quality, and semantic features of the candidate media content itself. The first content feature includes, but is not limited to, content topic tags, duration information, image quality attributes, and content popularity. For example, the first content feature could be the content category, media content length, and component type of the candidate media content.

[0036] In this paper, when a first object displays media content based on the media playback module in a first device, the first interaction feature of the first object in the current session round for the media content is obtained (e.g., basic information of the first object, and interaction behavior data such as playback, click, pause, like, and swipe operations of the first object on the media content), and the first device feature of the first device is obtained (e.g., device model, memory location, network type, and current battery level of the first device). The first interaction feature and the first device feature are used as first information. At the same time, second information corresponding to each media content is obtained, so as to predict at least one first content feature (e.g., content category, media content length, and component type) that matches the first object based on the second information, so as to determine at least one candidate media content corresponding to the first object.

[0037] S220. Based on the first processing module in the first machine learning model, the first information and the second information are analyzed and processed to obtain the first feature information.

[0038] The first machine learning model can be understood as a deep learning model that analyzes and processes first and second information to obtain first feature information. Alternatively, the first machine learning model can also be a network model dependent on an intelligent system. This first machine learning model includes at least a first processing module, which in turn includes at least a feature fusion network and a shared backbone network. The feature fusion network fuses the first and second information and outputs a fused feature vector based on the first and second information. This fused feature vector is then used as input to the shared backbone network, enabling it to perform further deep feature extraction. The shallowly extracted fused feature vector is further extracted to obtain high-dimensional deep semantic features, thus obtaining the first feature information. This first feature information can be understood as the deep fused feature vector output by the first machine learning model.

[0039] In practical applications, the first information and the second information are input into the first machine learning model. Based on the first processing module in the first machine learning model, the first information and the second information are used to extract and fuse features to obtain a shallow-level fused feature vector. Further, the shallow-level fused feature vector is used to extract deep features to obtain a fused feature vector of deep semantic features (i.e., the first feature information).

[0040] S230. Based on the first machine learning model, multiple second processing modules process the first feature information to obtain the third information. Each second processing module corresponds to a candidate display task. The second processing module is used to output the display strategy of the candidate display task. The candidate display task is related to the presentation of candidate media content and related elements. The related elements are related to the candidate media content.

[0041] The second processing module is a post-processor of the first processing module. For example, the second processing module can be a policy-value head, which includes two parallel sub-branches: a value head and an advantage head. The value head receives the first feature information and outputs a state characterizing the overall device environment state of the first device. The advantage head receives the first feature information and outputs the action advantage value corresponding to each candidate display task, representing the additional or negative gain (i.e., action advantage value) that the corresponding candidate display task can bring above the state baseline gain.

[0042] The third piece of information includes the display strategy for each candidate display task. For media content, there are multiple associated candidate elements. Each candidate element and the loading strategy of the media content can be considered a candidate display task. This can be understood as the display method of the candidate elements associated with the media content and the loading method of the media content both constituting a candidate display task. In other words, each candidate display task corresponds to an associated element, and each associated element corresponds to a set of control parameters for the display strategy. The display strategy can be understood as executable parameters used to control the display of candidate media content by the media playback module. For example, the associated element could be the display timing of the candidate media content, the pre-stream duration, the click hotspot range, the style selection of the semi-transparent overlay at the bottom of the screen, the pre-load size control of the media content resources, the control of the number of line breaks in the text, or the screen cropping and filling method of the media content, etc. Each associated element's display strategy corresponds to a candidate display task, and correspondingly, each candidate display task corresponds to a second processing module, i.e., a processing strategy header, to determine the appropriate display strategy corresponding to the associated element based on the second processing module.

[0043] For example, taking the pre-pull streaming duration of candidate media content as an example, the display strategy of the pre-pull streaming duration (i.e. the pre-loading task of media content) is used to characterize the pre-loading duration of candidate media content. For example, the control parameters of the display strategy corresponding to the pre-pull streaming duration can be set to 2 seconds, 4 seconds or 6 seconds. For example, the pre-pull streaming duration is determined to be 4 seconds based on the first interaction feature, the first device feature and the first content feature of the candidate media content.

[0044] In practical applications, the display of candidate media content is influenced by multi-dimensional related elements, including first device features, first interaction features, and first content features. To optimize the display effect of candidate media content, display strategies with different dimensions can be set, and corresponding candidate display tasks can be assigned to each strategy. In this paper, each second processing module in the first machine learning model handles one of the candidate display tasks.

[0045] Considering the significant differences in the characteristics of different first devices, different display strategies should be adopted when different first devices are used for the first object to display media content. For example, the device types of the first device may include Type 1 and Type 2, where Type 1 devices have superior performance compared to Type 2 devices. Based on this, Type 1 devices can support higher-quality media content playback requirements, while Type 2 devices, due to their inherent performance limitations, often face more constraints in media content playback.

[0046] Taking the associated elements of candidate media content, including pre-stream duration, media content clarity, and text line breaks, as an example, the pre-stream duration reflects the pre-loading time of the candidate media content before display; the media content clarity characterizes the clarity of the candidate media content during display; and the text line breaks indicate the upper limit of the number of text lines when displaying the candidate media content. The displayed text in the candidate media content can be subtitles or titles, etc. Each associated element and the loading strategy of the media content is treated as a candidate display task, and these three candidate display tasks are input into three second processing modules. The second processing modules then evaluate the display strategy suitable for each associated element. For example, for the first device of the first type, the second processing module corresponding to the pre-stream duration determines the pre-stream duration to be 4 seconds, the second processing module corresponding to the media content clarity determines the media content clarity to be ultra-high definition, and the second processing module corresponding to the text line breaks determines the text line break function to be 5 lines. Based on this, when displaying candidate media content, the operating parameters of the media content playback module are adjusted according to the display strategy corresponding to the associated elements determined by each second processing module, and the candidate media content is then displayed. For the first device of the second type, the second processing module corresponding to the pre-stream duration determines the pre-stream duration to be 2 seconds, the second processing module corresponding to the media content clarity determines the media content clarity to be standard definition, and the second processing module corresponding to the text line break number determines the text line break function to be 5 lines. Based on this, when displaying candidate media content, the operating parameters of the media content playback module are adjusted according to the display strategy corresponding to the associated elements determined by each second processing module, and the candidate media content is then displayed.

[0047] The reason for this setup is that, due to the superior performance of the first type of device compared to the second type, forcibly applying the display strategy for candidate media content from the first type to the second type of device would result in stuttering, frame drops, or overheating when displaying candidate media content on the second type of device. Conversely, applying the display strategy for candidate content from the second type of device to the first type of device would negatively impact the user experience of the first object associated with the first type of device. Therefore, this paper, by combining the characteristics of the first device, determines the display strategy for different types of first devices when displaying media content, thus enabling flexible adaptation to the display needs of related elements of media content across different types of first devices.

[0048] Therefore, this paper will generate a corresponding candidate display task for each display strategy, and process each candidate display task based on each second processing module pre-deployed in the first machine learning model to determine the appropriate display strategy (i.e., third information) for the media content corresponding to each candidate display task, so as to display the candidate media content corresponding to each candidate display task in sequence based on the third information.

[0049] S240. In response to displaying candidate media content, display candidate media content based on third-party information.

[0050] Based on the above example, the third information is sent to the media playback module of the first device so that when displaying candidate media content, the operating parameters of the media playback module are adjusted based on the third information, and each candidate media content is displayed in turn.

[0051] In this paper, during the process of a first object browsing media content, the interaction behavior between the first object and the media content played by the first device can be reflected based on the first interaction features, the first device features, and the first content features. This interaction behavior is incorporated as a component of the environmental state, enabling the first machine learning model to perceive real-time user behavior feedback and session temporal changes. Combined with device and content features, it infers a display strategy suitable for candidate media content. Furthermore, the first machine model evaluates the display strategies for at least one candidate display task to obtain a suitable display strategy (i.e., third information) for displaying candidate media content. Based on this third information, the operating parameters of the media playback device are adjusted so that the adjusted parameters allow the media playback device to display the media content. This solves the problem that existing technologies cannot meet the personalized display needs of media content. By combining the device state of the first device, the media content features, and the first object's interaction behavior with the media content, it achieves the effect of intelligently determining the media content to be displayed on the smart device and the display strategy suitable for the associated elements of that media content.

[0052] Figure 3 This is a flowchart illustrating an information display method in one scenario. The technical solution in this scenario can be combined with implementation methods in other scenarios. For identical or related parts, descriptions of other scenarios can be used, and will not be repeated here. Figure 3 As shown, the method in this case may specifically include: S310. Obtain first information and second information, wherein the first information is used to reflect the first interaction characteristics and the first device characteristics of the first object, and the second information is used to reflect the first content characteristics of the candidate media content.

[0053] In some cases, obtaining first information and second information includes: obtaining object features and behavioral features of a first object, wherein the behavioral features are used to reflect the statistical results of interaction information within multiple candidate time periods; obtaining first device features, wherein the first device features are related to a first device and include at least one or more of device type, device memory, device network type, and device battery level; obtaining media attributes of candidate media content, thereby obtaining first content features based on the media attributes, wherein the media attributes are related to the candidate media content; using the object features and behavioral features as first information; and using the first content features as second information.

[0054] For example, taking the first object browsing media content as an example, the first information may include the basic information of the first object. The first information may also include the first interaction characteristics between the first object and the playing media content during the browsing process. Furthermore, the first information may also include session context data during the first object's browsing process, such as the number of times the first object browses media content in the current session, the number of times buffering or frame drops occurred within the session, the cumulative data consumption within the session, and the duration of the current session. The first information may also include the basic device information of the first device, such as the device model, memory location, network type, and current battery level. Simultaneously, during the first object's browsing of media content, first content features are extracted based on the historical media content browsed by the first object, and these first content features are used as second information for matching candidate media content.

[0055] S320. Based on the first processing module in the first machine learning model, the first information and the second information are analyzed and processed to obtain the first feature information.

[0056] In this paper, the first processing module includes a first feature encoding unit, a second feature encoding unit, a feature fusion unit, and a feature sharing unit. Based on the first processing module in the first machine learning model, the module analyzes and processes first information and second information to obtain first feature information, including: encoding the first information and second information based on the first feature encoding unit to obtain a first encoded feature; encoding the first information and second information based on the second feature encoding unit to obtain a second encoded feature; fusing the first encoded feature and the second encoded feature based on the feature fusion unit to obtain a fused feature; and processing the fused feature based on the feature sharing unit to obtain the first feature information.

[0057] The system comprises the following components: a first feature encoding unit for embedding and encoding sparse features of the first and second information to obtain a first encoded feature; a second feature encoding unit for embedding and encoding dense features of the first and second information to obtain a second encoded feature; a feature fusion unit for fusing the first and second encoded features to obtain a fused feature; and a feature sharing unit for further feature extraction from the fused feature to obtain first feature information.

[0058] For details, see Figure 4 Taking a sparse feature embedding layer with a first feature encoding unit and a dense feature multilayer perceptron as an example, after obtaining the first information and the second information, the first information and the second information are input into the sparse feature embedding layer for encoding processing to obtain sparse feature information (i.e., the first encoded feature). At the same time, the first information and the second information are input into the dense feature multilayer perceptron for encoding processing to obtain dense feature information (i.e., the second encoded feature). Based on this, the first encoded feature and the second encoded feature are input into the feature fusion network (i.e. the feature fusion unit) for fusion processing to obtain the fused feature.

[0059] The feature fusion network supports at least one feature fusion method, such as factorization machine, cross network or multilayer perceptron.

[0060] Taking the factorization machine class as a feature fusion network as an example, when performing low-order feature cross-interaction on the first and second encoded features, the factorization machine class can calculate the inner product of the latent vectors of the concatenated first and second encoded features, explore the pairwise low-order interaction relationship between the first and second encoded features, and perform fusion processing on the first and second encoded features to obtain the fused features.

[0061] Taking cross-networks as a feature fusion network as an example, cross-networks can explicitly construct network layers with high-order feature crosses. The first and second encoded features are concatenated as inputs, and multi-order feature combinations are generated iteratively by performing feature multiplication operations layer by layer. The first and second encoded features are then fused to obtain the fused features.

[0062] Taking a multilayer perceptron as a feature fusion network as an example, the multilayer perceptron, as a deep neural network with stacked fully connected layers, inputs the concatenated first and second encoded features into each fully connected layer of the network. It relies on activation functions to complete implicit nonlinear transformation and feature interaction, and performs fusion processing on the first and second encoded features to obtain fused features.

[0063] It is understandable that the method of obtaining fused features based on the first and second coding features in this paper is not limited. In practical applications, the fusion processing method for the first and second coding features can be switched according to actual needs.

[0064] The purpose of this setup is to use sparse feature information to characterize discrete identifier information, which has a strong sample memory capability, while dense feature information is used to represent continuous statistical information, which has a good generalization and reasoning capability. Fusing these two features can combine memory capability and generalization capability, take into account both discrete and continuous amplitude information, explore the interaction relationship between sparse and dense cross-type features, and improve the model's prediction performance for both known and unknown samples.

[0065] Based on this, continue to see Figure 4 Based on the feature sharing unit, the fused features are further extracted at a deeper level to obtain the first feature information, and the deep semantic features of the first information and the second information are represented based on the first feature information.

[0066] S330. For multiple second processing modules, feature decoupling processing is performed on the first feature information based on the second processing modules to obtain third information, which includes the display strategy corresponding to each candidate display task.

[0067] Based on the examples above, see further. Figure 4The first machine learning model may include multiple policy value heads (i.e., second processing modules), such as policy 1 value head, policy 2 value head, ..., policy N value head. Each second processing module corresponds to a candidate display task, which corresponds to the associated elements of the media content, and outputs the display strategy corresponding to the candidate display task based on the second processing module. Taking one of the second processing modules in the first machine learning model as an example, the second processing module is a dual-output head module that shares a backbone network with the first processing module. This second processing module includes two parallel sub-branches, namely the value head and the advantage head. The second processing module receives first feature information and performs semantic decoupling processing on the first feature information to obtain the expected total gain corresponding to the first feature information. Among them, the value head receives the first feature information and outputs the state baseline gain representing the overall device environment state of the first device. This state baseline gain is independent of the specific candidate display task. At the same time, the advantage head receives the first feature information and outputs the action advantage value corresponding to each candidate display task, which is used to represent the additional gain or gain loss (i.e., action advantage value) that the corresponding candidate display task can bring above the baseline gain. It should be noted that for the same candidate display task, the corresponding device environment state is consistent, corresponding to the state in the decision space, while the same candidate display task corresponds to different action combinations in the decision space. Based on this, in order to determine the appropriate display strategy for the candidate display task, the second processing module combines the first feature in the first feature information to determine the state baseline gain, and on the basis of the state baseline gain, adds the action advantage value corresponding to each display strategy, thereby obtaining the Q-value vector of the candidate display task under different display strategies (wherein, the Q-value vector is used to characterize the expected total gain of the display strategy output by the second processing module), so as to determine the independent optimal action appropriate to the candidate display task (i.e., determine the display strategy appropriate to the candidate display task) based on the expected total gain.

[0068] In some cases, the first feature information is decoupled based on the second processing module to obtain the third information, including: decoupling the first feature information into a first state feature and a first action feature based on the second processing module; processing the first state feature and the first action feature based on a first function to obtain the second feature information; obtaining the fourth information based on the second feature information, which reflects the evaluation result corresponding to each candidate strategy of the candidate display task; and obtaining the third information based on the fourth information.

[0069] The first state feature characterizes the current objective conditions of the system, and includes, but is not limited to, the first device feature, the first interaction feature, and the first content feature. The first state feature corresponds to the state in the decision space. The first action feature characterizes the specific set of parameters for the display strategy of the candidate display task. The first action state corresponds to the action in the decision space.

[0070] The second feature information is used to characterize the feature information after merging the first state feature and the first action feature. The second feature information is also used to characterize the correlation between the first state feature and the first action feature.

[0071] The fourth information can be understood as the evaluation result after the second processing module performs attribute evaluation on the second feature information. The fourth information is used to reflect the total expected gain corresponding to the second feature information, that is, the expected total gain corresponding to the display strategy of the candidate display task.

[0072] Specifically, after the first feature information is input into the second processing module, it undergoes decoupling processing to obtain the first state feature and the first action feature. The purpose is to achieve semantic separation at the feature dimension, facilitating the representation of the device state information of the first device and the action information representing the decision candidate space, respectively, with targeted encoding and constraints applied. Based on this, the first state feature and the first action feature are merged, and the first action feature and the first state feature are calculated based on the first function to obtain the second feature information. The purpose of this second feature information is to establish the state-action feature interaction relationship of the candidate display task's display strategy, so as to evaluate the display strategy of the candidate display task based on the current system objective conditions of the first device, obtaining the fourth information. The third information is then constructed based on the fourth information output by all the second processing modules.

[0073] S340. In response to displaying candidate media content, display candidate media content based on third-party information.

[0074] Based on the above example, the third information is sent to the media playback module of the first device so that when the first object displays candidate media content through the media playback module, the operating parameters of the media playback module are adjusted based on the third information, and each candidate media content is displayed in turn.

[0075] In some cases, the information display method further includes: obtaining first feedback information, which indicates feedback information for displaying candidate media content based on third information; and adjusting the model parameters of a first machine learning model based on the first feedback information, the first information, and the second information.

[0076] The first feedback information is used to characterize the various real business feedback results generated by the first object in response to the candidate media content. The first feedback information can be understood as a multi-dimensional set of observations of user-side playback behavior and playback experience after adjusting the operating parameters of the media playback module of the first device. The first feedback information includes at least the interactive behavior data of the first object in response to a single media content (e.g., the dwell time of the first object in response to the media content, playback completion rate, and active operation behavior, etc.), the overall feedback information of a single session (e.g., the duration of a single session, the number of media content played within the session, active session exit behavior, or abnormal session exit behavior, etc.), and may also include objective evaluation indicators of the media content, such as the playback clarity of the media content, the pre-pull streaming duration, the click hotspot range, and the video masking type, etc.

[0077] In practical applications, in order to further improve the model processing capability of the first machine learning model and match appropriate display strategies for candidate display tasks of media playback modules of different first devices, the first feedback information of the first object for each candidate media content is collected, and the first feedback information is used as the training data of the first machine learning model. The first feedback information, the first information and the second information are combined to construct new training samples to further train the first machine learning model and adjust the model parameters of the first machine learning model.

[0078] In this paper, the first processing module in the first machine learning model performs shallow feature extraction on the first and second information to obtain a fused feature code. Further deep feature extraction is then performed on the fused feature code based on a feature sharing unit to obtain first feature information, thereby deeply mining the deep semantic features contained in the first and second information. Based on this, multiple second processing modules deployed in the first machine learning model evaluate the display strategies associated with each candidate media content's candidate display task to determine the appropriate display strategy for each candidate display task, obtaining third information. This third information is then used to adjust the operating parameters of the media playback device so that the adjusted parameters allow the media playback device to display the media content. This solves the problem that existing technologies cannot meet the personalized display needs of media content. By combining the device status of the first device, the characteristics of the media content, and the interactive behavior of the first object with the media content, it achieves the effect of matching appropriate display strategies to media content and personalized display of media content and its associated elements.

[0079] Figure 5 This is a flowchart illustrating an information display method in one scenario. The technical solution in this scenario can be combined with implementation methods in other scenarios. For identical or related parts, descriptions of other scenarios can be used, and will not be repeated here. Figure 5 As shown, the method in this case may specifically include: S410. Obtain fifth information and sixth information. The fifth information is used to reflect the second interactive features of the second object, the second device features, and the second content features of the first media content. The sixth information is used to reflect the set display strategy corresponding to the candidate display task.

[0080] The second object can be understood as a user browsing media content on a second device during the model training phase of the first machine learning model. The second object can interact with the media content displayed on the second device during the browsing process. This second device can be a smart device that supports media content playback; for example, it can be a mobile communication device or a computer client, or any device capable of network interaction. A media playback module can be deployed in the second device. This media playback module can support displaying locally pre-stored media content, real-time loaded media content, and completed media content.

[0081] The fifth information can be understood as a carrier of the original information set associated with the second object. The fifth information reflects two different dimensions of information associated with the second object: firstly, the first interaction characteristics between the second object and the media content being displayed on the second device; and secondly, the second device characteristics associated with the second object. The fifth information includes, but is not limited to, the basic information of the second object, as well as the original interaction data, device operation data, and session context data of the second object during its use of the second device. For example, taking the second object browsing media content on the second device as an example, the fifth information may include the basic information associated with the second object collected when the second object registered the media playback module. The fifth information may also include the first interaction characteristics between the second object and the media content being displayed on the second device during media content browsing, as well as session context data during the second object's browsing process, such as the number of sessions browsed by the second object through the media playback module of the second device in the current session round, the number of stutters or frame drops occurring within the session, the cumulative traffic consumption within the session, and the duration of the current session round. The fifth information may also include the basic device information of the second device, such as the device model, memory location, network type, and current battery level.

[0082] The sixth piece of information can be understood as the pre-defined display strategy for the first media content on the second device during the model training phase of the first machine learning model. This display strategy can be understood as a set of executable parameters used to control the media playback module on the second device. For example, different display strategies are set for the media content, and corresponding candidate display tasks are generated based on each strategy. These display strategies include, but are not limited to, controlling the timing of interface components displayed when showing media content, selecting the style of the semi-transparent overlay at the bottom of the screen, selecting the method for cropping and filling the media content, controlling the pre-stream duration of the media content, controlling the pre-load size of media content resources, adjusting click hotspots, and controlling the number of line breaks in the text.

[0083] The second interaction feature is used to reflect the behavioral interaction between the second object and the media content played on the second device. For example, the second interaction feature can be behavioral features of the second object regarding the playback duration, completion status, likes, pauses, skips, and playback of the media content currently being played on the second device.

[0084] The second device characteristics can be understood as reflecting the device status information of the second device, such as the device type, device model, internal storage location, and current power level. The second device characteristics are used to reflect at least the device status and / or the environmental operating status of the second device.

[0085] The second content feature can be understood as a feature that characterizes the inherent attributes, quality, and semantic qualities of the candidate media content itself. For example, the second content feature could be the content category, media content length, and component type of the candidate media content.

[0086] For details, see Figure 6In the model training phase of the first machine learning model, the model training-inference iterative closed loop includes offline training, model / configuration deployment, real-time client decision-making, multi-strategy execution, execution result tracking and reporting, and sample construction and feedback. Specifically, in the offline training phase of the first machine learning model, fifth and sixth information are obtained to construct model training samples. On the one hand, at least one second interaction feature and second device feature of the second device during the process of the second object displaying media content are collected. On the other hand, at least one second content feature is extracted based on the media data during the process of the second object browsing media, and this second content feature is used as the basis for matching candidate media content. The constructed training samples are then processed through sample construction and feedback, and sent to the offline training step. The model is updated by relying on imitation learning cold start + offline reinforcement learning. After the updated first machine learning model and configuration are deployed, the client gathers state features to carry out model inference to achieve real-time decision-making. Then, after multi-strategy execution and hard constraint secondary verification, the execution result tracking (i.e., the first feedback information) is reported, and the process re-enters sample construction and feedback, iteratively optimizing the first machine learning model.

[0087] S420. Obtain second feedback information, which reflects the feedback result of the second object to the sixth information.

[0088] The second feedback information is used to characterize the various real business feedback results generated by the second object after receiving the sixth information. The second feedback information can be understood as a multi-dimensional set of observations of user-side playback behavior and playback experience after adjusting the operating parameters of the media playback module of the second device based on the sixth information. The second feedback information includes at least the second object's interactive behavior data for a single media content (e.g., the second object's dwell time on the media content, playback completion rate, and active operation behavior), and overall feedback information for a single session (e.g., single session duration, number of media content played within the session, active session exit behavior, or abnormal session exit behavior). In addition, it may also include objective evaluation indicators of the media content, such as the playback clarity, pre-stream duration, click hotspot range, and video masking type.

[0089] In some cases, the first machine learning model is obtained by: inputting the first media content and the fifth information into the second machine learning model, outputting the first prediction result, obtaining the first attribute based on the fifth and sixth information, the first attribute being used to characterize the adaptation attribute between the prediction display strategy and the expectation; adjusting the model parameters of the second machine learning model based on the first attribute, the first prediction result, and the second feedback information; and obtaining the first machine learning model based on the second machine learning model in response to the completion of training.

[0090] The second machine learning model can be understood as a deep learning model that analyzes and processes information based on the fifth and sixth pieces of information. Alternatively, the second machine learning model can also be a network model that relies on an intelligent system. This second machine learning model is in the training phase, and the first machine learning model can be obtained by training the second machine learning model. The first media content refers to the media content used during the model training phase.

[0091] The first prediction result is used to reflect the predicted display strategy for the candidate display task. This predicted display strategy refers to the display strategy based on the output of the second machine learning model, which reflects the display strategy for the second device to display candidate media content.

[0092] The first attribute characterizes the degree of fit between the predictive display strategy and the candidate media content displayed on the second device. A higher first attribute indicates a better display effect of the candidate media content based on the predictive display strategy on the second device, while a lower first attribute indicates a worse display effect of the candidate media content based on the predictive display strategy on the second device.

[0093] In the actual training process of the second machine learning model, the training samples are online-collected logs of (state, action, reward, next state) quadruplets, which are independent of the current training strategy and can be reused repeatedly. The second machine learning model includes at least a third feature processing module and at least one fourth processing module. The third processing module includes a third feature extraction unit, a fourth feature extraction unit, and a shared fusion unit to obtain a deep-level fused feature vector based on the fifth and sixth information after obtaining the fifth and sixth information. During the data collection phase, when obtaining the fifth and sixth information, each value of each strategy is manually assigned the same probability of occurrence (e.g., two strategies each have two values, and the four combinations have the same sample size) to ensure that the training data is unbiased.

[0094] Based on this, the prediction and display strategies for at least one candidate display task are evaluated by each of the fourth processing modules to obtain the first attribute of each prediction and display strategy (i.e., the first prediction result). Further, the first prediction result, the first attribute, and the second feedback information are used as offline training samples for the second machine learning model to train its parameters until the second machine learning model is fully trained, thus obtaining the first machine learning model.

[0095] The purpose of obtaining the second feedback information and using it as part of the offline sample data is that, in practice, some prediction and display strategies exhibit strong autonomy. This is because a series of control conditions built into the intelligent device have higher priority than the decision-making center, leading to inconsistencies between assignment and execution, such as "the initiating a strategy but not executing it, while actually executing it." This results in systematic pollution of the training data (for example, three-quarters of the sample values ​​for a certain strategy are concentrated on one side, and the model naturally favors that value). Based on this, this paper sets up a system framework for the second machine learning model in two ways: First, it increases the control priority of the decision-making center, opening up business conditions previously fixed by human experience to the decision-making model for exploration, allowing the model to make decisions based on real-time performance characteristics. Second, it implements hard-constrained dual-channel modeling. That is, after obtaining the prediction and display strategy, the media playback module of the second device needs to further verify the strategy before executing it, ensuring that the device state of the second device can support the parameter control of the media playback module by the prediction and display strategy. Simultaneously, the proportion of prediction and display strategies of the second machine learning model that are rolled back by the media playback module of the second device is continuously statistically analyzed to prevent long-term deviation between the decision space and the actual executable space.

[0096] In some cases, the method further includes: in response to a new candidate display task, configuring a third processing module, the third processing module being related to the new candidate display task, the third processing module being a module in the first machine learning model; adjusting the parameters of the first machine learning model based on the second sample; and in response to training completion, obtaining a first machine learning model, the first machine learning model including a second processing module associated with the new candidate display task, the second processing module being a module with adjusted parameters of the third processing module.

[0097] The newly added candidate display task can be understood as a display task that adds associated elements to the media content. This media content can have multiple associated candidate elements. When adding new associated candidate elements, each new candidate element, the media content loading strategy, and the media content loading method can be considered a new candidate display task. For example, the original associated elements of the candidate media content might include the pre-stream duration, the display timing of the candidate media content, and the style selection of the semi-transparent overlay at the bottom of the screen. Based on this, a new associated element corresponding to the candidate media content is added. For example, the new associated element might be the corresponding image clarity of the candidate media content. At this point, a new candidate display task corresponding to this new associated element is generated, and a second processing module corresponding to the new candidate display task is added to the first machine learning model. This second processing module is used as a second sample for model training, and the new candidate task is then displayed with the corresponding image clarity when the candidate media content is played.

[0098] The third processing module can be understood as a post-processor module added to the first machine learning model for new candidate display tasks, and connected to the first processing module. For example, the third processing module can be a strategy-value head corresponding to the new display task.

[0099] Specifically, when new associated elements exist corresponding to candidate media tasks, a third processing module can be set in the first machine learning model. To ensure the accuracy of the first machine learning model in executing each candidate display task when the third processing module is added, the model parameters of the first machine learning model need to be adjusted to optimize the model. For example, the first machine learning model with the added third processing module is trained based on the second sample until the first machine learning model converges. Then, the model parameters of the first machine learning model are adjusted to obtain an updated first machine learning model, and the added third processing module is used as the second processing module in the updated first machine learning model.

[0100] The advantage of this setup is that, to facilitate the easy integration of any new strategy (i.e., newly added candidate display tasks), the second machine learning model defines a unified strategy integration protocol. Any candidate display task's display strategy only needs to implement the following conventions to be included in the joint decision-making process: (a) Registration and preprocessing: Before the decision is triggered, the candidate display task's display strategy returns its currently executable action result range, arranged by priority, for the framework to construct a set of legal actions; (b) Dynamic feature provision: The candidate display task's display strategy optionally returns its dependent real-time dynamic features, which are input into the second machine learning model along with the state; (c) Decision result acquisition: When the strategy takes effect, the decision result (including switches and parameters) is read through a unified interface according to the strategy identifier; (d) Tracking execution result reporting: After the strategy is executed, the actual execution result is reported through unified tracking points for sample feedback and detection; (e) Post-hoc detection: Each strategy automatically generates an execution result distribution dashboard to observe the consistency between the model's tendency and the actual execution. The entire integration process does not require modification of the model backbone; only a new value head needs to be added and the above four interface conventions implemented according to the protocol are required.

[0101] S430. Based on the first media content, the fifth information, and the second feedback information, a first sample is obtained.

[0102] In this paper, during the model training phase of the second machine learning model, offline training samples are constructed by collecting the operating environment of the smart device, the content features of the media content, the interaction features of the second object with the media content, and the feedback information of the second object with the adjusted media content. This ensures the reliability of offline training in a real client environment, thereby improving the accuracy of the first machine learning model after obtaining the first machine learning model based on the second machine learning model.

[0103] Figure 7 This is a schematic diagram of the structure of an information display device in one scenario. The device may include multiple modules, which can be hardware circuits configured to perform corresponding functions, software programs running on a processor, or a combination of both. Figure 7 As shown, the device includes: a first acquisition module 510, a feature acquisition module 520, a second acquisition module 530, and an information display module 540.

[0104] The first acquisition module 510 is used to acquire first information and second information. The first information is used to reflect the first interaction features and the first device features of the first object, and the second information is used to reflect the first content features of the candidate media content. The feature acquisition module 520 is used to analyze and process the first information and the second information based on the first processing module in the first machine learning model to obtain the first feature information; The second acquisition module 530 is used to process the first feature information based on multiple second processing modules in the first machine learning model to obtain the third information. Each second processing module corresponds to a candidate display task. The second processing module is used to present candidate media content and related elements. The related elements are related to the candidate media content. The information display module 540 is used to respond to the display of candidate media content and display candidate media content based on third-party information.

[0105] The aforementioned device, during the process of a first object browsing media content, can reflect the interactive behavior between the first object and the media content played by the first device based on first interaction features, first device features, and first content features. This interactive behavior is incorporated as a component of the environmental state, enabling the first machine learning model to perceive real-time user behavior feedback and session temporal changes. Combining device and content features, it infers a display strategy suitable for candidate media content. Furthermore, the first machine model evaluates the display strategies for at least one candidate display task to obtain a suitable display strategy (i.e., third information) for displaying candidate media content. Based on this third information, the operating parameters of the media playback device are adjusted so that the adjusted parameters allow the media playback device to display the media content. This solves the problem in existing technologies that cannot meet the personalized display needs of media content. By combining the device state of the first device, media content features, and the first object's interactive behavior with the media content, it achieves the effect of intelligently determining the media content to be displayed on the smart device and the display strategy suitable for the associated elements of that media content.

[0106] In one scenario, the feature acquisition unit is used to acquire the object features and behavioral features of the first object, whereby the behavioral features are used to reflect the statistical results of interaction information within multiple candidate time periods. A feature acquisition unit is used to acquire a first device feature, which is related to a first device. The first device feature includes at least one or more of the following: device type, device memory, device network type, and device power level. The first content feature acquisition unit is used to obtain the media attributes of the candidate media content, so as to obtain the first content feature based on the media attributes, wherein the media attributes are related to the candidate media content; The first information determining unit is used to take object features and behavioral features as first information and first content features as second information.

[0107] In one scenario, the first processing module includes a first feature encoding unit, a second feature encoding unit, a feature fusion unit, and a feature sharing unit, and a first encoded feature acquisition unit, used to encode the first information and the second information based on the first feature encoding unit to obtain the first encoded feature; The second coding feature obtaining unit is used to encode the first information and the second information based on the second feature coding unit to obtain the second coding feature; The feature fusion unit is used to fuse the first coding feature and the second coding feature based on the feature fusion unit to obtain the fused feature; The first feature information acquisition unit is used to process the fused features based on the feature sharing unit to obtain the first feature information.

[0108] In one scenario, the second acquisition module is used to perform feature decoupling processing on the first feature information based on the second processing modules for multiple second processing modules to obtain third information, which includes the display strategy corresponding to each candidate display task.

[0109] In one scenario, the second acquisition module includes: a decoupling unit, used to decouple the first feature information into a first state feature and a first action feature based on the second processing module; The second feature information acquisition unit is used to process the first state feature and the first action feature based on the first function to obtain the second feature information; The fourth information acquisition unit is used to acquire fourth information based on the second feature information. The fourth information is used to reflect the evaluation result corresponding to each candidate strategy of the candidate display task. The third information acquisition unit is used to acquire the third information based on the fourth information.

[0110] In one scenario, the information display device includes: a first feedback information acquisition module, configured to acquire first feedback information, wherein the first feedback information indicates feedback information for displaying candidate media content based on third information; The model adjustment module is used to adjust the model parameters of the first machine learning model based on the first feedback information, the first information, and the second information.

[0111] In one scenario, the information display device includes a first sample acquisition module, wherein... The first sample acquisition module includes: an information acquisition unit, used to acquire fifth information and sixth information, the fifth information being used to reflect the second interaction features, the second device features, and the second content features of the first media content of the second object, and the sixth information being used to reflect the set display strategy corresponding to the candidate display task; The second feedback information acquisition unit is used to acquire second feedback information, and the first feedback information is used to reflect the feedback result of the second object to the sixth information. The first sample acquisition unit is used to acquire a first sample based on the first media content, the fifth information, and the second feedback information.

[0112] In one scenario, the information display device includes: a first machine learning model acquisition module, wherein... The first machine learning acquisition module includes: a first prediction result output unit, which is used to input the first media content and the fifth information into the second machine learning model and output the first prediction result, which is used to reflect the prediction and display strategy of the candidate display task; The first attribute acquisition unit is used to obtain the first attribute based on the fifth and sixth information. The first attribute is used to characterize the matching attribute between the prediction display strategy and the expectation. The model parameter adjustment unit is used to adjust the model parameters of the second machine learning model based on the first attribute, the first prediction result, and the first feedback information. The first machine learning acquisition unit is used to obtain the first machine learning model based on the second machine learning model in response to the completion of training.

[0113] In one scenario, the information display device further includes: a third processing module configuration module, configured to configure a third processing module in response to a new candidate display task, the third processing module being related to the new candidate display task, and the third processing module being a module in the first machine learning model; The parameter adjustment module is used to adjust the parameters of the first machine learning model based on the second sample; A module is configured to obtain the first machine learning model in response to training completion. The first machine learning model includes a second processing module associated with the new candidate display task, and the second processing module is a module whose parameters are adjusted from those of the third processing module. The aforementioned information display device can execute the information display method provided in any of the embodiments described herein, and has the corresponding functional modules and beneficial effects for executing the information display method.

[0114] It is worth noting that the various units and modules included in the above-mentioned information display device are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the protection scope of the embodiments in this article.

[0115] The following is for reference. Figure 8 The diagram illustrates a structural schematic of an electronic device (e.g., a terminal device or server) 700 suitable for implementing the above-described methods. The terminal device referred to herein may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, personal digital assistants, tablet computers, portable multimedia players, in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital televisions and desktop computers. Figure 8 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments described herein.

[0116] like Figure 8 As shown, the electronic device 700 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory 702 or a program loaded from a storage device 708 into a random access memory 703. The random access memory 703 also stores various programs and data required for the operation of the electronic device 700. The processing unit 701, the read-only memory 702, and the random access memory 703 are interconnected via a bus 704. An input / output interface 705 is also connected to the bus 704.

[0117] Typically, the following devices can be connected to the input / output interface 705: input devices 706 including, for example, a touchscreen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 707 including, for example, a liquid crystal display, speaker, vibrator, etc.; storage devices 708 including, for example, magnetic tape, hard disk, etc.; and communication devices 709. Communication device 709 allows electronic device 700 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 8 An electronic device 700 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0118] In particular, according to embodiments of this document, the processes described in the above-referenced flowcharts can be implemented as computer software programs. For example, the technical solutions of this document include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 709, or installed from a storage device 708, or installed from a read-only memory 702. When the computer program is executed by a processing device 701, it performs the functions defined in the methods of the embodiments of this document.

[0119] The names of messages or information exchanged between multiple devices in this document are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0120] The electronic device provided in this embodiment and the information display method provided in the above technical solutions belong to the same inventive concept. Technical details not described in detail in this document can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0121] This article provides a computer storage medium on which a computer program is stored, which, when executed by a processor, implements the information display method provided in the above embodiments.

[0122] It should be noted that the computer-readable medium described above can be a computer-readable signal medium, 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, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical fiber, a portable compact disk read-only memory, an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, 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 document, 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, radio frequency, etc., or any suitable combination thereof.

[0123] Based on one or more scenarios described in this article, Example 1 provides an information display method, including: acquiring first information and second information, wherein the first information is used to reflect the first interactive features and the first device features of a first object, and the second information is used to reflect the first content features of candidate media content; Based on the analysis and processing of the first information and the second information by the first processing module in the first machine learning model, the first feature information is obtained; Based on the first machine learning model, multiple second processing modules process the first feature information to obtain the third information. Each second processing module corresponds to a candidate display task. The second processing module is used to output the display strategy of the candidate display task. The candidate display task is related to the presentation of candidate media content and related elements. The related elements are related to the candidate media content. In response to displaying candidate media content, candidate media content is displayed based on third-party information.

[0124] Based on one or more scenarios in this paper, Example 2 provides the method of Example 1, and also includes: optionally, obtaining the object features and behavioral features of the first object, wherein the behavioral features are used to reflect the statistical results of the interaction information within multiple candidate time periods; Obtain a first device feature, which is related to a first device. The first device feature includes at least one or more of the following: device type, device memory, device network type, and device power level. Obtain the media attributes of candidate media content, and based on the media attributes, obtain the first content feature. The media attributes are related to the candidate media content. The object characteristics and behavioral characteristics are used as the first information, and the first content characteristics are used as the second information.

[0125] According to one or more scenarios in this paper, Example 3 provides the method of Example 1, and further includes: Optionally, the first processing module includes a first feature encoding unit, a second feature encoding unit, a feature fusion unit and a feature sharing unit, and encodes the first information and the second information based on the first feature encoding unit to obtain the first encoded feature; The first information and the second information are encoded based on the second feature encoding unit to obtain the second encoded feature; The first and second encoded features are fused using a feature fusion unit to obtain fused features; The fused features are processed based on the feature sharing unit to obtain the first feature information.

[0126] Based on one or more scenarios in this paper, Example 4 provides the method of Example 2, and further includes: Optionally, for multiple second processing modules, feature decoupling processing is performed on the first feature information based on the second processing modules to obtain third information, the third information including the display strategy corresponding to each candidate display task.

[0127] According to one or more scenarios in this paper, Example 5 provides the method of Example 4, which also includes: optionally, decoupling the first feature information into a first state feature and a first action feature based on the second processing module; The first state features and the first action features are processed based on the first function to obtain the second feature information; Based on the second feature information, the fourth information is obtained. The fourth information is used to reflect the evaluation result corresponding to each candidate strategy in the candidate display task. Based on the fourth piece of information, the third piece of information is obtained.

[0128] According to one or more scenarios in this article, Example 6 provides the method of Example 1, which also includes: optionally, obtaining first feedback information, the first feedback information indicating feedback information for displaying candidate media content based on third information; Based on the first feedback information, the first information, and the second information, the model parameters of the first machine learning model are adjusted.

[0129] According to one or more scenarios in this article, Example 7 provides the method of Example 1, which also includes: optionally, obtaining fifth information and sixth information, wherein the fifth information is used to reflect the second interaction characteristics, the second device characteristics and the second content characteristics of the first media content of the second object, and the sixth information is used to reflect the set display strategy corresponding to the candidate display task; Obtain the second feedback information; the first feedback information is used to reflect the feedback result of the second object to the sixth information. The first sample is obtained based on the first media content, the fifth information, and the second feedback information.

[0130] According to one or more scenarios in this paper, Example 8 provides the method of Example 1, which also includes: optionally, inputting the first media content and the fifth information into the second machine learning model and outputting a first prediction result, the first prediction result being used to reflect the prediction and display strategy of the candidate display task; Based on the fifth and sixth information, the first attribute is obtained. The first attribute is used to characterize the adaptation attribute between the prediction display strategy and the expectation. Based on the first attribute, the first prediction result, and the first feedback information, adjust the model parameters of the second machine learning model; In response to the completion of training, the first machine learning model is obtained based on the second machine learning model.

[0131] Based on one or more scenarios in this document, Example 9 provides the method of Example 8, and also includes: Optionally, Configure a third processing module, which is related to the task of adding candidate display. The third processing module is a module in the first machine learning model. The parameters of the first machine learning model are adjusted based on the second sample. Upon completion of training, a first machine learning model is obtained. The first machine learning model includes a second processing module associated with the new candidate display task. The second processing module is a module whose parameters are adjusted based on those of the third processing module.

[0132] According to one or more scenarios described herein, [Example 9] provides an information display device, including: a first acquisition module, used to acquire first information and second information, wherein the first information is used to reflect a first interactive feature and a first device feature of a first object, and the second information is used to reflect a first content feature of candidate media content; The feature acquisition module is used to analyze and process the first information and the second information based on the first processing module in the first machine learning model to obtain the first feature information. The second acquisition module is used to process the first feature information based on multiple second processing modules in the first machine learning model to obtain the third information. Each second processing module corresponds to a candidate display task. The second processing module is used to output the display strategy of the candidate display task. The candidate display task is related to the presentation of candidate media content and related elements. The related elements are related to the candidate media content. The information display module is used to respond to and display candidate media content, based on third-party information.

[0133] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol, such as Hypertext Transfer Protocol, and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (LANs), wide area networks (WANs), the Internet (e.g., the Internet), and peer-to-peer networks, as well as any currently known or future-developed networks.

[0134] 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.

[0135] The aforementioned computer-readable medium carries one or more programs. When the aforementioned one or more programs are executed by the electronic device, the electronic device causes the electronic device to: acquire first information and second information, wherein the first information reflects a first interactive feature and a first device feature of a first object, and the second information reflects a first content feature of candidate media content; analyze and process the first information and the second information based on a first processing module in a first machine learning model to obtain first feature information; process the first feature information based on multiple second processing modules in the first machine learning model to obtain third information, wherein each second processing module corresponds to a candidate display task, and the second processing module is used to output a display strategy for the candidate display task, wherein the candidate display task is related to the presentation of candidate media content and associated elements, and the associated elements are related to the candidate media content; and display the candidate media content based on the third information in response to displaying the candidate media content.

[0136] Computer program code for performing the operations described herein can be written in one or more programming languages ​​or a combination thereof, including but not limited to object-oriented programming languages, as well as conventional procedural programming languages—such as the "C" programming 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 separate 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 local area networks (LANs) or wide area networks (WANs), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0137] 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 document. 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 the 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, may 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.

[0138] The modules or units described herein can be implemented in software or hardware. The names of modules or units do not necessarily constitute a limitation on the module or unit itself.

[0139] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that can be used include at least one of the following: field-programmable gate arrays, application-specific integrated circuits (ASICs), application-specific standard products, systems-on-a-chip (SoCs), complex programmable logic devices, and so on.

[0140] In the context of this document, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory, read-only memory, erasable programmable read-only memory (flash memory), optical fibers, portable compact disk read-only memory, optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0141] The above description is merely a preferred embodiment and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure herein 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 herein that have similar functions.

[0142] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain circumstances, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of this document. Certain features described in the context of individual implementations may also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation may also be implemented individually or in any suitable sub-combination in multiple implementations.

[0143] Although the subject matter has been described using a programming language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative forms of implementing the claims.

Claims

1. An information display method, comprising: Acquire first information and second information, wherein the first information is used to reflect the first interaction characteristics and the first device characteristics of the first object, and the second information is used to reflect the first content characteristics of the candidate media content; Based on the analysis and processing of the first information and the second information by the first processing module in the first machine learning model, the first feature information is obtained; The first feature information is processed by multiple second processing modules in the first machine learning model to obtain third information. Each second processing module corresponds to a candidate display task. The second processing module is used to output the display strategy of the candidate display task. The candidate display task is related to the presentation of the candidate media content and related elements. The related elements are related to the candidate media content. In response to displaying the candidate media content, the candidate media content is displayed based on the third information.

2. The information display method according to claim 1, wherein obtaining the first information and the second information includes: Obtain the object features and behavioral features of the first object, wherein the behavioral features are used to reflect the statistical results of interaction information within multiple candidate time periods; Obtain a first device feature, which is related to a first device, and the first device feature includes at least one or more of the following: device type, device memory, device network type, and device battery level. Obtain media attributes to obtain the first content feature based on the media attributes, wherein the media attributes are related to the candidate media content; The object characteristics and the behavioral characteristics are used as the first information; The first content feature is used as the second information.

3. The information display method according to claim 1, wherein the first processing module includes a first feature encoding unit, a second feature encoding unit, a feature fusion unit, and a feature sharing unit, and the first processing module in the first machine learning model analyzes and processes the first information and the second information to obtain the first feature information, including: The first information and the second information are encoded based on the first feature encoding unit to obtain the first encoded feature; The first information and the second information are encoded based on the second feature encoding unit to obtain the second encoding feature; The feature fusion unit performs fusion processing on the first encoded feature and the second encoded feature to obtain the fused feature; The fused features are processed by the feature sharing unit to obtain the first feature information.

4. The information display method according to claim 1, wherein the step of processing the first feature information based on the multiple second processing modules in the first machine learning model to obtain the third information includes: For multiple second processing modules, feature decoupling processing is performed on the first feature information based on the second processing modules to obtain the third information, which includes the display strategy corresponding to each candidate display task.

5. The information display method according to claim 4, wherein the step of performing feature decoupling processing on the first feature information based on the second processing module to obtain the third information includes: Based on the second processing module, the first feature information is decoupled into a first state feature and a first action feature; The first state feature and the first action feature are processed based on the first function to obtain the second feature information; Based on the second feature information, fourth information is obtained, which is used to reflect the evaluation result corresponding to each candidate strategy of the candidate display task; Based on the fourth information, the third information is obtained.

6. The information display method according to claim 1, further comprising: Obtain first feedback information, which indicates feedback information for displaying candidate media content based on the third information; Based on the first feedback information, the first information, and the second information, the model parameters of the first machine learning model are adjusted.

7. The information display method according to claim 1, wherein the first sample is obtained in the following manner, and the first sample is used to train the first machine learning model: The fifth information and the sixth information are obtained. The fifth information is used to reflect the second interaction characteristics, the second device characteristics, and the second content characteristics of the first media content of the second object. The sixth information is used to reflect the set display strategy corresponding to the candidate display task. Obtain second feedback information, which reflects the feedback result of the second object to the sixth information; The first sample is obtained based on the first media content, the fifth information, and the second feedback information.

8. The information display method according to claim 7, wherein the first machine learning model is obtained in the following manner: The first media content and the fifth information are input into the second machine learning model, and the first prediction result is output. The first prediction result is used to reflect the prediction and display strategy of the candidate display task. Based on the fifth and sixth information, a first attribute is obtained, which is used to characterize the compatibility between the prediction display strategy and the expected attributes. Based on the first attribute, the first prediction result, and the second feedback information, adjust the model parameters of the second machine learning model; In response to the completion of training, the first machine learning model is obtained based on the second machine learning model.

9. The information display method according to claim 8, further comprising: In response to the new candidate display task, a third processing module is configured. The third processing module is related to the new candidate display task and is a module in the first machine learning model. The parameters of the first machine learning model are adjusted based on the second sample; In response to training completion, a first machine learning model is obtained, the first machine learning model including a second processing module associated with the new candidate display task, the second processing module being a module with adjusted parameters of the third processing module.

10. An information display device, comprising: The first acquisition module is used to acquire first information and second information. The first information is used to reflect the first interaction features and the first device features of the first object, and the second information is used to reflect the first content features of the candidate media content. The feature acquisition module is used to analyze and process the first information and the second information based on the first processing module in the first machine learning model to obtain first feature information; The second acquisition module is used to process the first feature information based on multiple second processing modules in the first machine learning model to obtain third information. Each second processing module corresponds to a candidate display task. The second processing module is used to output the display strategy of the candidate display task. The candidate display task is related to the presentation of the candidate media content and related elements. The related elements are related to the candidate media content. The information display module is used to display the candidate media content based on the third information in response to the display of the candidate media content.

11. An electronic device, the electronic device comprising: One or more processors; Storage device, used to store program instructions When the processor reads and executes the program instructions, it implements the information display method as described in any one of claims 1-9.

12. A non-transitory computer-readable storage medium, wherein computer program instructions are stored internally in the storage medium; When the processor executes the computer program instructions, it performs the information display method as described in any one of claims 1-9.

13. A computer program product comprising a computer program that, when loaded and run by a processor of an electronic device, controls the processor to execute the information display method as described in any one of claims 1-9.

Images