Multimedia resource delivery methods, devices, electronic devices, storage media and products

Abstract

This disclosure relates to a multimedia resource push method, apparatus, electronic device, storage medium, and product. The method includes: obtaining a target association graph; obtaining initial object category indicators for multiple preset objects; based on the initial object category indicator data, performing bidirectional transmission of indicator data between nodes with preset association edges to obtain target object category indicator data for multiple preset objects and target interest indicator data for historical multimedia resources; filtering the first object set based on the target object category indicator data of preset objects in the first object set to obtain target objects; determining associated multimedia resources of the first object category in historical multimedia resources based on the target interest indicator data; and pushing associated multimedia resources to the target objects. Using embodiments of this disclosure, objects with abnormal interests in the preset object set can be filtered out in a timely manner, and the multimedia resources preferred by each type of object can be accurately identified, thus improving the efficiency of multimedia resource push.

Description

Technical Field

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

[0002] With the rapid development of Internet technology, information flow content services have become widely popular among users, generating a large amount of multimedia content.

[0003] The content distribution methods used in related technologies typically classify users based on static characteristics such as region, age, and gender, and then distribute the multimedia content that each type of user has watched the most to users in the corresponding category who have not yet watched it.

[0004] However, among users with different static characteristics, there are often users with inconsistent interests and preferences, resulting in inaccurate identification of user interests and preferences and low efficiency in resource delivery. Summary of the Invention

[0005] This disclosure provides a multimedia resource push method, apparatus, electronic device, storage medium, and product to at least solve the problem of low resource push efficiency in related technologies. The technical solution of this disclosure is as follows:

[0006] According to a first aspect of the present disclosure, a multimedia resource push method is provided, comprising:

[0007] Obtain a target association graph, which includes a first node and a second node connected by a preset association edge. The first node represents multiple preset objects, and the second node represents historical multimedia resources that the multiple preset objects have acted upon. The preset association edge is generated based on the historical behavior data of the multiple preset objects.

[0008] Obtain initial object category index data for multiple preset objects;

[0009] Based on the initial object category index data, bidirectional transmission of index data is performed between nodes with the preset associated edges to obtain target object category index data of multiple preset objects and target interest index data of the historical multimedia resources.

[0010] Based on the target object category index data of the preset objects in the first object set, the first object set is filtered to obtain the target objects. The first object set refers to the set of preset objects that belong to the first object category among the multiple preset objects.

[0011] Based on the target interest index data, determine the associated multimedia resources of the first object category in the historical multimedia resources;

[0012] The associated multimedia resources are pushed to the target object.

[0013] In one possible design, based on the initial object category index data, bidirectional transmission of index data is performed between nodes with the preset associated edges to obtain target object category index data for multiple preset objects and target interest index data for the historical multimedia resources, including:

[0014] The object category index data of multiple preset objects are forward-transmitted through the preset association edge to update the interest index of the historical multimedia resources, thereby obtaining the interest index difference data of the historical multimedia resources. The object category index data is the initial object category index data during the first forward transmission, and the interest index difference data refers to the difference data of the interest index of the historical multimedia resources before and after the forward transmission.

[0015] The interest index data of the historical multimedia resources are reverse-transmitted through the preset association edge to obtain the category index difference data of multiple preset objects. The category index difference data refers to the difference data of the object category index of the multiple preset objects before and after the reverse transmission.

[0016] When both the interest index difference data and the category index difference data meet the preset difference conditions, the target object category index data of multiple preset objects and the target interest index data of the historical multimedia resources are obtained.

[0017] In one possible design, the step of forward propagating object category index data of multiple preset objects through the preset association edge to update the interest index of the historical multimedia resources and obtaining the interest index difference data of the historical multimedia resources includes:

[0018] The current object category index data is forward-transmitted through the preset association edge to obtain the updated interest index data of the historical multimedia resources. The current object category index is the initial object category index data during the first forward transmission, and the current object category index is the updated object category index data obtained from the previous reverse transmission during subsequent forward transmissions.

[0019] Use the updated interest metric data as the current interest metric data;

[0020] If the number of forward passes is greater than or equal to the first preset number, the first difference data between the current interest index data and the updated interest index data obtained from the previous forward pass is determined.

[0021] In one possible design, the interest index data of the historical multimedia resources is passed back through the preset association edge to obtain category index difference data of multiple preset objects, including:

[0022] If the second difference data does not meet the second preset condition, the current interest index data is passed back through the preset association edge to obtain updated object category index data. The second difference data is the difference data between the current object category index data and the updated object category index data obtained from the previous back transmission.

[0023] Use the updated object category index data as the current object category index data;

[0024] If the number of reverse propagations is greater than or equal to the second preset number, a third difference data is determined between the current object category index data and the updated object category index data obtained from the previous reverse propagation.

[0025] In one possible design, when both the interest index difference data and the category index difference data satisfy preset difference conditions, obtaining the target object category index data of multiple preset objects and the target interest index data of the historical multimedia resources includes:

[0026] If the first difference data satisfies the first preset condition and the third difference data satisfies the second preset condition, the current interest index data is used as the target interest index data, and the current object category index data is used as the target object category index data.

[0027] In one possible design, the method further includes:

[0028] If the number of forward transmissions is less than the first preset number or the first difference data does not meet the first preset condition, the process jumps to the step of transmitting the current interest index data in reverse through the preset association edge to obtain the updated object category index data.

[0029] If the number of reverse transmissions is less than the second preset number or the third difference data does not meet the second preset condition, the process jumps to the step of forward transmission of the current object category index data through the preset association edge to obtain the updated interest index data of the historical multimedia resources.

[0030] In one possible design, the step of forward propagating the current object category index data through the preset association edge to obtain the updated interest index data of the historical multimedia resources includes:

[0031] Based on the current object category index data, determine the transmitted data on the preset associated edge;

[0032] The transmitted data is transmitted to the second node via the preset associated edge;

[0033] Based on the transmitted data received by the second node, updated interest index data of the historical multimedia resources are obtained.

[0034] In one possible design, determining the transmitted data on the preset associated edge based on the current object category index data includes:

[0035] Traverse the first node in the target association graph;

[0036] When traversing to any first node, determine the first preset associated edge that connects to the currently traversed first node;

[0037] Determine the weight information of the first preset associated edge;

[0038] Based on the current object category index data corresponding to the first traversed node and the weight information, determine the transmitted data on the first preset associated edge;

[0039] The transmitted data on the first preset associated edge connected to each first node is determined as the transmitted data on the preset associated edge.

[0040] In one possible design, obtaining the updated interest index data of the historical multimedia resources based on the transmitted data received by the second node includes:

[0041] Traverse the second node in the target association graph;

[0042] When traversing to any second node, obtain the second data passed to the currently traversed second node;

[0043] Based on the second transmitted data, the updated interest index data corresponding to the currently traversed second node is generated;

[0044] The updated interest index data obtained during the traversal process will be used as the updated interest index data for the historical multimedia resources.

[0045] In one possible design, generating updated interest metric data corresponding to the currently traversed second node based on the second transmitted data includes:

[0046] Based on the second transmitted data, determine the second cumulative transmitted data on the currently traversed second node;

[0047] Determine the number of second preset associated edges connected to the currently traversed second node;

[0048] Based on the number of the second cumulative transmitted data and the second preset associated edges, the updated interest index data corresponding to the currently traversed second node is determined.

[0049] In one possible design, the step of passing the current interest index data back through the preset association edge to obtain updated object category index data includes:

[0050] Based on the current interest index data, determine the transmitted data on the preset associated edge;

[0051] The transmitted data is transmitted to the first node via the preset associated edge;

[0052] Based on the transmitted data received by the first node, the updated object category index data is obtained.

[0053] In one possible design, determining the transmitted data on the preset associated edge based on the current interest index data includes:

[0054] Traverse the second node in the target association graph;

[0055] When traversing to any second node, determine the second preset associated edge that connects to the currently traversed second node;

[0056] Determine the weight information of the second preset associated edge;

[0057] Based on the current interest index data corresponding to the currently traversed second node and the weight information, determine the transmitted data on the second preset associated edge;

[0058] The transmitted data on the second preset associated edge connected to each second node is determined as the transmitted data on the preset associated edge.

[0059] In one possible design, obtaining the updated object category index data based on the transmitted data received by the first node includes:

[0060] Traverse the first node in the target association graph;

[0061] When traversing to any first node, obtain the first data passed to the currently traversed first node;

[0062] Based on the first transmitted data, an updated object category index corresponding to the first traversed node is generated;

[0063] The updated object category index data obtained during the traversal process is used as the updated object category index data for the plurality of preset objects.

[0064] In one possible design, generating the updated object category index corresponding to the currently traversed first node based on the first transmitted data includes:

[0065] Based on the first transmitted data, determine the first cumulative transmitted data on the first node currently traversed;

[0066] Determine the number of the first preset associated edges connected to the currently traversed first node;

[0067] Based on the number of the first cumulative transmitted data and the first preset associated edges, the updated object category index data corresponding to the first traversed node is determined.

[0068] In one possible design, filtering the first object set based on target object category index data of preset objects in the first object set to obtain target objects includes:

[0069] The objects in the first object set whose target object category index data do not meet the preset index conditions are identified as abnormal objects. The preset index conditions are the object category index constraints corresponding to the first object category.

[0070] The abnormal object is removed from the first object set to obtain the target object.

[0071] In one possible design, determining the associated multimedia resources for the first object category in the historical multimedia resources based on the target interest index data includes:

[0072] The multimedia resources in the historical multimedia resources whose target interest index data meets the first interest index condition are identified as the associated multimedia resources, where the first interest index condition is the interest index constraint condition corresponding to the first object category.

[0073] In one possible design, the method further includes:

[0074] Based on the target object category index data of the abnormal object, determine the second object category corresponding to the abnormal object;

[0075] Based on the target interest index data, determine the associated multimedia resources of the second object category in the historical multimedia resources;

[0076] Push the associated multimedia resources of the second object category to the abnormal object.

[0077] In one possible design, the method further includes:

[0078] Obtain object feature data of multiple preset objects;

[0079] Based on the object feature data, determine the object categories corresponding to the multiple preset objects.

[0080] In one possible design, obtaining the initial object category indexes of the multiple preset objects includes:

[0081] Based on the object categories corresponding to the plurality of preset objects, initial object category index data for the plurality of preset objects is determined. According to a second aspect of the present disclosure, a multimedia resource push device is provided, comprising:

[0082] The association graph acquisition module is configured to acquire a target association graph, which includes a first node and a second node connected by preset association edges. The first node represents multiple preset objects, and the second node represents historical multimedia resources that the multiple preset objects have acted upon. The preset association edges are generated based on the historical behavior data of the multiple preset objects.

[0083] The indicator acquisition module is configured to acquire initial object category indicator data for multiple preset objects;

[0084] The indicator bidirectional transmission module is configured to perform bidirectional transmission of indicator data between nodes with the preset associated edges based on the initial object category indicator data, so as to obtain target object category indicator data of multiple preset objects and target interest indicator data of the historical multimedia resources.

[0085] The interest filtering module is configured to filter the first object set based on the target object category index data of preset objects in the first object set to obtain the target object. The first object set refers to the set of preset objects belonging to the first object category among multiple preset objects.

[0086] The resource determination module is configured to determine the associated multimedia resources of the first object category in the historical multimedia resources based on the target interest index data.

[0087] The resource push module is configured to push the associated multimedia resources to the target object.

[0088] In one possible design, the bidirectional indicator transmission module includes:

[0089] The first difference data determination submodule is configured to perform forward transmission of object category index data of multiple preset objects through the preset association edge, update the interest index of the historical multimedia resources, and obtain the interest index difference data of the historical multimedia resources. The object category index data is the initial object category index data during the first forward transmission, and the interest index difference data refers to the difference data of the interest index of the historical multimedia resources before and after the forward transmission.

[0090] The second difference data determination submodule is configured to perform reverse transmission of the interest index data of the historical multimedia resources through the preset association edge to obtain the category index difference data of multiple preset objects. The category index difference data refers to the difference data of the object category index of the multiple preset objects before and after the reverse transmission.

[0091] The indicator data determination module is configured to obtain target object category indicator data of multiple preset objects and target interest indicator data of historical multimedia resources when both the interest indicator difference data and the category indicator difference data meet preset difference conditions.

[0092] In one possible design, the first difference data determining submodule includes:

[0093] The indicator forward transmission submodule is configured to perform forward transmission of the current object category indicator data through the preset association edge to obtain the updated interest indicator data of the historical multimedia resources. The current object category indicator is the initial object category indicator data during the first forward transmission, and the current object category indicator is the updated object category indicator data obtained from the previous reverse transmission during subsequent forward transmissions.

[0094] The interest metric update submodule is configured to use the updated interest metric data as the current interest metric data.

[0095] The difference data determination submodule is configured to determine the first difference data between the current interest index data and the updated interest index data obtained in the previous forward pass, provided that the number of forward passes is greater than or equal to a first preset number.

[0096] In one possible design, the second difference data determining submodule includes:

[0097] The indicator reverse transmission submodule is configured to, when the second difference data does not meet the second preset condition, reversely transmit the current interest indicator data through the preset association edge to obtain updated object category indicator data. The second difference data is the difference data between the current object category indicator data and the updated object category indicator data obtained from the previous reverse transmission.

[0098] The category index update submodule is configured to use the updated object category index data as the current object category index data;

[0099] The difference data determination submodule is further configured to determine the third difference data between the current object category index data and the updated object category index data obtained from the previous reverse pass, provided that the number of reverse passes is greater than or equal to the second preset number.

[0100] In one possible design, the indicator data determination module is specifically configured to perform:

[0101] If the first difference data satisfies the first preset condition and the third difference data satisfies the second preset condition, the current interest index data is used as the target interest index data, and the current object category index data is used as the target object category index data.

[0102] In one possible design, the indicator reverse transmission submodule is further configured to execute the step of reversing the current interest indicator data through the preset association edge to obtain updated object category indicator data if the number of forward transmissions is less than the first preset number or the first difference data does not meet the first preset condition.

[0103] The forward transmission submodule is further configured to, if the number of reverse transmissions is less than the second preset number or the third difference data does not meet the second preset condition, jump to the step of forward transmission of the current object category index data through the preset association edge to obtain the updated interest index data of the historical multimedia resources.

[0104] In one possible design, the indicator forward propagation submodule includes:

[0105] The data transmission determination unit is configured to determine the transmission data on the preset associated edge based on the current object category index data;

[0106] The data transmission unit is configured to transmit the transmitted data to the second node through the preset associated edge;

[0107] The interest index update unit is configured to execute the updated interest index data of the historical multimedia resources based on the transmitted data received by the second node.

[0108] In one possible design, the data transmission determination unit includes:

[0109] The node traversal sub-unit is configured to perform a traversal of the first node in the target association graph;

[0110] The associated edge determination sub-unit is configured to determine the first preset associated edge connected to the currently traversed first node when any first node is traversed.

[0111] The weight determination subunit is configured to determine the weight information of the first preset associated edge;

[0112] The data transmission determination subunit is configured to determine the transmission data on the first preset associated edge based on the current object category index data corresponding to the currently traversed first node and the weight information.

[0113] The data collection subunit is configured to execute the data transfer on the first preset associated edge connecting each first node, and to determine the data transfer on the preset associated edge.

[0114] In one possible design, the interest metric update unit includes:

[0115] The node traversal sub-unit is configured to perform a traversal of the second node in the target association graph;

[0116] The data acquisition subunit is configured to acquire the second data to be acquired when any second node is reached during traversal.

[0117] The interest index generation subunit is configured to generate updated interest index data corresponding to the currently traversed second node based on the second transmitted data.

[0118] The interest index collection subunit is configured to use the updated interest index data obtained during the traversal process as the updated interest index data for the historical multimedia resources.

[0119] In one possible design, the interest index generation subunit is specifically configured to perform the following actions: determining the second cumulative transmitted data on the currently traversed second node based on the second transmitted data; determining the number of second preset associated edges connected to the currently traversed second node; and determining the updated interest index data corresponding to the currently traversed second node based on the second cumulative transmitted data and the number of the second preset associated edges.

[0120] In one possible design, the indicator back-passing submodule includes:

[0121] The data transmission determination unit is configured to determine the data to be transmitted on the preset associated edge based on the current interest index data;

[0122] The data transmission unit is configured to transmit the transmitted data to the first node through the preset associated edge;

[0123] The category index update unit is configured to execute the updated object category index data based on the transmitted data received by the first node.

[0124] In one possible design, the data transmission determination unit includes:

[0125] The node traversal sub-unit is configured to perform a traversal of the second node in the target association graph;

[0126] The associated edge determination sub-unit is configured to determine a second preset associated edge connecting to the currently traversed second node when any second node is traversed.

[0127] The weight determination subunit is configured to determine the weight information of the second preset associated edge;

[0128] The data transmission determination subunit is configured to determine the data to be transmitted on the second preset associated edge based on the current interest index data corresponding to the currently traversed second node and the weight information.

[0129] The data collection subunit is configured to execute the transfer of data on the second preset associated edge connecting each second node, and to determine the transfer data on the preset associated edge.

[0130] In one possible design, the category index update unit includes:

[0131] The node traversal sub-unit is configured to perform a traversal of the first node in the target association graph;

[0132] The data acquisition subunit is configured to acquire the first data to be acquired when any first node is reached during traversal.

[0133] The category index generation subunit is configured to generate an updated object category index corresponding to the currently traversed first node based on the first transmitted data.

[0134] The category index collection subunit is configured to use the updated object category index data obtained during the traversal process as the updated object category index data for the plurality of preset objects.

[0135] In one possible design, the category index generation subunit is specifically configured to perform: based on the first transmitted data, determine the first cumulative transmitted data on the currently traversed first node; determine the number of first preset associated edges connected to the currently traversed first node; and based on the first cumulative transmitted data and the number of the first preset associated edges, determine the updated object category index data corresponding to the currently traversed first node.

[0136] In one possible design, the interest filtering module includes:

[0137] The abnormal interest object determination unit is configured to determine the target object category index data in the first object set as an abnormal object if the data does not meet the preset index conditions. The preset index conditions are the object category index constraints corresponding to the first object category.

[0138] The abnormal interest object removal unit is configured to remove the abnormal object from the first object set to obtain the target object.

[0139] In one possible design, the resource determination module is specifically configured to determine the multimedia resources in the historical multimedia resources whose target interest index data meets the first interest index condition as the associated multimedia resources, where the first interest index condition is the interest index constraint condition corresponding to the first object category.

[0140] In one possible design, the device further includes:

[0141] The object category determination module is configured to execute target object category index data based on the abnormal object to determine a second object category corresponding to the abnormal object;

[0142] The resource determination module is further configured to determine the associated multimedia resources of the second object category in the historical multimedia resources based on the target interest index data.

[0143] The resource push module is also configured to push associated multimedia resources of the second object category to the abnormal object.

[0144] In one possible design, the device further includes:

[0145] The feature data acquisition module is configured to acquire object feature data of multiple preset objects;

[0146] The object classification module is configured to determine the object category corresponding to multiple preset objects based on the object feature data.

[0147] In one possible design, the indicator acquisition module is specifically configured to determine initial object category indicator data for a plurality of preset objects based on the object categories corresponding to the preset objects.

[0148] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the method as described in any one of the first aspects above.

[0149] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided such that, when instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform any of the methods described in the first aspect of the present disclosure.

[0150] According to a fifth aspect of the present disclosure, a computer program product including instructions is provided that, when run on a computer, causes the computer to perform any of the methods described in the first aspect of the present disclosure.

[0151] The technical solutions provided by the embodiments of this disclosure have at least the following beneficial effects:

[0152] By obtaining the preset association edge between the first node representing the preset object and the second node representing historical multimedia resources in the target association graph, and based on the initial object category index of the preset object, bidirectional transmission of index data is performed between nodes with preset association edges in the target association graph to obtain the object category index data of the preset object and the interest index data of the preset object on historical multimedia resources. Then, the preset object set is filtered according to the object category index data to redetermine the target users belonging to the object category. Finally, the multimedia resources associated with the object category determined based on the interest index data are pushed to the target users. This can filter out objects with abnormal interests in the preset object set. While ensuring the consistency of the interests and preferences of each object in the preset object set, it can accurately identify the multimedia resources preferred by each type of object and push them accordingly, thereby improving the efficiency of multimedia resource push and reducing the operating pressure of loading multimedia resources on the server and terminal.

[0153] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0154] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0155] Figure 1 This is a schematic diagram illustrating an application environment according to an exemplary embodiment;

[0156] Figure 2 This is a flowchart illustrating a multimedia resource push method according to an exemplary embodiment;

[0157] Figure 3 This is a schematic diagram of a target association diagram in an initial state according to an exemplary embodiment;

[0158] Figure 4 This is a flowchart illustrating a multimedia resource push method according to an exemplary embodiment;

[0159] Figure 5 This is a flowchart illustrating a multimedia resource push method according to an exemplary embodiment;

[0160] Figure 6 This is a schematic diagram illustrating a target association graph in a bidirectional data transmission state according to an exemplary embodiment;

[0161] Figure 7 This is a schematic diagram of a target association graph in an end state according to an exemplary embodiment;

[0162] Figure 8 This is a block diagram illustrating a multimedia resource push device according to an exemplary embodiment;

[0163] Figure 9 This is a block diagram illustrating an electronic device for pushing multimedia resources according to an exemplary embodiment. Detailed Implementation

[0164] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0165] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0166] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data used for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties.

[0167] Please see Figure 1 , Figure 1 This is a schematic diagram illustrating an application environment according to an exemplary embodiment, such as... Figure 1 As shown, the application environment may include terminal 100 and server 200.

[0168] Terminal 100 can be used to provide multimedia resource push services to any user. Specifically, terminal 100 can be, but is not limited to, electronic devices such as smartphones, desktop computers, tablets, laptops, smart speakers, digital assistants, augmented reality (AR) / virtual reality (VR) devices, and smart wearable devices, or software running on the aforementioned electronic devices, such as applications. Optionally, the operating system running on the electronic device can be, but is not limited to, Android, iOS, Linux, Windows, etc.

[0169] In an optional embodiment, server 200 can provide background services to terminal 100, pushing multimedia resources to be displayed to terminal 100. Specifically, server 200 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0170] In addition, it should be noted that, Figure 1The example shown is merely one application environment provided by this disclosure. In practical applications, other application environments may also be included, such as more terminals.

[0171] In the embodiments described in this specification, the terminal 100 and the server 200 can be directly or indirectly connected through wired or wireless communication, and this disclosure does not impose any restrictions.

[0172] Figure 2 This is a flowchart illustrating a multimedia resource delivery method according to an exemplary embodiment, such as... Figure 2 As shown, the multimedia resource push method can be applied to electronic devices such as terminals or servers, and includes the following steps (210-260).

[0173] In step S210, the target association graph is obtained.

[0174] The target association graph includes a first node and a second node connected by a preset association edge. The target association graph formed by the first node and the second node can be a bipartite graph. In a bipartite graph, there is no preset association edge connection relationship between the first node and the second node, but there can be a preset association edge connection relationship between the first node and the second node.

[0175] The aforementioned first node represents multiple preset objects. Optionally, the preset objects are users, each first node corresponds to a user account, and each user account can point to a user. Thus, the aforementioned first node can be used to represent a user.

[0176] The second node represents historical multimedia resources that have been accessed by multiple preset objects. These preset objects' actions can be related operations on the multimedia resources, such as watching, liking, commenting, forwarding, or deleting. Optionally, each second node corresponds to one multimedia object. The multimedia resources can be composed of video, audio, text / images, or multiple modalities.

[0177] The preset association edges are generated based on the historical behavior data of multiple preset objects. This historical behavior data refers to the data generated by the operations performed by multiple preset objects on multimedia resources. The preset association edges between the first node and the second node in the target association graph can be determined based on the historical behavior data. For example, if preset object A has performed an operation on multimedia resource B, a preset association edge is generated between the first node corresponding to preset object A and the second node corresponding to multimedia resource B. This preset association edge can represent that preset object A has performed an operation on multimedia resource B.

[0178] In step S220, initial object category index data of multiple preset objects are obtained.

[0179] Acquire object feature data for multiple preset objects. This object feature data can be feature data representing user characteristics. Alternatively, the object feature data can be static feature data of the preset objects, representing their static characteristics. Optionally, static features include attribute information of the preset objects such as the user's age, gender, and city of residence.

[0180] Based on object feature data, the system determines the object categories corresponding to multiple preset objects. During initialization, the system needs to classify multiple preset objects according to their feature data, with preset objects having the same static characteristics belonging to the same object category. For example, preset objects aged 18-23 are assigned to the first object category, and preset objects of other age groups are assigned to the second object category. This embodiment does not limit the division of object categories and can be adaptively adjusted according to specific application scenarios.

[0181] Based on the object categories corresponding to multiple preset objects, initial object category index data for these preset objects is determined. Before bidirectional transmission of index data between nodes in the target association graph, the index data corresponding to each node in the target association graph is first set. Specifically, the first node, corresponding to the preset object, has object category index data, which characterizes the probability that the preset object belongs to the preset object category; the second node, corresponding to the multimedia resource, has interest index data, which characterizes the degree of interest of users of the preset object category in the multimedia resource.

[0182] In one possible implementation, different initial object category metric data can be set for preset objects of different object categories. This object category metric data can be a specific score used to differentiate between different users. For example, the initial score of preset objects in a first object category is higher than the initial score of preset objects in a second object category.

[0183] Optionally, initial interest index data for the second node can also be set. This initial interest index data can be a score, with the score being positively correlated with the preset object's level of interest in multimedia resources. In specific applications, during initialization, the initial interest index data corresponding to the second node is set to 0.

[0184] In one example, such as Figure 3 As shown, it exemplifies a schematic diagram of a target association graph. In Figure 3 In the target association diagram shown, circles represent the first node, and one first node represents one user; triangles represent the second node, and one second node represents one multimedia resource. Figure 3The image displays six first nodes: Node 311, Node 312, Node 313, Node 314, Node 315, and Node 316. Nodes 311, 312, and 313 belong to the first object category, while nodes 314, 315, and 316 belong to the second object category. Furthermore, nodes 311, 312, and 313 share the same color, but are darker than nodes 314, 315, and 316. This color indicates that first nodes belonging to the same object category have the same initial score, with the initial score for the first object category being higher than that for the second object category. The initial score is positively correlated with the node's color intensity.

[0185] Figure 3 The system also displays six second nodes: second node 321, second node 322, second node 323, second node 324, second node 325, and second node 326. During initialization, the initial score of each second node is set to 0.

[0186] Figure 3 In the above, the preset objects represented by the first node 311 have performed corresponding operations on the multimedia resources corresponding to the second nodes 321, 322, and 324 respectively. Therefore, there are preset associated edges between the first node 311 and the second nodes 321, 322, and 324. Other preset associated edges can be determined by analogy based on the same reasoning.

[0187] In step S230, based on the initial object category index data, the index data is transmitted bidirectionally between nodes with preset associated edges to obtain the target object category index data of multiple preset objects and the target interest index data of historical multimedia resources.

[0188] In one possible implementation, BIMUI (Bidirectional Interest Model For Users And Items), a graph model-based mining algorithm, is used to iteratively train a target association graph model composed of preset objects and multimedia resources based on operational behavior relationships. This determines the target object category index data of multiple preset objects and the target interest index data of historical multimedia resources. These two types of index data are used to remove and weaken the abnormal data corresponding to each object category, ultimately obtaining the preferred videos corresponding to the preset objects for each object category.

[0189] The general process of bidirectional transmission of indicator data is as follows: based on the connection relationship of the preset association edges in the target association graph, the initial object category indicator data of the first node is transmitted to the second node for the first time to generate the interest indicator data of the second node. Then, the generated interest indicator data is transmitted back to the first node to update the object category indicator data of the first node. This process is iterated multiple times, and the indicator data of the two nodes are updated according to a certain learning rate until the indicator data of the two nodes meet the convergence condition, thus obtaining the target object category indicator data and target interest indicator data.

[0190] The target object category indicator data mentioned above is the stable object category indicator data corresponding to each preset object after the bidirectional transmission of indicator data. There may be differences between the target object category indicator data of each preset object and the initial object category indicator data. The target object category indicator data is data that can reflect user preferences.

[0191] The aforementioned target interest index data are stable interest index data corresponding to each multimedia resource after the bidirectional transmission of index data is completed. The target interest index data can accurately reflect the degree of preference of the preset object category for multimedia resources.

[0192] The bidirectional transmission process of the above-mentioned indicator data can be referred to the description in the following embodiments.

[0193] In an exemplary embodiment, the bidirectional transmission process of indicator data includes the following:

[0194] The object category index data of multiple preset objects are forward-propagated through preset association edges to update the interest index of historical multimedia resources, thus obtaining the interest index difference data of historical multimedia resources. The object category index data is the initial object category index data during the first forward propagation, while the interest index difference data refers to the difference in the interest index of historical multimedia resources before and after the forward propagation.

[0195] Interest index data from historical multimedia resources is passed backwards through pre-defined association edges to obtain category index difference data for multiple pre-defined objects. Here, category index difference data refers to the difference in object category indices before and after the backward pass.

[0196] When both the interest index difference data and the category index difference data meet the preset difference conditions, the target object category index data of multiple preset objects and the target interest index data of historical multimedia resources are obtained.

[0197] The above process will be further explained below.

[0198] In an exemplary embodiment, such as Figure 4As shown, the implementation process of step 230 above includes the following steps (S23a~S23j).

[0199] In step S23a, the current object category index data is forward-transmitted through a preset association edge to obtain the updated interest index data of historical multimedia resources.

[0200] The aforementioned forward transmission direction is from the first node to the second node. During the bidirectional transmission of indicator data, there is at least one forward transmission and at least one reverse transmission. Each forward transmission updates the interest indicator of the second node; each reverse transmission updates the object category indicator of the first node. The current object category indicator data is the object category indicator data updated in the latest reverse transmission. The current object category indicator data is the initial object category indicator data during the first forward transmission. The current object category indicator data is the updated object category indicator data obtained in the previous reverse transmission during subsequent forward transmissions.

[0201] The direction of the reverse propagation is from the second node side to the first node side. For an introduction to the reverse propagation, please refer to the following content, which will not be elaborated here.

[0202] The forward propagation process in step S23a is described here through an example.

[0203] In an exemplary embodiment, such as Figure 5 As shown, the implementation process of step S23a above includes the following steps (S23a1 to S23a3).

[0204] In step S23a1, the transmitted data on the preset associated edge is determined based on the current object category index data.

[0205] Traverse the first node in the target association graph.

[0206] When traversing to any first node, determine the first preset associated edge connecting it to the currently traversed first node. The aforementioned first preset associated edge refers to a preset associated edge that has a connection relationship with the first node. If the first node does not have a preset associated edge connecting it, then the first node does not have the aforementioned first preset associated edge, and traversal can proceed to the next first node.

[0207] The weight information of the first preset associated edges is determined. This weight information includes the weight values ​​of each first preset associated edge connected to the currently traversed first node. These weight values ​​can be preset weight values ​​or weight values ​​determined based on the corresponding operation behavior on the preset associated edge. For example, if a preset associated edge represents a "like" action, then the weight value corresponding to that preset associated edge should be higher than the weight value of a preset associated edge that only represents a "view" action. The weight values ​​of the first preset associated edges of the first node can be the same or different. These weight values ​​can be a ratio or a coefficient; this embodiment does not limit this.

[0208] Based on the current object category index data and weight information corresponding to the first traversed node, the transmitted data on the first preset associated edge is determined. Optionally, the current object category index data of the first node is multiplied by the weight values ​​of each of the first preset associated edges to obtain the transmitted data on each of the first preset associated edges. The transmitted data on each of the first preset associated edges may be the same or different, and the method for determining the transmitted data on each preset associated edge in this embodiment is not limited.

[0209] The data transmitted on the first preset associated edge connecting each first node is defined as the data transmitted on the preset associated edge. The data transmitted on the first preset associated edge of each first node is the data that all preset associated edges need to transmit to the second node during this forward transmission process.

[0210] In step S23a2, the transmitted data is transmitted to the second node through a preset associated edge.

[0211] In step S23a3, based on the transmitted data received by the second node, updated interest index data of historical multimedia resources are obtained.

[0212] Traverse the second node in the target association graph.

[0213] When traversing to any second node, obtain the second transmitted data passed to the currently traversed second node. The transmitted data is passed to the second node through a preset associated edge, and each second node will receive the transmitted data from the second preset associated edge connected to the second node.

[0214] Based on the second transmission data, updated interest index data corresponding to the currently traversed second node is generated. In one possible implementation, the process of generating the updated interest index data includes: determining the second cumulative transmission data on the currently traversed second node based on the second transmission data; determining the number of second preset associated edges connected to the currently traversed second node; and determining the updated interest index data corresponding to the currently traversed second node based on the second cumulative transmission data and the number of second preset associated edges. The second cumulative transmission data can be determined by summing the transmission data of each second preset associated edge. The second cumulative transmission data can be a sum of data, and the ratio of this sum to the number of second preset associated edges can be used as the updated interest index data.

[0215] The updated interest index data obtained during the traversal process will be used as the updated interest index data for historical multimedia resources.

[0216] In step S23b, the updated interest index data is used as the current interest index data.

[0217] In step S23c, it is determined whether the number of forward passes is greater than or equal to the first preset number. If yes, then step S23d is executed; otherwise, the process jumps to step S23h.

[0218] The aforementioned first preset number of iterations can be determined based on the transmission rate; for example, the convergence check can be performed after the first preset number of iterations. The aforementioned first preset number of iterations can be a positive integer greater than or equal to 2.

[0219] In step S23d, the first difference data between the current interest index data and the updated interest index data obtained in the previous forward pass is determined.

[0220] The above interest index difference data includes the first difference data.

[0221] If the number of forward passes is greater than or equal to a first preset number, a first difference data is determined between the current interest indicator data and the updated interest indicator data obtained from the previous forward pass. This first difference data can be the data difference between the current interest indicator data and the previous interest indicator data.

[0222] In step S23e, it is determined whether the first difference data meets the first preset condition. If yes, then step S23f is executed; if no, then step S23h is executed.

[0223] The first preset condition can be that the first difference data is less than the first threshold. The first difference data satisfies the first preset condition if the first difference data of each second node is less than the first threshold.

[0224] According to steps S23c and S23e above, if the number of forward transmissions is less than the first preset number or the first difference data does not meet the first preset condition, the process jumps to the step of transmitting the current interest index data in reverse through the preset association edge to obtain the updated object category index data.

[0225] In step S23f, the current interest index data is used as the target interest index data.

[0226] If the first difference data meets the first preset condition, the current interest index data is used as the target interest index data. Meeting the first preset condition indicates that the data change of each second node is less than a certain threshold, signifying that the scores on the second node side have converged. Therefore, the current interest index data can be used as the target interest index data. It is important to note that the aforementioned current interest index data is not necessarily the final target interest index data. This is because when the scores on the second node side converge, the scores on the first node side may not necessarily converge. Subsequent steps S23g are needed to determine whether the scores on the first node side have converged, and then corresponding steps are executed until the scores on both sides of the node converge.

[0227] In one example, such as Figure 6 As shown, Figure 6 This is a schematic diagram illustrating a target association graph in a bidirectional data transmission state according to an exemplary embodiment. Figure 6 In the process, the initial scores corresponding to each first node, i.e., the object category index data, are processed through... Figure 6 The pre-defined associated edges are passed to the second node. According to... Figure 6 As can be seen from the node colors of each second node, second nodes 321 and 324 have higher scores, and the multimedia resources they represent can be regarded as multimedia resources that the preset objects of the first object category prefer; second nodes 322 and 326 have moderate scores, and the multimedia resources they represent can be regarded as multimedia resources that the preset objects of the first object category may prefer; second nodes 323 and 325 have lower scores, and the multimedia resources they represent can be regarded as multimedia resources that the preset objects of the first object category do not prefer.

[0228] In step S23g, it is determined whether the second difference data meets the second preset condition. If not, step S23h is executed; if yes, step S23m is executed.

[0229] The second difference data is the difference between the current object category indicator data and the updated object category indicator data obtained from the previous reverse pass. The aforementioned category indicator difference data includes the second difference data.

[0230] The aforementioned second preset condition can be that the second difference data is less than the second threshold. The second difference data can satisfy the second preset condition if the second difference data of each first node is less than the second threshold.

[0231] If the second difference data meets the second preset condition, it means that the score on the first node side has also converged. Then, after executing step S23m, the target object category index data and target interest index data are directly obtained.

[0232] If the second difference data does not meet the second preset condition, it means that although the score on the second node side has reached convergence, the score on the first node side has not reached convergence. Further data transfer is required through subsequent steps until the scores on both nodes converge.

[0233] In step S23h, the current interest index data is transmitted in reverse through a preset association edge to obtain the updated object category index data.

[0234] If the second difference data does not meet the second preset condition, the current interest index data is passed back through a preset association edge to obtain updated object category index data. That is, if the node score on the first node side has not yet converged, the back propagation is performed to update the object category index data of the first node.

[0235] The reverse transmission process described above is similar to the forward transmission process, except that the data being transmitted and the direction of transmission have changed. For a detailed explanation, please refer to the following embodiments.

[0236] In an exemplary embodiment, such as Figure 5 As shown, the implementation process of step S23h includes the following steps (S23h1~S23h3).

[0237] In step S23h1, the transmitted data on the preset associated edges is determined based on the current interest index data.

[0238] Traverse the second node in the target association graph.

[0239] When traversing to any second node, determine the second preset associated edge connecting it to the currently traversed second node. The aforementioned second preset associated edge refers to a preset associated edge that has a connection relationship with the second node. If a second node does not have a preset associated edge connecting it, then that second node does not have the aforementioned second preset associated edge, and traversal can proceed to the next second node.

[0240] The weight information of the second preset associated edges is determined. This weight information includes the weight values ​​of each second preset associated edge connected to the currently traversed second node. These weight values ​​can be preset weight values ​​or weight values ​​determined based on the corresponding operation behavior on the preset associated edge. For example, if a preset associated edge represents a "like" action, then the weight value corresponding to that preset associated edge should be higher than the weight value of a preset associated edge that only represents a "view" action. The weight values ​​of the second preset associated edges of the second node can be the same or different. These weight values ​​can be a ratio or a coefficient; this embodiment does not limit this.

[0241] Based on the current interest index data and weight information corresponding to the currently traversed second node, the transmitted data on the second preset association edge is determined. Optionally, the current interest index data of the second node is multiplied by the weight values ​​of each second preset association edge to obtain the transmitted data on each second preset association edge. The transmitted data on each second preset association edge can be the same or different, and the method for determining the transmitted data on each preset association edge in this embodiment is not limited.

[0242] The data transmitted on the second preset associated edge connected to each second node is defined as the data transmitted on the preset associated edge. The data transmitted on the second preset associated edge of each second node is the data that all preset associated edges need to transmit to the second node during this reverse transmission process.

[0243] In step S23h2, the transmitted data is transmitted to the first node through a preset associated edge.

[0244] In step S23h3, updated object category index data is obtained based on the transmitted data received by the first node.

[0245] Traverse the first node in the target association graph.

[0246] When traversing to any first node, obtain the first transmitted data passed to the currently traversed first node. The transmitted data is passed to the first node through a preset associated edge, and each first node will receive the transmitted data from the first preset associated edge connected to it.

[0247] Based on the first transmitted data, updated object category index data corresponding to the currently traversed first node is generated. In one possible implementation, the process of generating the updated object category index data includes: determining the first cumulative transmitted data on the currently traversed first node based on the first transmitted data; determining the number of first preset associated edges connected to the currently traversed first node; and determining the updated object category index data corresponding to the currently traversed first node based on the first cumulative transmitted data and the number of first preset associated edges. The first cumulative transmitted data can be determined by summing the transmitted data of each first preset associated edge. The first cumulative transmitted data can be a sum of data, and the ratio of this sum to the number of first preset associated edges can be used as the updated object category index data.

[0248] The updated object category index data obtained during the traversal process will be used as the updated object category index data for multiple preset objects.

[0249] In step S23i, the updated object category index data is used as the current object category index data.

[0250] In step S23j, it is determined whether the number of reverse transmissions is greater than or equal to the second preset number. If yes, then step S23k is executed; otherwise, the process jumps to step S23a.

[0251] The aforementioned second preset number of iterations can be determined based on the transmission rate; for example, the convergence check can be performed after the second preset number of iterations. The aforementioned second preset number of iterations can be a positive integer greater than or equal to 2.

[0252] In step S23k, the third difference data between the current object category index data and the updated object category index data obtained from the previous reverse pass is determined.

[0253] If the number of reverse propagations is greater than or equal to the second preset number, a third difference data is determined between the current object category index data and the updated object category index data obtained from the previous reverse propagation. The aforementioned category index difference data includes the third difference data.

[0254] The third difference data and the second difference data mentioned above are both difference data between the current object category index data and the updated object category index data obtained from the previous reverse pass. However, the third difference data and the second difference data do not correspond to the difference data before and after the same reverse pass.

[0255] In step S231, it is determined whether the third difference data meets the second preset condition. If yes, then step S23m is executed; otherwise, the process jumps to step S23a.

[0256] If the third difference data satisfies the second preset condition, it indicates that the score on the first node side has also converged. Since the score on the first node side has converged, after executing step S23m, the target object category index data and target interest index data are directly obtained. The above process ensures that, when the first difference data satisfies the first preset condition and the third difference data satisfies the second preset condition, the current interest index data is used as the target interest index data, and the current object category index data is used as the target object category index data. In the aforementioned step S23e, a judgment was made on whether the first difference data satisfies the first preset condition. If the judgment result is yes, step S23f is executed, and the current interest index data is used as the target interest index data. Then, the process continues to step S23l. If the third difference data satisfies the second preset condition, the following step S23m can be executed, and the current object category index data is used as the target object category index data. Therefore, only when the first difference data satisfies the first preset condition and the third difference data also satisfies the second preset condition can the current interest index data be used as the final target interest index data, and the current object category index data be used as the final target object category index data. The aforementioned preset difference conditions include, but are not limited to, the first preset condition, the second preset condition, and the third preset condition.

[0257] If the third difference data does not meet the second preset condition, it means that although the score on the second node side has reached convergence, the score on the first node side has not reached convergence. Further data transfer is required through subsequent steps until the scores on both nodes converge.

[0258] According to steps S23j and S23l above, if the number of reverse transmissions is less than the second preset number or the third difference data does not meet the second preset condition, the process jumps to the step of forward transmission of the current object category index data through the preset association edge to obtain the updated interest index data of historical multimedia resources.

[0259] In step S23m, the current object category index data is used as the target object category index data.

[0260] If the third difference data meets the second preset condition, the current object category index data will be used as the target object category index data.

[0261] The following section will further explain the bidirectional transmission process of indicator data implemented in steps S23a to S23m, based on the specific transmission sequence.

[0262] The initial object category index data is passed forward for the first time through a preset association edge to obtain the second interest index data of the historical multimedia resources. The forward transmission direction is from the first node side to the second node side.

[0263] The second interest index data of historical multimedia resources is transmitted in reverse for the first time through a preset association edge to obtain the second object category index data of multiple preset objects. The transmission direction of the reverse transmission is from the second node side to the first node side.

[0264] The i-th object category index data of multiple preset objects is passed forward for the i-th time through preset association edges to obtain the (i+1)-th interest index data of historical multimedia resources, where i is a positive integer greater than or equal to 2.

[0265] If the first difference between the (i+1)th interest indicator data and the ith interest indicator data meets the first preset condition, and the second difference between the ith object category indicator data and the (i-1)th object category indicator data meets the second preset condition, then the ith object category indicator data is taken as the target object category indicator data, and the (i+1)th interest indicator data is taken as the target interest indicator data.

[0266] If the first difference data between the (i+1)th interest indicator data and the ith interest indicator data does not meet the first preset condition, or the second difference data between the ith object category indicator data and the (i-1)th object category indicator data does not meet the second preset condition, then the (i+1)th interest indicator data of the historical multimedia resources will be reversed for the ith time through the preset association edge to obtain the (i+1)th object category indicator data of multiple preset objects.

[0267] If the third difference data between the (i+1)th object category indicator data and the i-th object category indicator data meets the second preset condition, and the first difference data between the (i+1)th interest indicator data and the i-th interest indicator data meets the first preset condition, then the (i+1)th object category indicator data is used as the target object category indicator data, and the (i+1)th interest indicator data is used as the target interest indicator data.

[0268] If the third difference data between the (i+1)th object category index data and the ith object category index data does not meet the second preset condition, or the first difference data between the (i+1)th interest index data and the ith interest index data does not meet the first preset condition, then i is incremented by 1, and execution begins from the (i+1)th interest index data of historical multimedia resources obtained by passing the ith object category index data of multiple preset objects through preset association edges in the i-th forward transmission.

[0269] In step S240, the first object set is filtered based on the target object category index data of the preset objects in the first object set to obtain the target objects.

[0270] The first object set refers to the set of preset objects belonging to the first object category among multiple preset objects. The above-mentioned abnormal interest filtering refers to the process of filtering out preset objects in the first object set whose interests differ significantly from other preset objects.

[0271] In one example, such as Figure 7 As shown, Figure 7 This is a schematic diagram illustrating a target association diagram in an end state according to an exemplary embodiment. Figure 3 Compared to the node colors of the first nodes in the initial state shown, Figure 7 The lighter color of the first node 313 indicates a decrease in its score. Clearly, its score differs significantly from that of the first nodes 311 and 312, classifying it as an anomaly in the first object set of the first object category, requiring its removal from the first object set. Conversely, the darker color of the first node 314 indicates it should be added to the first node set, resulting in a new first node set. Furthermore, the darker colors of the second nodes 321 and 324 suggest higher scores, making them suitable as target objects represented by the new first node set.

[0272] In an exemplary embodiment, such as Figure 4 As shown, the implementation process of step S240 above includes the following steps (S241 to S242).

[0273] In step S241, preset objects in the first object set whose target object category index data do not meet the preset index conditions are identified as abnormal objects.

[0274] The preset indicator conditions are the object category indicator constraints corresponding to the first object category. Optionally, the target object category indicator data is a score representing the object category. Accordingly, the preset indicator conditions can be that the score is within the score range corresponding to the target object category, or that the score is greater than a preset score threshold.

[0275] In step S242, the abnormal object is removed from the first object set to obtain the target object.

[0276] The abnormal object is removed from the first object set, resulting in an updated first object set. The target object mentioned above is the preset object in the updated first object set.

[0277] In one possible implementation, preset objects whose target object category index data meets preset index conditions from preset objects outside the first object set are moved into the first object set to obtain an updated first object set.

[0278] In step S250, based on the target interest index data, the associated multimedia resources of the first object category are determined in the historical multimedia resources.

[0279] In an exemplary embodiment, such as Figure 4 As shown, step S250 above can be replaced by step S251 below.

[0280] In step S251, multimedia resources whose target interest index data meets the first interest index condition are identified as associated multimedia resources.

[0281] The first interest indicator condition is the interest indicator constraint condition corresponding to the first object category. Optionally, the target interest indicator data is a score representing the interest of a preset object to multimedia resources. Accordingly, the first interest indicator condition can be that the interest score is within the interest score range corresponding to the first object category, or the interest score is greater than a preset interest score threshold.

[0282] In step S260, associated multimedia resources are pushed to the target object.

[0283] In an exemplary embodiment, such as Figure 4 As shown, the above method also includes the following steps (S270~S290).

[0284] In step S270, a second object category corresponding to the abnormal object is determined based on the target object category index data of the abnormal object.

[0285] If the target object category index data of the abnormal object meets the category index conditions corresponding to the second object category, then the object category of the abnormal object is redefined as the second object category. Optionally, the target object category index data is a score representing the object category. Accordingly, the category index conditions corresponding to the second object category can be that the score is within the score range corresponding to the second object category, or that the score is greater than a preset score threshold corresponding to the second object category.

[0286] In step S280, based on the target interest index data, the associated multimedia resources of the second object category are determined in the historical multimedia resources.

[0287] Multimedia resources in the historical multimedia resources whose target interest index data meets the conditions of the second interest index are identified as associated multimedia resources of the second object category.

[0288] The second interest indicator condition is the interest indicator constraint condition corresponding to the second object category. Optionally, the target interest indicator data is a score representing the interest of a preset object in multimedia resources. Accordingly, the above-mentioned second interest indicator condition can be that the interest score is within the interest score range corresponding to the second object category, or the interest score is greater than a preset interest score threshold.

[0289] In step S290, the associated multimedia resources of the second object category are pushed to the abnormal object.

[0290] In summary, the technical solution provided in this application obtains a preset association edge between a first node representing a preset object and a second node representing historical multimedia resources in a target association graph. Based on the initial object category index of the preset object, bidirectional transmission of index data is performed between nodes with preset association edges in the target association graph to obtain object category index data of the preset object and interest index data of the preset object on historical multimedia resources. Then, the preset object set is filtered according to the object category index data to redetermine the target users belonging to the object category. Finally, the multimedia resources associated with the object category determined based on the interest index data are pushed to the target users. This can filter out objects with abnormal interests in the preset object set. While ensuring the consistency of the interests of each object in the preset object set, it accurately identifies the multimedia resources preferred by each type of object and pushes them accordingly, improving the efficiency of multimedia resource push and reducing the operating pressure of loading multimedia resources on the server and terminal.

[0291] Figure 8 This is a block diagram illustrating a multimedia resource delivery device according to an exemplary embodiment. (Refer to...) Figure 8 The device 800 includes: a relational graph acquisition module 810, an indicator acquisition module 820, an indicator bidirectional transmission module 830, an interest filtering module 840, a resource determination module 850, and a resource push module 860.

[0292] The association graph acquisition module 810 is configured to acquire a target association graph, which includes a first node and a second node connected by preset association edges. The first node represents multiple preset objects, and the second node represents historical multimedia resources that the multiple preset objects have acted upon. The preset association edges are generated based on the historical behavior data of the multiple preset objects.

[0293] The indicator acquisition module 820 is configured to acquire initial object category indicator data for multiple preset objects;

[0294] The bidirectional indicator transmission module 830 is configured to perform bidirectional transmission of indicator data between nodes with the preset associated edges based on the initial object category indicator, so as to obtain target object category indicator data of multiple preset objects and target interest indicator data of the historical multimedia resources.

[0295] Interest filtering module 840 is configured to filter the first object set based on target object category index data of preset objects in the first object set to obtain target objects. The first object set refers to the set of preset objects belonging to the first object category among multiple preset objects.

[0296] The resource determination module 850 is configured to determine the associated multimedia resources of the first object category in the historical multimedia resources based on the target interest index data.

[0297] The resource push module 860 is configured to push the associated multimedia resources to the target object.

[0298] In one possible design, the bidirectional indicator transmission module 830 includes:

[0299] The first difference data determination submodule is configured to perform forward transmission of object category index data of multiple preset objects through the preset association edge, update the interest index of the historical multimedia resources, and obtain the interest index difference data of the historical multimedia resources. The object category index data is the initial object category index data during the first forward transmission, and the interest index difference data refers to the difference data of the interest index of the historical multimedia resources before and after the forward transmission.

[0300] The second difference data determination submodule is configured to perform reverse transmission of the interest index data of the historical multimedia resources through the preset association edge to obtain the category index difference data of multiple preset objects. The category index difference data refers to the difference data of the object category index of the multiple preset objects before and after the reverse transmission.

[0301] The indicator data determination module is configured to obtain target object category indicator data of multiple preset objects and target interest indicator data of historical multimedia resources when both the interest indicator difference data and the category indicator difference data meet preset difference conditions.

[0302] In one possible design, the first difference data determining submodule includes:

[0303] The indicator forward transmission submodule is configured to perform forward transmission of the current object category indicator data through the preset association edge to obtain the updated interest indicator data of the historical multimedia resources. The current object category indicator is the initial object category indicator data during the first forward transmission, and the current object category indicator is the updated object category indicator data obtained from the previous reverse transmission during subsequent forward transmissions.

[0304] The interest metric update submodule is configured to use the updated interest metric data as the current interest metric data.

[0305] The difference data determination submodule is configured to determine the first difference data between the current interest index data and the updated interest index data obtained in the previous forward pass, provided that the number of forward passes is greater than or equal to a first preset number.

[0306] In one possible design, the second difference data determining submodule includes:

[0307] The indicator reverse transmission submodule is configured to, when the second difference data does not meet the second preset condition, reversely transmit the current interest indicator data through the preset association edge to obtain updated object category indicator data. The second difference data is the difference data between the current object category indicator data and the updated object category indicator data obtained from the previous reverse transmission.

[0308] The category index update submodule is configured to use the updated object category index data as the current object category index data;

[0309] The difference data determination submodule is further configured to determine the third difference data between the current object category index data and the updated object category index data obtained from the previous reverse pass, provided that the number of reverse passes is greater than or equal to the second preset number.

[0310] In one possible design, the indicator data determination module is specifically configured to perform:

[0311] If the first difference data satisfies the first preset condition and the third difference data satisfies the second preset condition, the current interest index data is used as the target interest index data, and the current object category index data is used as the target object category index data.

[0312] In one possible design, the indicator reverse transmission submodule is further configured to execute the step of reversing the current interest indicator data through the preset association edge to obtain updated object category indicator data if the number of forward transmissions is less than the first preset number or the first difference data does not meet the first preset condition.

[0313] The forward transmission submodule is further configured to, if the number of reverse transmissions is less than the second preset number or the third difference data does not meet the second preset condition, jump to the step of forward transmission of the current object category index data through the preset association edge to obtain the updated interest index data of the historical multimedia resources.

[0314] In one possible design, the indicator forward propagation submodule includes:

[0315] The data transmission determination unit is configured to determine the transmission data on the preset associated edge based on the current object category index data;

[0316] The data transmission unit is configured to transmit the transmitted data to the second node through the preset associated edge;

[0317] The interest index update unit is configured to execute the updated interest index data of the historical multimedia resources based on the transmitted data received by the second node.

[0318] In one possible design, the data transmission determination unit includes:

[0319] The node traversal sub-unit is configured to perform a traversal of the first node in the target association graph;

[0320] The associated edge determination sub-unit is configured to determine the first preset associated edge connected to the currently traversed first node when any first node is traversed.

[0321] The weight determination subunit is configured to determine the weight information of the first preset associated edge;

[0322] The data transmission determination subunit is configured to determine the transmission data on the first preset associated edge based on the current object category index data corresponding to the currently traversed first node and the weight information.

[0323] The data collection subunit is configured to execute the data transfer on the first preset associated edge connecting each first node, and to determine the data transfer on the preset associated edge.

[0324] In one possible design, the interest metric update unit includes:

[0325] The node traversal sub-unit is configured to perform a traversal of the second node in the target association graph;

[0326] The data acquisition subunit is configured to acquire the second data to be acquired when any second node is reached during traversal.

[0327] The interest index generation subunit is configured to generate updated interest index data corresponding to the currently traversed second node based on the second transmitted data.

[0328] The interest index collection subunit is configured to use the updated interest index data obtained during the traversal process as the updated interest index data for the historical multimedia resources.

[0329] In one possible design, the interest index generation subunit is specifically configured to perform the following actions: determining the second cumulative transmitted data on the currently traversed second node based on the second transmitted data; determining the number of second preset associated edges connected to the currently traversed second node; and determining the updated interest index data corresponding to the currently traversed second node based on the second cumulative transmitted data and the number of the second preset associated edges.

[0330] In one possible design, the indicator back-passing submodule includes:

[0331] The data transmission determination unit is configured to determine the data to be transmitted on the preset associated edge based on the current interest index data;

[0332] The data transmission unit is configured to transmit the transmitted data to the first node through the preset associated edge;

[0333] The category index update unit is configured to execute the updated object category index data based on the transmitted data received by the first node.

[0334] In one possible design, the data transmission determination unit includes:

[0335] The node traversal sub-unit is configured to perform a traversal of the second node in the target association graph;

[0336] The associated edge determination sub-unit is configured to determine a second preset associated edge connecting to the currently traversed second node when any second node is traversed.

[0337] The weight determination subunit is configured to determine the weight information of the second preset associated edge;

[0338] The data transmission determination subunit is configured to determine the data to be transmitted on the second preset associated edge based on the current interest index data corresponding to the currently traversed second node and the weight information.

[0339] The data collection subunit is configured to execute the transfer of data on the second preset associated edge connecting each second node, and to determine the transfer data on the preset associated edge.

[0340] In one possible design, the category index update unit includes:

[0341] The node traversal sub-unit is configured to perform a traversal of the first node in the target association graph;

[0342] The data acquisition subunit is configured to acquire the first data to be acquired when any first node is reached during traversal.

[0343] The category index generation subunit is configured to generate an updated object category index corresponding to the currently traversed first node based on the first transmitted data.

[0344] The category index collection subunit is configured to use the updated object category index data obtained during the traversal process as the updated object category index data for the plurality of preset objects.

[0345] In one possible design, the category index generation subunit is specifically configured to perform: based on the first transmitted data, determine the first cumulative transmitted data on the currently traversed first node; determine the number of first preset associated edges connected to the currently traversed first node; and based on the first cumulative transmitted data and the number of the first preset associated edges, determine the updated object category index data corresponding to the currently traversed first node.

[0346] In one possible design, the interest filtering module 840 includes:

[0347] The abnormal interest object determination unit is configured to determine the target object category index data in the first object set as an abnormal object if the data does not meet the preset index conditions. The preset index conditions are the object category index constraints corresponding to the first object category.

[0348] The abnormal interest object removal unit is configured to remove the abnormal object from the first object set to obtain the target object.

[0349] In one possible design, the resource determination module 850 is specifically configured to determine the multimedia resources in the historical multimedia resources whose target interest index data meets the first interest index condition as the associated multimedia resources, where the first interest index condition is the interest index constraint condition corresponding to the first object category.

[0350] In one possible design, the device further includes:

[0351] The object category determination module is configured to execute target object category index data based on the abnormal object to determine a second object category corresponding to the abnormal object;

[0352] The resource determination module 850 is further configured to perform the task of determining the associated multimedia resources of the second object category in the historical multimedia resources based on the target interest index data.

[0353] The resource push module 860 is also configured to push associated multimedia resources of the second object category to the abnormal object.

[0354] In one possible design, the device further includes:

[0355] The feature data acquisition module is configured to acquire object feature data of multiple preset objects;

[0356] The object classification module is configured to determine the object category corresponding to multiple preset objects based on the object feature data.

[0357] In one possible design, the indicator acquisition module 820 is specifically configured to determine initial object category indicator data for a plurality of preset objects based on the object categories corresponding to the plurality of preset objects.

[0358] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0359] In summary, the technical solution provided in this application obtains a preset association edge between a first node representing a preset object and a second node representing historical multimedia resources in a target association graph. Based on the initial object category index of the preset object, bidirectional transmission of index data is performed between nodes with preset association edges in the target association graph to obtain object category index data of the preset object and interest index data of the preset object on historical multimedia resources. Then, the preset object set is filtered according to the object category index data to redetermine the target users belonging to the object category. Finally, the multimedia resources associated with the object category determined based on the interest index data are pushed to the target users. This can filter out objects with abnormal interests in the preset object set. While ensuring the consistency of the interests of each object in the preset object set, it accurately identifies the multimedia resources preferred by each type of object and pushes them accordingly, improving the efficiency of multimedia resource push and reducing the operating pressure of loading multimedia resources on the server and terminal.

[0360] Figure 9 This is a block diagram illustrating an electronic device for pushing multimedia resources according to an exemplary embodiment. The electronic device may be a terminal, and its internal structure diagram may be as follows: Figure 9 As shown, the electronic device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a multimedia resource delivery method. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the device's casing, or an external keyboard, touchpad, or mouse.

[0361] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present disclosure and does not constitute a limitation on the electronic device to which the present disclosure is applied. A specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0362] In an exemplary embodiment, an electronic device is also provided, including: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the multimedia resource push method as described in the embodiments of this disclosure.

[0363] In an exemplary embodiment, a computer-readable storage medium is also provided, wherein when the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the multimedia resource push method of the present disclosure embodiments.

[0364] In an exemplary embodiment, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute the multimedia resource push method of the present disclosure embodiments.

[0365] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0366] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0367] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for pushing multimedia resources, characterized in that, include: Obtain a target association graph, which includes a first node and a second node connected by a preset association edge. The first node represents multiple preset objects, and the second node represents historical multimedia resources that the multiple preset objects have acted upon. The preset association edge is generated based on the historical behavior data of the multiple preset objects. Obtain initial object category index data for multiple preset objects; The object category index data of multiple preset objects are forward-transmitted through the preset association edge to update the interest index of the historical multimedia resources, thereby obtaining the interest index difference data of the historical multimedia resources. The object category index data is the initial object category index data during the first forward transmission, and the interest index difference data refers to the difference data of the interest index of the historical multimedia resources before and after the forward transmission. The interest index data of the historical multimedia resources are passed in reverse through the preset association edge to obtain the category index difference data of multiple preset objects. The category index difference data refers to the difference data of the object category index of multiple preset objects before and after the reverse transmission. When both the interest index difference data and the category index difference data meet the preset difference conditions, target object category index data of multiple preset objects and target interest index data of the historical multimedia resources are obtained. Based on the target object category index data of the preset objects in the first object set, the first object set is filtered to obtain the target objects. The first object set refers to the set of preset objects that belong to the first object category among the multiple preset objects. Based on the target interest index data, determine the associated multimedia resources of the first object category in the historical multimedia resources; The associated multimedia resources are pushed to the target object.

2. The multimedia resource push method according to claim 1, characterized in that, The step of forward propagating object category index data of multiple preset objects through the preset association edge to update the interest index of the historical multimedia resources and obtain the interest index difference data of the historical multimedia resources includes: The current object category index data is forward-transmitted through the preset association edge to obtain the updated interest index data of the historical multimedia resources. The current object category index is the initial object category index data during the first forward transmission, and the current object category index is the updated object category index data obtained from the previous reverse transmission during subsequent forward transmissions. Use the updated interest metric data as the current interest metric data; If the number of forward passes is greater than or equal to the first preset number, the first difference data between the current interest index data and the updated interest index data obtained from the previous forward pass is determined.

3. The multimedia resource push method according to claim 2, characterized in that, The step of passing the interest index data of the historical multimedia resources in reverse through the preset association edge to obtain the category index difference data of multiple preset objects includes: If the second difference data does not meet the second preset condition, the current interest index data is passed back through the preset association edge to obtain updated object category index data. The second difference data is the difference data between the current object category index data and the updated object category index data obtained from the previous back transmission. Use the updated object category index data as the current object category index data; If the number of reverse propagations is greater than or equal to the second preset number, a third difference data is determined between the current object category index data and the updated object category index data obtained from the previous reverse propagation.

4. The multimedia resource push method according to claim 3, characterized in that, When both the interest index difference data and the category index difference data meet preset difference conditions, the target object category index data of multiple preset objects and the target interest index data of the historical multimedia resources are obtained, including: If the first difference data satisfies the first preset condition and the third difference data satisfies the second preset condition, the current interest index data is used as the target interest index data, and the current object category index data is used as the target object category index data.

5. The multimedia resource push method according to claim 3, characterized in that, The method further includes: If the number of forward transmissions is less than the first preset number or the first difference data does not meet the first preset condition, the process jumps to the step of transmitting the current interest index data in reverse through the preset association edge to obtain the updated object category index data. If the number of reverse transmissions is less than the second preset number or the third difference data does not meet the second preset condition, the process jumps to the step of forward transmission of the current object category index data through the preset association edge to obtain the updated interest index data of the historical multimedia resources.

6. The multimedia resource push method according to claim 2, characterized in that, The step of forward propagating the current object category index data through the preset association edge to obtain the updated interest index data of the historical multimedia resources includes: Based on the current object category index data, determine the transmitted data on the preset associated edge; The transmitted data is transmitted to the second node via the preset associated edge; Based on the transmitted data received by the second node, updated interest index data of the historical multimedia resources are obtained.

7. The multimedia resource push method according to claim 6, characterized in that, The step of determining the transmitted data on the preset association edge based on the current object category index data includes: Traverse the first node in the target association graph; When traversing to any first node, determine the first preset associated edge that connects to the currently traversed first node; Determine the weight information of the first preset associated edge; Based on the current object category index data corresponding to the first traversed node and the weight information, determine the transmitted data on the first preset associated edge; The transmitted data on the first preset associated edge connected to each first node is determined as the transmitted data on the preset associated edge.

8. The multimedia resource push method according to claim 6, characterized in that, The process of obtaining updated interest index data for the historical multimedia resources based on the transmitted data received by the second node includes: Traverse the second node in the target association graph; When traversing to any second node, obtain the second data passed to the currently traversed second node; Based on the second transmitted data, the updated interest index data corresponding to the currently traversed second node is generated; The updated interest index data obtained during the traversal process will be used as the updated interest index data for the historical multimedia resources.

9. The multimedia resource push method according to claim 8, characterized in that, The step of generating updated interest index data corresponding to the currently traversed second node based on the second transmitted data includes: Based on the second transmitted data, determine the second cumulative transmitted data on the currently traversed second node; Determine the number of second preset associated edges connected to the currently traversed second node; Based on the number of the second cumulative transmitted data and the second preset associated edges, the updated interest index data corresponding to the currently traversed second node is determined.

10. The multimedia resource push method according to claim 3, characterized in that, The step of passing the current interest index data back through the preset association edge to obtain updated object category index data includes: Based on the current interest index data, determine the transmitted data on the preset associated edge; The transmitted data is transmitted to the first node via the preset associated edge; Based on the transmitted data received by the first node, the updated object category index data is obtained.

11. The multimedia resource push method according to claim 10, characterized in that, The step of determining the transmitted data on the preset association edge based on the current interest index data includes: Traverse the second node in the target association graph; When traversing to any second node, determine the second preset associated edge that connects to the currently traversed second node; Determine the weight information of the second preset associated edge; Based on the current interest index data corresponding to the currently traversed second node and the weight information, determine the transmitted data on the second preset associated edge; The transmitted data on the second preset associated edge connected to each second node is determined as the transmitted data on the preset associated edge.

12. The multimedia resource push method according to claim 10, characterized in that, The process of obtaining the updated object category index data based on the transmitted data received by the first node includes: Traverse the first node in the target association graph; When traversing to any first node, obtain the first passed data passed to the currently traversed first node; Based on the first transmitted data, an updated object category index corresponding to the first traversed node is generated; The updated object category index data obtained during the traversal process is used as the updated object category index data for the plurality of preset objects.

13. The multimedia resource push method according to claim 12, characterized in that, The step of generating the updated object category index corresponding to the currently traversed first node based on the first transmitted data includes: Based on the first transmitted data, determine the first cumulative transmitted data on the first node currently traversed; Determine the number of the first preset associated edges connected to the currently traversed first node; Based on the number of the first cumulative transmitted data and the first preset associated edges, the updated object category index data corresponding to the first traversed node is determined.

14. The multimedia resource push method according to any one of claims 1 to 13, characterized in that, The step of filtering the first object set based on target object category index data of preset objects in the first object set to obtain target objects includes: The objects in the first object set whose target object category index data do not meet the preset index conditions are identified as abnormal objects. The preset index conditions are the object category index constraints corresponding to the first object category. The abnormal object is removed from the first object set to obtain the target object.

15. The multimedia resource push method according to any one of claims 1 to 13, characterized in that, The step of determining the associated multimedia resources of the first object category in the historical multimedia resources based on the target interest index data includes: The multimedia resources in the historical multimedia resources whose target interest index data meets the first interest index condition are identified as the associated multimedia resources, where the first interest index condition is the interest index constraint condition corresponding to the first object category.

16. The multimedia resource push method according to claim 14, characterized in that, The method further includes: Based on the target object category index data of the abnormal object, determine the second object category corresponding to the abnormal object; Based on the target interest index data, determine the associated multimedia resources of the second object category in the historical multimedia resources; Push the associated multimedia resources of the second object category to the abnormal object.

17. The multimedia resource push method according to any one of claims 1 to 13, characterized in that, The method further includes: Obtain object feature data of multiple preset objects; Based on the object feature data, determine the object categories corresponding to the multiple preset objects.

18. The multimedia resource push method according to claim 17, characterized in that, The step of obtaining the initial object category index of the multiple preset objects includes: Based on the object categories corresponding to the multiple preset objects, determine the initial object category index data for the multiple preset objects.

19. A multimedia resource push device, characterized in that, include: The association graph acquisition module is configured to acquire a target association graph, which includes a first node and a second node connected by preset association edges. The first node represents multiple preset objects, and the second node represents historical multimedia resources that the multiple preset objects have acted upon. The preset association edges are generated based on the historical behavior data of the multiple preset objects. The indicator acquisition module is configured to acquire initial object category indicator data for multiple preset objects; The first difference data determination submodule is configured to perform forward transmission of object category index data of multiple preset objects through the preset association edge, update the interest index of the historical multimedia resources, and obtain the interest index difference data of the historical multimedia resources. The object category index data is the initial object category index data during the first forward transmission, and the interest index difference data refers to the difference data of the interest index of the historical multimedia resources before and after the forward transmission. The second difference data determination submodule is configured to perform reverse transmission of the interest index data of the historical multimedia resources through the preset association edge to obtain the category index difference data of multiple preset objects. The category index difference data refers to the difference data of the object category index of multiple preset objects before and after the reverse transmission. The indicator data determination module is configured to obtain target object category indicator data of multiple preset objects and target interest indicator data of historical multimedia resources when both the interest indicator difference data and the category indicator difference data meet preset difference conditions. The interest filtering module is configured to filter the first object set based on the target object category index data of preset objects in the first object set to obtain the target object. The first object set refers to the set of preset objects belonging to the first object category among multiple preset objects. The resource determination module is configured to determine the associated multimedia resources of the first object category in the historical multimedia resources based on the target interest index data. The resource push module is configured to push the associated multimedia resources to the target object.

20. The multimedia resource push device according to claim 19, characterized in that, The first difference data determination submodule includes: The indicator forward transmission submodule is configured to perform forward transmission of the current object category indicator data through the preset association edge to obtain the updated interest indicator data of the historical multimedia resources. The current object category indicator is the initial object category indicator data during the first forward transmission, and the current object category indicator is the updated object category indicator data obtained from the previous reverse transmission during subsequent forward transmissions. The interest metric update submodule is configured to use the updated interest metric data as the current interest metric data. The difference data determination submodule is configured to determine the first difference data between the current interest index data and the updated interest index data obtained in the previous forward pass, provided that the number of forward passes is greater than or equal to a first preset number.

21. The multimedia resource push device according to claim 20, characterized in that, The second difference data determination submodule includes: The indicator reverse transmission submodule is configured to, when the second difference data does not meet the second preset condition, reversely transmit the current interest indicator data through the preset association edge to obtain updated object category indicator data. The second difference data is the difference data between the current object category indicator data and the updated object category indicator data obtained from the previous reverse transmission. The category index update submodule is configured to use the updated object category index data as the current object category index data; The difference data determination submodule is further configured to determine the third difference data between the current object category index data and the updated object category index data obtained from the previous reverse pass, provided that the number of reverse passes is greater than or equal to the second preset number.

22. The multimedia resource push device according to claim 21, characterized in that, The indicator data determination module is specifically configured to execute: If the first difference data satisfies the first preset condition and the third difference data satisfies the second preset condition, the current interest index data is used as the target interest index data, and the current object category index data is used as the target object category index data.

23. The multimedia resource push device according to claim 21, characterized in that, The indicator reverse transmission submodule is further configured to execute the step of transmitting the current interest indicator data in reverse through the preset association edge to obtain the updated object category indicator data when the number of forward transmissions is less than the first preset number or the first difference data does not meet the first preset condition. The forward transmission submodule is further configured to, if the number of reverse transmissions is less than the second preset number or the third difference data does not meet the second preset condition, jump to the step of forward transmission of the current object category index data through the preset association edge to obtain the updated interest index data of the historical multimedia resources.

24. The multimedia resource push device according to claim 20, characterized in that, The positive propagation submodule of the indicator includes: The data transmission determination unit is configured to determine the transmission data on the preset associated edge based on the current object category index data; The data transmission unit is configured to transmit the transmitted data to the second node through the preset associated edge; The interest index update unit is configured to execute the updated interest index data of the historical multimedia resources based on the transmitted data received by the second node.

25. The multimedia resource push device according to claim 24, characterized in that, The data transmission determination unit includes: The node traversal sub-unit is configured to perform a traversal of the first node in the target association graph; The associated edge determination sub-unit is configured to determine the first preset associated edge connected to the currently traversed first node when any first node is traversed. The weight determination subunit is configured to determine the weight information of the first preset associated edge; The data transmission determination subunit is configured to determine the transmission data on the first preset associated edge based on the current object category index data corresponding to the currently traversed first node and the weight information. The data collection subunit is configured to execute the data transfer on the first preset associated edge connecting each first node, and to determine the data transfer on the preset associated edge.

26. The multimedia resource push device according to claim 24, characterized in that, The interest metric update unit includes: The node traversal sub-unit is configured to perform a traversal of the second node in the target association graph; The data acquisition subunit is configured to acquire the second data to be acquired when any second node is reached during traversal. The interest index generation subunit is configured to generate updated interest index data corresponding to the currently traversed second node based on the second transmitted data. The interest index collection subunit is configured to use the updated interest index data obtained during the traversal process as the updated interest index data for the historical multimedia resources.

27. The multimedia resource push device according to claim 26, characterized in that, The interest index generation subunit is specifically configured to perform the following: based on the second transmitted data, determine the second cumulative transmitted data on the currently traversed second node; Determine the number of second preset associated edges connected to the currently traversed second node; based on the second cumulative transmitted data and the number of second preset associated edges, determine the updated interest index data corresponding to the currently traversed second node.

28. The multimedia resource push device according to claim 21, characterized in that, The indicator reverse transmission submodule includes: The data transmission determination unit is configured to determine the data to be transmitted on the preset associated edge based on the current interest index data; The data transmission unit is configured to transmit the transmitted data to the first node via the preset associated edge; The category index update unit is configured to execute the updated object category index data based on the transmitted data received by the first node.

29. The multimedia resource push device according to claim 28, characterized in that, The data transmission determination unit includes: The node traversal sub-unit is configured to perform a traversal of the second node in the target association graph; The associated edge determination sub-unit is configured to determine a second preset associated edge connecting to the currently traversed second node when any second node is traversed. The weight determination subunit is configured to determine the weight information of the second preset associated edge; The data transmission determination subunit is configured to determine the data to be transmitted on the second preset associated edge based on the current interest index data corresponding to the currently traversed second node and the weight information. The data collection subunit is configured to execute the transfer of data on the second preset associated edge connecting each second node, and to determine the transfer data on the preset associated edge.

30. The multimedia resource push device according to claim 28, characterized in that, The category indicator update unit includes: The node traversal sub-unit is configured to perform a traversal of the first node in the target association graph; The data acquisition subunit is configured to acquire the first data to be acquired when any first node is reached during traversal. The category index generation subunit is configured to generate an updated object category index corresponding to the currently traversed first node based on the first transmitted data. The category index collection subunit is configured to use the updated object category index data obtained during the traversal process as the updated object category index data for the plurality of preset objects.

31. The multimedia resource push device according to claim 30, characterized in that, The category index generation subunit is specifically configured to perform: based on the first transmitted data, determine the first cumulative transmitted data on the currently traversed first node; Determine the number of first preset associated edges connected to the currently traversed first node; based on the first cumulative transmitted data and the number of the first preset associated edges, determine the updated object category index data corresponding to the currently traversed first node.

32. The multimedia resource push device according to any one of claims 19 to 31, characterized in that, The interest filtering module includes: The abnormal interest object determination unit is configured to determine the target object category index data in the first object set as an abnormal object if the data does not meet the preset index conditions. The preset index conditions are the object category index constraints corresponding to the first object category. The abnormal interest object removal unit is configured to remove the abnormal object from the first object set to obtain the target object.

33. The multimedia resource push device according to any one of claims 19 to 31, characterized in that, The resource determination module is specifically configured to determine the multimedia resources in the historical multimedia resources whose target interest index data meets the first interest index condition as the associated multimedia resources, where the first interest index condition is the interest index constraint condition corresponding to the first object category.

34. The multimedia resource push device according to claim 32, characterized in that, The device further includes: The object category determination module is configured to execute target object category index data based on the abnormal object to determine a second object category corresponding to the abnormal object; The resource determination module is further configured to determine the associated multimedia resources of the second object category in the historical multimedia resources based on the target interest index data. The resource push module is also configured to push associated multimedia resources of the second object category to the abnormal object.

35. The multimedia resource push device according to any one of claims 19 to 31, characterized in that, The device further includes: The feature data acquisition module is configured to acquire object feature data of multiple preset objects; The object classification module is configured to determine the object category corresponding to multiple preset objects based on the object feature data.

36. The multimedia resource push device according to claim 35, characterized in that, The indicator acquisition module is specifically configured to determine the initial object category indicator data of the multiple preset objects based on the object categories corresponding to the multiple preset objects.

37. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the multimedia resource push method as described in any one of claims 1 to 18.

38. A computer-readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor of the electronic device, the electronic device is able to perform the multimedia resource push method as described in any one of claims 1 to 18.

39. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the multimedia resource push method according to any one of claims 1 to 18.

Images