Content recommendation method, device and equipment, computer readable storage medium and product

Abstract

The embodiment of the invention provides a content recommendation method and device, equipment, a computer readable storage medium and a product. The method comprises the steps that target media content is acquired; performing content recall operation in a preset resource pool according to the target media content to obtain a plurality of recalled media contents; for each recalled media content, identifying an information increment of the recalled media content relative to a target media content through a preset information increment identification model; and performing content recommendation operation on the multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents. Since the information increment can accurately represent the content difference between the target media content and the recalled media content and the increment information, the content recommendation based on the information increment can effectively avoid recommending repeated and redundant content to the user, and can recommend the media content with higher quality and more content to the user.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of data processing technology, and in particular, to a content recommendation method, device, equipment, computer-readable storage medium, and product. Background Art

[0002] With the continuous development of technology, users can browse media content through application software preset in terminal devices.

[0003] In the related art, in order to enable users to browse more media content in application software, content recommendation operations can generally be performed to users in a targeted manner. Currently, the recommended content is generally recalled based on content similarity, and the recommendation operation is performed based on the recalled content.

[0004] However, when the above method is used for content recommendation, since the recalled content is often highly repetitive, the media content recommended to the user may be repetitive and redundant. Summary of the invention

[0005] The embodiments of the present disclosure provide a content recommendation method, apparatus, device, computer-readable storage medium, and product, which are used to solve the technical problem of redundant content recommendation in current content recommendation methods.

[0006] In a first aspect, an embodiment of the present disclosure provides a content recommendation method, including:

[0007] Acquire target media content;

[0008] Performing a content recall operation in a preset resource pool according to the target media content to obtain a plurality of recalled media contents;

[0009] For each recalled media content, identifying the information increment of the recalled media content relative to the target media content through a preset information increment identification model;

[0010] A content recommendation operation is performed on the multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents.

[0011] In a second aspect, an embodiment of the present disclosure provides a content recommendation device, including:

[0012] An acquisition module, used for acquiring target media content;

[0013] A recall module, configured to perform a content recall operation in a preset resource pool according to the target media content, and obtain a plurality of recalled media contents;

[0014] A determination module, configured to identify, for each recalled media content, an information increment of the recalled media content relative to the target media content through a preset information increment identification model;

[0015] The recommendation module is used to perform a content recommendation operation on the multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents.

[0016] In a third aspect, an embodiment of the present disclosure provides an electronic device, including: a processor and a memory;

[0017] The memory stores computer-executable instructions;

[0018] The processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the content recommendation method described in the first aspect and various possible designs of the first aspect.

[0019] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer execution instructions. When a processor executes the computer execution instructions, the content recommendation method described in the first aspect and various possible designs of the first aspect is implemented.

[0020] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including a computer program, which, when executed by a processor, implements the content recommendation method as described in the first aspect and various possible designs of the first aspect.

[0021] The content recommendation method, device, equipment, computer-readable storage medium and product provided in this embodiment can identify the information increment of each recalled media content relative to the target media content based on the preset information increment identification model after recalling multiple recalled media content based on the target media content, so that the recalled media content can be recommended based on the information increment. Since the information increment can accurately represent the content difference and incremental information between the target media content and the recalled media content, content recommendation based on the information increment can effectively avoid recommending repeated and redundant content to users, and can recommend better quality and more meaningful media content to users. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0023] Figure 1 A diagram of the system architecture on which the present disclosure is based;

[0024] Figure 2A flowchart of a content recommendation method provided by an embodiment of the present disclosure;

[0025] Figure 3 A schematic diagram of an application scenario provided for an embodiment of the present disclosure;

[0026] Figure 4 A flowchart of a content recommendation method provided by another embodiment of the present disclosure;

[0027] Figure 5 A schematic diagram of the structure of the information increment recognition model provided by the embodiment of the present disclosure;

[0028] Figure 6 A flowchart of a content recommendation method provided by another embodiment of the present disclosure;

[0029] Figure 7 A schematic diagram of the structure of a content recommendation device provided by an embodiment of the present disclosure;

[0030] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0032] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0033] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0034] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0035] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0036] In order to solve the technical problem of redundant content recommendation in current content recommendation methods, the present disclosure provides a content recommendation method, apparatus, device, computer-readable storage medium and product.

[0037] It should be noted that the content recommendation method, apparatus, device, computer-readable storage medium and product provided by the present disclosure can be applied in any content recommendation scenario.

[0038] Current content recommendation methods generally recall content based on the user's historical consumption content and directly perform content recommendation operations based on the recalled content. However, since the recall operation is generally based on features such as similarity when recalling data, the recalled content is often redundant and cumbersome, and therefore cannot meet the user's demand for fresh and novel content.

[0039] In the process of solving the above technical problems, the inventors found through research that in order to avoid pushing redundant and repeated recommended content to users, an information increment recognition model can be pre-trained. The information increment recognition model can accurately identify the information increment of the recalled media content relative to the target media content. The information increment can accurately characterize the content difference and incremental information between the target media content and the recalled media content, so that content recommendation based on the information increment can effectively avoid recommending repeated and redundant content to users, and can recommend better quality and more meaningful media content to users.

[0040] Figure 1 This is the system architecture diagram on which the present disclosure is based, such as Figure 1 As shown, the system architecture based on the present disclosure includes at least a server 11, a data server 12 and a terminal device 13. The server 11 can be connected to the data server 12 and the terminal device 13 in communication. The server 11 is provided with a content recommendation device, which can be written in languages ​​such as C / C++, Java, Shell or Python. The data server 12 can be a cloud server or a server cluster, which stores a resource pool for content recall. The terminal device 13 can be, for example, a desktop computer, a tablet computer, a mobile phone, etc.

[0041] Based on the above system architecture, after determining the target media content, the server 11 can recall multiple recalled media contents associated with the target media content from the data server 12. The server 11 can also identify the information increment of each recalled media content relative to the target media content based on the preset information increment identification model. The server 11 can then determine multiple target recalled media contents to be recommended based on the information increment and push them to the terminal device 13.

[0042] Figure 2 A flow chart of a content recommendation method provided by an embodiment of the present disclosure is shown in FIG. Figure 2 As shown, the method includes:

[0043] Step 201: Acquire target media content.

[0044] The execution subject of this embodiment is a content recommendation device. The content recommendation device can be coupled to a server. The server can be connected to the terminal device and the data server for communication.

[0045] In this embodiment, in order to implement the content recommendation operation, the target media content may be first obtained, wherein the target media content includes but is not limited to video media content, audio media content, graphic media content, text media content, etc., and the present disclosure does not limit this.

[0046] Optionally, the target media content may be selected by the user according to actual needs, or may be media content pre-stored in a preset storage path, which is not limited in the present disclosure.

[0047] Step 202: Perform a content recall operation in a preset resource pool according to the target media content to obtain a plurality of recalled media contents.

[0048] In this embodiment, after the target media content is acquired, a content recall operation can be performed in a preset resource pool based on the target media content to obtain multiple recalled media contents. A large amount of preset media contents are pre-stored in the resource pool. The preset media contents include but are not limited to video media contents, audio media contents, graphic media contents, text media contents, etc.

[0049] Optionally, the determination of the recalled media content may be implemented based on the similarity between the target media content and the preset media content.

[0050] Alternatively, a recall model may be pre-trained, and the recalled media content related to the target media content may be recalled from the resource pool based on the recall model, which is not limited in the present disclosure.

[0051] Step 203: For each recalled media content, identify the information increment of the recalled media content relative to the target media content through a preset information increment identification model.

[0052] In this embodiment, since the recalled media content is recalled based on the target media content, there is a certain correlation between the recalled media content and the target media content. In order to avoid pushing duplicate and redundant media content to the user, after recalling multiple recalled media content, the information increment of the recalled media content relative to the target media content can also be determined. Among them, the information increment can characterize the difference content, new content, etc. of the recalled content relative to the media content. Therefore, an information increment recognition model can be pre-trained, and the information increment model can accurately identify the information increment of the recalled media content relative to the target media content.

[0053] Therefore, for each recalled media content, the information increment of the recalled media content relative to the target media content may be identified by using a preset information increment identification model.

[0054] Step 204: Perform a content recommendation operation on the multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents.

[0055] In this embodiment, in order to recommend recommended content that is more in line with actual needs and has novel content to users, after determining the information increments of multiple recalled media content, content recommendation operations can be performed to users based on the information increments.

[0056] For example, the recalled media content having information increment relative to the target media content may be recommended to the user, or the recalled media content having information increment relative to the target media content meeting a preset condition may be recommended to the user.

[0057] As an implementable manner, for each recalled media content, parameters such as whether it is associated with a manually set label, its associated information increment, its associated publisher information, and its corresponding content volume can be determined respectively. Among them, the label can be added to the recalled media content by the publisher or other user when determining that the content quality of the recalled media content is high.

[0058] Therefore, the recommendation priorities corresponding to multiple recalled contents can be determined based on the above parameters. For example, the priority of recalled media contents associated with manually set tags> recalled media contents with information increment relative to target media contents> recalled media contents whose publisher information is in the preset publisher whitelist> recalled media contents whose content volume is lower than the preset threshold. Therefore, the content recommendation operation can be implemented according to the recommendation priority.

[0059] The content recommendation method provided in this embodiment, after recalling multiple recalled media contents based on the target media contents, identifies the information increment of each recalled media content relative to the target media contents based on the preset information increment identification model, so that the recalled media contents can be recommended based on the information increment. Since the information increment can accurately represent the content difference and increment information between the target media contents and the recalled media contents, content recommendation based on the information increment can effectively avoid recommending repeated and redundant contents to users, and can recommend better quality and more meaningful media contents to users.

[0060] Optionally, based on any of the above embodiments, step 202 includes:

[0061] The vector distances between the target media content and each preset media content in the resource pool are determined respectively.

[0062] The preset media content whose vector distance satisfies the preset recall condition is determined as the recalled media content.

[0063] In this embodiment, the identification and recall operation of the recalled media content can be realized based on the similarity between each preset media content in the resource pool and the target media content, so that the recalled multiple recalled media contents can better meet the actual needs of the user.

[0064] Optionally, the vector distance between the target media content and each preset media content in the resource pool can be determined respectively. The target media content and each preset media content in the resource pool can be vectorized respectively to obtain a first vector corresponding to the target media content and a second vector corresponding to the preset media content. The vector distance between the first vector and the second vector is calculated using a preset vector distance formula. Any vector distance calculation formula can be used to calculate the vector distance, and the present disclosure does not limit this.

[0065] Furthermore, the preset media content whose vector distance satisfies the preset recall condition is determined as the recalled media content. The recall condition may be that the vector distance is less than a preset distance threshold, which may be set by the user according to actual needs or may be a default experience value, and the present disclosure does not limit this.

[0066] The content recommendation method provided in this embodiment performs content recall operations based on the vector distance between the target media content and each preset media content in the resource pool, thereby being able to recall media content that is more similar to the target media content and more in line with the actual needs of the user, thereby improving the accuracy of content recall.

[0067] Optionally, based on any of the above embodiments, after step 202, the method further includes:

[0068] The publisher information associated with each of the recalled media contents is determined, and a preset publisher list is obtained.

[0069] For each of the recalled media contents, it is determined whether the publisher list includes publisher information associated with the recalled media content, and a processing result is obtained.

[0070] A content filtering operation is performed on the plurality of recalled media contents based on the processing result.

[0071] Alternatively, after step 202, the method further includes:

[0072] A content recognition operation is performed on the plurality of recalled media contents through a preset recognition model to determine the media contents to be filtered that are generated through an artificial intelligence algorithm.

[0073] A content filtering operation is performed on the media contents to be filtered among the plurality of recalled media contents.

[0074] In this embodiment, after a plurality of recalled media contents are recalled based on the target media contents, in order to improve the content quality of the recalled media contents recommended to the user, a preliminary filtering operation may be performed on the plurality of recalled media contents.

[0075] Optionally, a publisher list may be preset, wherein the publisher list may store identification information of publishers that the user is interested in. The publisher information associated with each recalled media content may be determined, and a preset publisher list may be obtained. For each recalled media content, it is determined whether the publisher list includes the publisher information associated with the recalled media content, and a processing result is obtained. Content filtering operations are performed on multiple recalled media contents based on the processing result. For example, if the processing result is that the publisher list does not include the publisher information associated with the recalled media content, it indicates that the recalled media content is not the media content published by the publisher that the user is interested in, and therefore the recalled media content may be filtered.

[0076] As an implementable method, in order to improve the user's browsing experience for recalled media content, the recalled media content generated by AI can also be filtered.

[0077] Optionally, a recognition model can be pre-trained, and the recognition model can accurately recognize the media content generated by AI. Therefore, after obtaining multiple recalled media contents, a content recognition operation can be performed on the multiple recalled media contents through a preset recognition model to determine the media content to be filtered generated by the artificial intelligence algorithm. Content filtering operations are performed on the media content to be filtered in the multiple recalled media contents.

[0078] Figure 3 A schematic diagram of an application scenario provided by an embodiment of the present disclosure, such as Figure 3As shown, after acquiring the target media content 31, a content recall operation can be performed according to the target media content 31 to obtain multiple recalled media content 32. For each recalled media content, determine whether the recalled media content 32 meets the preset screening condition 33. Among them, the screening condition 33 may be that the publisher of the recalled media content is in a preset whitelist, and the recalled media content is not AI-created media content. If satisfied, the recalled media content 32 is input into a preset information increment recognition model 34 to determine its information increment relative to the target media content 31. Otherwise, the recalled media content 32 can be discarded. After determining the information increment of each recalled media content that meets the preset screening condition, a content recommendation operation 35 can be performed based on the information increment.

[0079] The content recommendation method provided in this embodiment can effectively screen high-quality content in the recalled media content by preliminarily screening the recalled media content based on preset screening conditions after recalling multiple recalled media content, thereby improving the quality of recommended content.

[0080] Figure 4 This is a flowchart of a content recommendation method provided by another embodiment of the present disclosure. Based on any of the above embodiments, the information increment recognition model is a three-tower extension model, including a multimodal feature extraction module, a feature splicing module, and a classification module. Figure 4 As shown, step 203 includes:

[0081] Step 401: input the target media content, the recalled media content and the associated data into the multimodal feature extraction module to obtain a first feature vector corresponding to the target media content, a second feature vector corresponding to the recalled media content and a third feature vector corresponding to the associated data, wherein the associated data includes parameter information and interaction information associated with the target media content and the recalled media content.

[0082] Step 402: Perform a vector splicing operation on the first feature vector, the second feature vector and the third feature vector through the splicing module to obtain a target feature vector.

[0083] Step 403: input the target feature vector into the classification module to obtain a classification result output by the classification module, wherein the classification result includes the probability of different preset information increments of the recalled media content relative to the target media content.

[0084] Step 404: Use the preset information increment in the classification result that has a probability of satisfying a preset condition as the information increment of the recalled media content relative to the target media content.

[0085] In this embodiment, the information increment recognition model can be a three-tower extension model, which includes a multimodal feature extraction module, a feature splicing module, and a classification module. Among them, the multimodal feature extraction module can include a first extraction unit, a second extraction unit, and a third extraction unit.

[0086] Based on the above model architecture, after obtaining multiple recalled media contents, for each recalled media content, the target media content, the recalled media content and the associated data can be input into the multimodal feature extraction module to obtain a first feature vector corresponding to the target media content, a second feature vector corresponding to the recalled media content and a third feature vector corresponding to the associated data, wherein the associated data includes parameter information and interaction information associated with the target media content and the recalled media content. Taking the target media content as video media content as an example, the associated data may include comment content, like information, collection information, publisher information, and video parameters associated with the target media content; as well as comment content, like information, collection information, publisher information, and video parameters associated with the recalled media content. The associated data may be any type of data related to the target media content and the recalled media content, and the present disclosure does not limit this.

[0087] Furthermore, the first feature vector, the second feature vector, and the third feature vector can be subjected to a vector splicing operation through a splicing module to obtain a target feature vector. The target feature vector is input into a classification module to obtain a classification result output by the classification module, wherein the classification result includes the probabilities of different preset information increments of the recalled media content relative to the target media content. The preset information increments in the classification results whose probabilities meet preset conditions are used as the information increments of the recalled media content relative to the target media content. For example, the preset information increment with the highest probability can be determined as the information increment of the recalled media content relative to the target media content.

[0088] Figure 5 A schematic diagram of the structure of the information increment recognition model provided by the embodiment of the present disclosure, such as Figure 5As shown, the information increment identification model 51 can be a three-tower extension model, which includes a multimodal feature extraction module 52, a feature splicing module 53 and a classification module 54. The multimodal feature extraction module 52 includes a first extraction unit 55, a second extraction unit 56 and a third extraction unit 57. The extraction unit can specifically be a DNN model. The multimodal feature extraction module 52 can perform feature extraction on the multimodal input content to obtain a first feature vector, a second feature vector and a third feature vector. The feature splicing module 53 can perform a vector splicing operation on the first feature vector, the second feature vector and the third feature vector to obtain a target feature vector. The classification module 54 can be a DNN Classifier, which can perform a classification operation on the target feature vector to determine the information increment of the recalled media content relative to the target media content.

[0089] The content recommendation method provided in this embodiment presets an information increment recognition model, thereby comprehensively considering the multimodal feature vectors associated with target media content, recalled media content, and associated data to identify information increments, thereby improving the accuracy of information increment recognition, and further being able to recommend more diverse media content to users based on the accurate information increments.

[0090] Further, based on any of the above embodiments, before step 203, the method further includes:

[0091] Obtain a preset training data set, wherein the training data set includes multiple sets of training data pairs, the training data pairs include target content, recalled content based on the target content, associated data corresponding to the target content and the recalled content, and preset annotation information, and the annotation information is the information increment of the recalled content relative to the target content.

[0092] The target content, the recalled content, and the associated data in the training data pair are input into a preset model to be trained to obtain the predicted information increment output by the model to be trained.

[0093] The loss value corresponding to the model to be trained is determined based on the prediction information increment and the annotation information.

[0094] Determine whether the model to be trained meets a preset convergence condition based on the loss value.

[0095] If it is satisfied, the current model to be trained is determined as the information increment recognition model.

[0096] If it is not satisfied, the parameters of the model to be trained are adjusted based on the loss value, and the step of inputting the target content, recall content, and related data in the training data pair into the preset model to be trained is returned to obtain the predicted information increment output by the model to be trained, until the model to be trained meets the preset convergence conditions, and the information increment recognition model is obtained.

[0097] In this embodiment, in order to identify the information increment of the recalled media content relative to the target media content based on the information increment identification model, it is first necessary to train the information increment identification model.

[0098] Optionally, a preset training data set can be obtained, wherein the training data set includes multiple sets of training data pairs, the training data pairs include target content, recalled content based on the target content, associated data corresponding to the target content and the recalled content, and preset annotation information, and the annotation information is the information increment of the recalled content relative to the target content.

[0099] Furthermore, the target content, recalled content, and related data in the training data pair can be input into a preset model to be trained. The model to be trained can extract multimodal features based on the input content, perform splicing operations on the extracted multimodal features, and perform classification based on the splicing results to obtain the predicted information increment of the recalled content output by the model to be trained relative to the target content.

[0100] Furthermore, the loss value corresponding to the model to be trained can be determined based on the prediction information increment and the annotation information. Among them, any loss calculation formula can be used to calculate the loss value, and the present disclosure does not limit this. Determine whether the model to be trained meets the preset convergence condition based on the loss value. The convergence condition includes but is not limited to the loss value being lower than a preset threshold, or the difference in loss values ​​obtained from two rounds of training is lower than a preset threshold, etc. Alternatively, the convergence condition can also be that the training duration reaches a preset duration threshold, or the number of training rounds reaches a preset number of thresholds, etc. Users can set the convergence conditions according to actual needs, and the present disclosure does not limit this.

[0101] Optionally, if the model to be trained meets the preset convergence condition, it is determined that the current model to be trained has been trained, and the current model to be trained can be determined as the information increment recognition model. On the contrary, if it does not meet the condition, the parameters of the model to be trained are adjusted based on the loss value, and the step of inputting the target content, recall content, and associated data in the training data pair into the preset model to be trained is returned to obtain the predicted information increment output by the model to be trained, until the model to be trained meets the preset convergence condition, and the information increment recognition model is obtained.

[0102] The content recommendation method provided in this embodiment iteratively trains the training model based on the training data set in advance, so that the trained incremental recognition model can accurately recognize the information increment of the recalled content relative to the target content based on the target content, recalled content, and related data, thereby improving the recognition accuracy of the model.

[0103] Further, based on any of the above embodiments, step 204 includes:

[0104] Determine a plurality of target recall contents whose information increment is greater than a preset increment threshold.

[0105] A content recommendation operation is performed according to the multiple target recalled contents.

[0106] In this embodiment, in order to recommend recommended content that is more in line with actual needs and has novel content to users, after determining the information increments of multiple recalled media content, content recommendation operations can be performed to users based on the information increments.

[0107] Optionally, since the information increment can characterize the difference content, new content, etc. of the recalled content relative to the media content, in order to avoid recommending redundant and repeated media content to the user, multiple target recalled contents whose information increment is greater than a preset increment threshold can be determined based on the information increment corresponding to each recalled media content. The preset increment threshold can be set by the user according to actual needs, or can be a default experience value, which is not limited in the present disclosure.

[0108] Furthermore, after determining multiple target recall contents whose information increment is greater than a preset increment threshold, content recommendation operations can be performed based on the multiple target recall contents, thereby making the recommended media contents more in line with the user's browsing preferences while avoiding recommending duplicate and redundant contents to the user.

[0109] Further, based on any of the above embodiments, step 204 includes:

[0110] The recommendation order corresponding to the plurality of recalled media contents is determined according to the order of information increment from large to small.

[0111] A content recommendation operation is performed on the plurality of recalled media contents according to the recommendation order.

[0112] In this embodiment, in order to recommend recommended content that is more in line with actual needs and has novel content to users, after determining the information increments of multiple recalled media content, content recommendation operations can be performed to users based on the information increments.

[0113] Optionally, the recommendation order corresponding to the plurality of recalled media contents may be determined in descending order of information increments. Content recommendation operations are performed on the plurality of recalled media contents in the recommendation order. Thus, the user can browse the recalled media contents that are of interest and novel in a more prominent position.

[0114] Alternatively, as an implementable manner, after determining the recommendation order corresponding to a plurality of recalled media contents, the first N recalled media contents may be selected for content recommendation operation.

[0115] The content recommendation method provided in this embodiment makes recommendations on recalled media content based on information increments, thereby making the recommended media content more in line with the user's browsing preferences while avoiding recommending duplicate and redundant content to the user, thereby improving the quality and accuracy of content recommendations.

[0116] Figure 6 A flowchart of a content recommendation method provided by another embodiment of the present disclosure is provided. Based on any of the above embodiments, Figure 6 As shown, the method also includes:

[0117] Step 601: Display multiple media contents on a preset media content display interface.

[0118] Step 602: In response to a trigger operation of a user on any media content, a plurality of recalled contents associated with the media content are recalled from the resource pool based on a preset recall algorithm.

[0119] Step 603: For each recalled content, determine the information increment of the recalled content relative to the media content.

[0120] Step 604: Determine a display order corresponding to the multiple recalled contents based on the information increments corresponding to the multiple recalled contents.

[0121] Step 605: Display the multiple recalled contents in the media content display interface according to the display order.

[0122] In this embodiment, the user can browse the media content stream in the media content display interface, wherein the media content includes but is not limited to text media content, graphic media content, video media content, etc. The user can trigger any media content according to actual needs to achieve a further understanding of the media content.

[0123] Accordingly, in order to make the media content displayed in the media content display interface more in line with the actual needs of the user, in response to the user's trigger operation on any media content, multiple recalled contents associated with the media content are recalled in the resource pool based on a preset recall algorithm. Among them, the content recall operation can be implemented based on the similarity between the media content and the pre-stored media content in the resource pool through the recall algorithm. Alternatively, any algorithm capable of implementing content recall can be used to implement the recall operation of the recalled content, and the present disclosure does not limit this.

[0124] Furthermore, in order to recommend more practical and novel content to users, after recalling multiple media content, the information increment of the recalled content relative to the media content can also be determined. The information increment can represent the difference content, new content, etc. of the recalled content relative to the media content.

[0125] Therefore, after determining the information increment of the recalled content, a content recommendation operation can be performed to the user based on the information increment. For example, recalled content whose information increment meets a preset condition can be recommended to the user, or recalled content that has information increment relative to the media content, that is, is different from the media content, can be recommended to the user.

[0126] As an implementable method, the display order of the multiple recalled contents can be determined based on the information increments corresponding to the multiple recalled contents. For example, the multiple recalled contents can be sorted in order of information increments from large to small, so that after the multiple recalled contents are displayed in this order, the user can browse the recalled contents that are of interest and novel in a more prominent position.

[0127] The content recommendation method provided in this embodiment recalls recalled content similar to the media content based on the user's trigger operation, and determines the display order corresponding to the recalled content based on the information increment of each recalled content relative to the media content, thereby enabling the user to browse the content of interest more intuitively.

[0128] Figure 7 A schematic diagram of the structure of a content recommendation device provided by an embodiment of the present disclosure, such as Figure 7 As shown, the device includes: an acquisition module 71, a recall module 72, a determination module 73 and a recommendation module 74. The acquisition module 71 is used to acquire target media content. The recall module 72 is used to perform a content recall operation in a preset resource pool according to the target media content to obtain multiple recalled media contents. The determination module 73 is used to identify the information increment of the recalled media content relative to the target media content through a preset information increment recognition model for each recalled media content. The recommendation module 74 is used to perform a content recommendation operation on multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents.

[0129] Further, based on any of the above embodiments, the recall module is used to: respectively determine the vector distance between the target media content and each preset media content in the resource pool, and determine the preset media content whose vector distance meets the preset recall condition as the recalled media content.

[0130] Further, based on any of the above embodiments, the device also includes: a determination module, which is used to determine the publisher information associated with each of the recalled media contents, and obtain a preset publisher list. A processing module, which is used to determine, for each of the recalled media contents, whether the publisher information associated with the recalled media content is included in the publisher list, and obtain a processing result. A filtering module, which is used to perform content filtering operations on the multiple recalled media contents based on the processing results. Alternatively, after performing a content recall operation in a preset resource pool according to the target media content and obtaining multiple recalled media contents, it also includes: an identification module, which is used to perform content recognition operations on the multiple recalled media contents through a preset recognition model, and determine the media content to be filtered generated by an artificial intelligence algorithm. A filtering module, which is used to perform content filtering operations on the media content to be filtered among the multiple recalled media contents.

[0131] Further, on the basis of any of the above embodiments, the information increment identification model is a three-tower extension model, including a multimodal feature extraction module, a feature splicing module, and a classification module. The determination module is used to: input the target media content, the recalled media content, and the associated data into the multimodal feature extraction module to obtain a first feature vector corresponding to the target media content, a second feature vector corresponding to the recalled media content, and a third feature vector corresponding to the associated data, wherein the associated data includes parameter information and interaction information associated with the target media content and the recalled media content. The first feature vector, the second feature vector, and the third feature vector are subjected to a vector splicing operation by the splicing module to obtain a target feature vector. The target feature vector is input into the classification module to obtain a classification result output by the classification module, wherein the classification result includes the probability of different preset information increments of the recalled media content relative to the target media content. The preset information increment whose probability in the classification result meets the preset condition is used as the information increment of the recalled media content relative to the target media content.

[0132] Further, on the basis of any of the above embodiments, the device also includes: an acquisition module, used to acquire a preset training data set, wherein the training data set includes multiple sets of training data pairs, and the training data pairs include target content, recalled content based on the target content, associated data corresponding to the target content and the recalled content, and preset annotation information, and the annotation information is the information increment of the recalled content relative to the target content. A training module, used to input the target content, recalled content, and associated data in the training data pair into a preset model to be trained to obtain the predicted information increment output by the model to be trained. A determination module, used to determine the loss value corresponding to the model to be trained based on the predicted information increment and the annotation information. A judgment module, used to determine whether the model to be trained meets the preset convergence condition based on the loss value. A processing module, used to determine the current model to be trained as the information increment recognition model if it meets the condition. The adjustment module is used to adjust the parameters of the model to be trained based on the loss value if it is not satisfied, and return to execute the step of inputting the target content, recall content, and related data in the training data pair into the preset model to be trained to obtain the predicted information increment output by the model to be trained, until the model to be trained meets the preset convergence conditions, and the information increment recognition model is obtained.

[0133] Further, based on any of the above embodiments, the recommendation module is used to: determine a plurality of target recall contents whose information increment is greater than a preset increment threshold, and perform a content recommendation operation according to the plurality of target recall contents.

[0134] Further, based on any of the above embodiments, the recommendation module is used to: determine a recommendation order corresponding to the plurality of recalled media contents according to the order of information increment from large to small, and perform content recommendation operations on the plurality of recalled media contents according to the recommendation order.

[0135] Further, based on any of the above embodiments, the device also includes: a display module, which is used to display multiple media contents in a preset media content display interface. A recall module, which is used to respond to a user's trigger operation on any media content and recall multiple recalled contents associated with the media content in the resource pool based on a preset recall algorithm. A determination module, which is used to determine, for each recalled content, the information increment of the recalled content relative to the media content. A processing module, which is used to determine the display order corresponding to the multiple recalled contents based on the information increment corresponding to the multiple recalled contents. A display module, which is used to display the multiple recalled contents in the media content display interface according to the display order.

[0136] The device provided in this embodiment can be used to execute the technical solution of the above method embodiment. Its implementation principle and technical effect are similar, and this embodiment will not be repeated here.

[0137] In order to implement the above embodiments, the embodiments of the present disclosure further provide a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the content recommendation method described in any of the above embodiments is implemented.

[0138] In order to implement the above embodiments, the embodiments of the present disclosure further provide a computer program product, including a computer program, and when the computer program is executed by a processor, the content recommendation method as described in any of the above embodiments is implemented.

[0139] In order to implement the above embodiment, the present disclosure also provides an electronic device, including: a processor and a memory;

[0140] The memory stores computer-executable instructions;

[0141] The processor executes the computer-executable instructions stored in the memory, so that the processor performs the content recommendation method as described in any of the above embodiments.

[0142] Figure 8 The electronic device 800 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure, and the electronic device 800 may be a terminal device or a server. The terminal device may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, personal digital assistants (PDAs), tablet computers, portable multimedia players (PMPs), vehicle terminals (such as vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 8 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0143] like Figure 8 As shown, the electronic device 800 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage device 808 to a random access memory (RAM) 803. Various programs and data required for the operation of the electronic device 800 are also stored in the RAM 803. The processing device 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0144] Typically, the following devices may be connected to the I / O interface 805: input devices 806 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 807 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 808 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 809. The communication device 809 may allow the electronic device 800 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 8 The electronic device 800 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0145] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device 809, or installed from a storage device 808, or installed from a ROM 802. When the computer program is executed by the processing device 801, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0146] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0147] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0148] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device executes the method shown in the above embodiment.

[0149] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0150] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0151] The units involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a unit does not limit the unit itself in some cases. For example, the first acquisition unit may also be described as a "unit for acquiring at least two Internet Protocol addresses".

[0152] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0153] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0154] In a first aspect, according to one or more embodiments of the present disclosure, a content recommendation method is provided, comprising:

[0155] Acquire target media content;

[0156] Performing a content recall operation in a preset resource pool according to the target media content to obtain a plurality of recalled media contents;

[0157] For each recalled media content, identifying the information increment of the recalled media content relative to the target media content through a preset information increment identification model;

[0158] A content recommendation operation is performed on the multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents.

[0159] According to one or more embodiments of the present disclosure, performing a content recall operation in a preset resource pool according to the target media content to obtain a plurality of recalled media contents includes:

[0160] respectively determining vector distances between the target media content and each preset media content in the resource pool;

[0161] The preset media content whose vector distance satisfies the preset recall condition is determined as the recalled media content.

[0162] According to one or more embodiments of the present disclosure, after performing a content recall operation in a preset resource pool according to the target media content and obtaining a plurality of recalled media contents, the method further includes:

[0163] Determining publisher information associated with each of the recalled media contents, and obtaining a preset publisher list;

[0164] For each of the recalled media contents, determining whether the publisher list includes publisher information associated with the recalled media content, and obtaining a processing result;

[0165] performing a content filtering operation on the plurality of recalled media contents based on the processing result;

[0166] Alternatively, after performing a content recall operation in a preset resource pool according to the target media content and obtaining a plurality of recalled media contents, the method further includes:

[0167] Performing content recognition operations on the plurality of recalled media contents through a preset recognition model recognition to determine the media contents to be filtered generated by an artificial intelligence algorithm;

[0168] A content filtering operation is performed on the media contents to be filtered among the plurality of recalled media contents.

[0169] According to one or more embodiments of the present disclosure, the information increment recognition model is a three-tower extension model, including a multimodal feature extraction module, a feature splicing module, and a classification module;

[0170] The identifying the information increment of the recalled media content relative to the target media content by using a preset information increment identification model includes:

[0171] Inputting the target media content, the recalled media content and the associated data into the multimodal feature extraction module to obtain a first feature vector corresponding to the target media content, a second feature vector corresponding to the recalled media content and a third feature vector corresponding to the associated data, wherein the associated data includes parameter information and interaction information associated with the target media content and the recalled media content;

[0172] Performing a vector splicing operation on the first feature vector, the second feature vector and the third feature vector by the splicing module to obtain a target feature vector;

[0173] Inputting the target feature vector into the classification module to obtain a classification result output by the classification module, wherein the classification result includes probabilities of different preset information increments of the recalled media content relative to the target media content;

[0174] The preset information increment in the classification result that has a probability of satisfying a preset condition is used as the information increment of the recalled media content relative to the target media content.

[0175] According to one or more embodiments of the present disclosure, before identifying the information increment of the recalled media content relative to the target media content through a preset information increment identification model, the method further includes:

[0176] Acquire a preset training data set, wherein the training data set includes a plurality of training data pairs, the training data pairs include target content, recalled content based on the target content, associated data corresponding to the target content and the recalled content, and preset annotation information, wherein the annotation information is an information increment of the recalled content relative to the target content;

[0177] Inputting the target content, the recalled content, and the associated data in the training data pair into a preset model to be trained, and obtaining the predicted information increment output by the model to be trained;

[0178] Determine the loss value corresponding to the model to be trained based on the prediction information increment and the annotation information;

[0179] Determining whether the model to be trained meets a preset convergence condition according to the loss value;

[0180] If satisfied, the current model to be trained is determined as the information increment recognition model;

[0181] If it is not satisfied, the parameters of the model to be trained are adjusted based on the loss value, and the step of inputting the target content, recall content, and related data in the training data pair into the preset model to be trained is returned to obtain the predicted information increment output by the model to be trained, until the model to be trained meets the preset convergence conditions, and the information increment recognition model is obtained.

[0182] According to one or more embodiments of the present disclosure, the performing a content recommendation operation on the multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents includes:

[0183] Determine a plurality of target recall contents whose information increment is greater than a preset increment threshold;

[0184] A content recommendation operation is performed according to the multiple target recalled contents.

[0185] According to one or more embodiments of the present disclosure, the performing a content recommendation operation on the multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents includes:

[0186] Determine a recommendation order corresponding to the plurality of recalled media contents according to the order of information increment from large to small;

[0187] A content recommendation operation is performed on the plurality of recalled media contents according to the recommendation order.

[0188] According to one or more embodiments of the present disclosure, the method further includes:

[0189] Displaying multiple media contents on a preset media content display interface;

[0190] In response to a trigger operation of a user on any media content, recalling a plurality of recalled contents associated with the media content from the resource pool based on a preset recall algorithm;

[0191] For each recalled content, determining an information increment of the recalled content relative to the media content;

[0192] Determining a display order corresponding to the plurality of recalled contents based on the information increments corresponding to the plurality of recalled contents;

[0193] The plurality of recalled contents are displayed in the media content display interface according to the display order.

[0194] In a second aspect, according to one or more embodiments of the present disclosure, a content recommendation device is provided, comprising:

[0195] An acquisition module, used for acquiring target media content;

[0196] A recall module, configured to perform a content recall operation in a preset resource pool according to the target media content, and obtain a plurality of recalled media contents;

[0197] A determination module, configured to identify, for each recalled media content, an information increment of the recalled media content relative to the target media content through a preset information increment identification model;

[0198] The recommendation module is used to perform a content recommendation operation on the multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents.

[0199] According to one or more embodiments of the present disclosure, the recall module is used to:

[0200] respectively determining vector distances between the target media content and each preset media content in the resource pool;

[0201] The preset media content whose vector distance satisfies the preset recall condition is determined as the recalled media content.

[0202] According to one or more embodiments of the present disclosure, the device further includes:

[0203] A determination module, used to determine publisher information associated with each of the recalled media contents, and to obtain a preset publisher list;

[0204] A processing module, for determining, for each of the recalled media contents, whether the publisher list includes publisher information associated with the recalled media contents, and obtaining a processing result;

[0205] A filtering module, configured to perform content filtering operations on the plurality of recalled media contents based on the processing result;

[0206] Alternatively, after performing a content recall operation in a preset resource pool according to the target media content and obtaining a plurality of recalled media contents, the method further includes:

[0207] an identification module, configured to perform a content identification operation on the plurality of recalled media contents through a preset identification model, and determine the media contents to be filtered generated by an artificial intelligence algorithm;

[0208] The filtering module is used to perform content filtering operations on the media contents to be filtered among the plurality of recalled media contents.

[0209] According to one or more embodiments of the present disclosure, the information increment recognition model is a three-tower extension model, including a multimodal feature extraction module, a feature splicing module, and a classification module;

[0210] The determining module is used to:

[0211] Inputting the target media content, the recalled media content and the associated data into the multimodal feature extraction module to obtain a first feature vector corresponding to the target media content, a second feature vector corresponding to the recalled media content and a third feature vector corresponding to the associated data, wherein the associated data includes parameter information and interaction information associated with the target media content and the recalled media content;

[0212] Performing a vector splicing operation on the first feature vector, the second feature vector and the third feature vector by the splicing module to obtain a target feature vector;

[0213] Inputting the target feature vector into the classification module to obtain a classification result output by the classification module, wherein the classification result includes probabilities of different preset information increments of the recalled media content relative to the target media content;

[0214] The preset information increment in the classification result that has a probability of satisfying a preset condition is used as the information increment of the recalled media content relative to the target media content.

[0215] According to one or more embodiments of the present disclosure, the device further includes:

[0216] An acquisition module, configured to acquire a preset training data set, wherein the training data set includes a plurality of training data pairs, wherein the training data pairs include target content, recalled content based on the target content, associated data corresponding to the target content and the recalled content, and preset annotation information, wherein the annotation information is an information increment of the recalled content relative to the target content;

[0217] A training module, used to input the target content, the recalled content, and the associated data in the training data pair into a preset model to be trained, and obtain the prediction information increment output by the model to be trained;

[0218] A determination module, used to determine the loss value corresponding to the model to be trained based on the prediction information increment and the annotation information;

[0219] A determination module, used to determine whether the model to be trained meets a preset convergence condition according to the loss value;

[0220] A processing module, configured to determine the current model to be trained as the information increment recognition model if the conditions are met;

[0221] The adjustment module is used to adjust the parameters of the model to be trained based on the loss value if it is not satisfied, and return to execute the step of inputting the target content, recall content, and related data in the training data pair into the preset model to be trained to obtain the predicted information increment output by the model to be trained, until the model to be trained meets the preset convergence conditions, and the information increment recognition model is obtained.

[0222] According to one or more embodiments of the present disclosure, the recommendation module is used to:

[0223] Determine a plurality of target recall contents whose information increment is greater than a preset increment threshold;

[0224] A content recommendation operation is performed according to the multiple target recalled contents.

[0225] According to one or more embodiments of the present disclosure, the recommendation module is used to:

[0226] Determine a recommendation order corresponding to the plurality of recalled media contents according to the order of information increment from large to small;

[0227] A content recommendation operation is performed on the plurality of recalled media contents according to the recommendation order.

[0228] According to one or more embodiments of the present disclosure, the device further includes:

[0229] A display module, used to display multiple media contents on a preset media content display interface;

[0230] A recall module, configured to, in response to a trigger operation of a user on any media content, recall a plurality of recalled contents associated with the media content in the resource pool based on a preset recall algorithm;

[0231] A determination module, configured to determine, for each recalled content, an information increment of the recalled content relative to the media content;

[0232] A processing module, configured to determine a display order corresponding to the plurality of recalled contents based on information increments corresponding to the plurality of recalled contents;

[0233] A display module is used to display the multiple recalled contents in the media content display interface according to the display order.

[0234] In a third aspect, according to one or more embodiments of the present disclosure, there is provided an electronic device, comprising: at least one processor and a memory;

[0235] The memory stores computer-executable instructions;

[0236] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the content recommendation method described in the first aspect and various possible designs of the first aspect.

[0237] In a fourth aspect, according to one or more embodiments of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer execution instructions. When a processor executes the computer execution instructions, the content recommendation method described in the first aspect and various possible designs of the first aspect is implemented.

[0238] In a fifth aspect, according to one or more embodiments of the present disclosure, a computer program product is provided, including a computer program, which, when executed by a processor, implements the content recommendation method as described in the first aspect and various possible designs of the first aspect.

[0239] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0240] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0241] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

Claims

1. A content recommendation method, characterized in that: include: Acquire target media content; Performing a content recall operation in a preset resource pool according to the target media content to obtain a plurality of recalled media contents; For each recalled media content, identifying the information increment of the recalled media content relative to the target media content through a preset information increment identification model; A content recommendation operation is performed on the multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents.

2. The method according to claim 1, characterized in that The performing of a content recall operation in a preset resource pool according to the target media content to obtain a plurality of recalled media contents includes: respectively determining vector distances between the target media content and each preset media content in the resource pool; The preset media content whose vector distance satisfies the preset recall condition is determined as the recalled media content.

3. The method according to claim 1, characterized in that After performing a content recall operation in a preset resource pool according to the target media content and obtaining a plurality of recalled media contents, the method further includes: Determining publisher information associated with each of the recalled media contents, and obtaining a preset publisher list; For each of the recalled media contents, determining whether the publisher list includes publisher information associated with the recalled media content, and obtaining a processing result; performing a content filtering operation on the plurality of recalled media contents based on the processing result; Alternatively, after performing a content recall operation in a preset resource pool according to the target media content and obtaining a plurality of recalled media contents, the method further includes: Performing content recognition operations on the plurality of recalled media contents through a preset recognition model recognition to determine the media contents to be filtered generated by an artificial intelligence algorithm; A content filtering operation is performed on the media contents to be filtered among the plurality of recalled media contents.

4. The method according to claim 1, characterized in that: The information increment recognition model is a three-tower extension model, including a multimodal feature extraction module, a feature splicing module, and a classification module; The identifying the information increment of the recalled media content relative to the target media content by using a preset information increment identification model includes: Inputting the target media content, the recalled media content and the associated data into the multimodal feature extraction module to obtain a first feature vector corresponding to the target media content, a second feature vector corresponding to the recalled media content and a third feature vector corresponding to the associated data, wherein the associated data includes parameter information and interaction information associated with the target media content and the recalled media content; Performing a vector splicing operation on the first feature vector, the second feature vector and the third feature vector by the splicing module to obtain a target feature vector; Inputting the target feature vector into the classification module to obtain a classification result output by the classification module, wherein the classification result includes probabilities of different preset information increments of the recalled media content relative to the target media content; The preset information increment in the classification result that has a probability of satisfying a preset condition is used as the information increment of the recalled media content relative to the target media content.

5. The method according to claim 1, characterized in that Before identifying the information increment of the recalled media content relative to the target media content by using the preset information increment identification model, the method further includes: Acquire a preset training data set, wherein the training data set includes a plurality of training data pairs, the training data pairs include target content, recalled content based on the target content, associated data corresponding to the target content and the recalled content, and preset annotation information, wherein the annotation information is an information increment of the recalled content relative to the target content; Inputting the target content, the recalled content, and the associated data in the training data pair into a preset model to be trained, and obtaining the predicted information increment output by the model to be trained; Determine the loss value corresponding to the model to be trained based on the prediction information increment and the annotation information; Determining whether the model to be trained meets a preset convergence condition according to the loss value; If satisfied, the current model to be trained is determined as the information increment recognition model; If it is not satisfied, the parameters of the model to be trained are adjusted based on the loss value, and the step of inputting the target content, recall content, and related data in the training data pair into the preset model to be trained is returned to obtain the predicted information increment output by the model to be trained, until the model to be trained meets the preset convergence conditions, and the information increment recognition model is obtained.

6. The method according to claim 1, characterized in that The performing a content recommendation operation on the plurality of recalled media contents based on the information increments corresponding to the plurality of recalled media contents includes: Determine a plurality of target recall contents whose information increment is greater than a preset increment threshold; A content recommendation operation is performed according to the multiple target recalled contents.

7. The method according to claim 1, characterized in that The performing a content recommendation operation on the plurality of recalled media contents based on the information increments corresponding to the plurality of recalled media contents includes: Determine a recommendation order corresponding to the plurality of recalled media contents according to the order of information increment from large to small; A content recommendation operation is performed on the plurality of recalled media contents according to the recommendation order.

8. The method according to any one of claims 1 to 7, characterized in that: The method further comprises: Displaying multiple media contents on a preset media content display interface; In response to a trigger operation of a user on any media content, recalling a plurality of recalled contents associated with the media content from the resource pool based on a preset recall algorithm; For each recalled content, determining an information increment of the recalled content relative to the media content; Determining a display order corresponding to the plurality of recalled contents based on the information increments corresponding to the plurality of recalled contents; The plurality of recalled contents are displayed in the media content display interface according to the display order.

9. A content recommendation device, characterized in that: include: An acquisition module, used for acquiring target media content; A recall module, configured to perform a content recall operation in a preset resource pool according to the target media content, and obtain a plurality of recalled media contents; A determination module, configured to identify, for each recalled media content, an information increment of the recalled media content relative to the target media content through a preset information increment identification model; The recommendation module is used to perform a content recommendation operation on the multiple recalled media contents based on the information increments corresponding to the multiple recalled media contents.

10. An electronic device, characterized in that: include: Processor and memory; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the content recommendation method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the processor executes the computer-executable instructions, the content recommendation method according to any one of claims 1 to 8 is implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the content recommendation method according to any one of claims 1 to 8 is implemented.

Images