Resource display method and device, electronic equipment and storage medium

Abstract

The invention provides a resource display method and device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence, in particular to the fields of large models, deep learning, big data, intelligent recommendation and the like. The specific implementation scheme is as follows: determining a target period matched with a candidate resource from at least two candidate periods; determining a period preference value of the object for the target period; according to the target resource features of the candidate resources and the target object features of the objects, determining resource preference values of the objects to the candidate resources; determining a display state of the candidate resource according to the resource preference value and the period preference value; and displaying the resources based on the display state.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, and in particular to the fields of large models, deep learning, big data, and intelligent recommendations. More specifically, the present disclosure provides a resource display method, device, electronic device, storage medium, and computer program product. Background Art

[0002] Recommendation systems can recommend resources to users based on their historical browsing history. However, current recommendation systems mainly recommend resources based on users’ short-term interests, and the accuracy of recommendations needs to be improved. Summary of the invention

[0003] The present disclosure provides a resource display method, device, electronic device, storage medium and computer program product.

[0004] According to one aspect of the present disclosure, a resource display method is provided, including: determining a target period that matches a candidate resource from at least two candidate periods; determining a period preference value of an object for the target period; determining a resource preference value of the object for the candidate resource based on target resource features of the candidate resource and target object features of the object; determining a display status of the candidate resource based on the resource preference value and the period preference value; and displaying the resource based on the display status.

[0005] According to another aspect of the present disclosure, a resource display device is provided, comprising: a target period determination module, a first determination module, a second determination module, a state determination module and a display module. The target period determination module is used to determine a target period that matches a candidate resource from at least two candidate periods. The first determination module is used to determine a period preference value of an object for a target period. The second determination module is used to determine a resource preference value of an object for a candidate resource based on a target resource feature of the candidate resource and a target object feature of the object. The state determination module is used to determine a display state of a candidate resource based on a resource preference value and a period preference value; and the display module is used to display resources based on the display state.

[0006] According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method provided by the present disclosure.

[0007] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable a computer to execute the method provided by the present disclosure.

[0008] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the method provided by the present disclosure is implemented.

[0009] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0011] Figure 1 is a schematic diagram of an application scenario of the resource display method and device according to an embodiment of the present disclosure;

[0012] Figure 2 is a schematic flow chart of a resource display method according to an embodiment of the present disclosure;

[0013] Figure 3 is a schematic diagram of a resource display method according to an embodiment of the present disclosure;

[0014] Figure 4 is a schematic structural block diagram of a resource display device according to an embodiment of the present disclosure; and

[0015] Figure 5 It is a structural block diagram of an electronic device used to implement the resource display method of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0016] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0017] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0018] In the technical solution of the present disclosure, the user's authorization or consent is obtained before obtaining or collecting the user's personal information.

[0019] In some embodiments, the recommendation system mainly uses the user's short-term (e.g., 1 month) behavior data to estimate the user's interest value in the candidate resources, and then determines the resources to be recommended to the user based on the interest value. The recommendation system lacks the user's long-term (e.g., 1 year) behavior data when making resource recommendations.

[0020] In actual applications, for example, users prefer comedy resources in the past year, but prefer food resources in the past month and browse a lot of food resources. Since the recommendation system mainly recommends food resources to users based on short-term behavioral data, the recommendation system mainly recommends food resources to users, but rarely recommends martial arts resources. However, users are not completely uninterested in martial arts resources, so this recommendation method has poor accuracy and poor user experience.

[0021] The disclosed embodiments aim to provide a resource display method, which uses a period preference value to influence whether to display candidate resources, so that resources that meet the user's interests at different periods have a certain display rate, avoid recommending only resources that are of short-term interest to the object, and achieve regulation of resources that meet the object's interests at different periods, thereby improving the accuracy of recommendations and further enhancing the user experience.

[0022] The technical solution provided by the embodiment of the present disclosure can be applied to resource recommendation scenarios of types such as videos. The technical solution provided by the present disclosure will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0023] Figure 1 It is a schematic diagram of an application scenario of the resource display method and device according to an embodiment of the present disclosure.

[0024] It should be noted that Figure 1 What is shown is merely an example of a system architecture to which the embodiments of the present disclosure can be applied, in order to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios.

[0025] like Figure 1 As shown, the system architecture 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0026] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptops, desktop computers, etc.

[0027] The server 105 may be a server that provides various services, such as a background management server (only an example) that provides support for websites browsed by users using the terminal devices 101, 102, and 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as resources to be displayed determined according to the user's attribute characteristics, etc.) to the terminal device.

[0028] It should be noted that the resource display method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the resource display device provided in the embodiment of the present disclosure can generally be set in the server 105. The resource display method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the resource display device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105.

[0029] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0030] Figure 2 It is a schematic flow chart of a resource display method according to an embodiment of the present disclosure.

[0031] like Figure 2 As shown, the resource display method 200 may include operations S210 to S240.

[0032] In operation S210, a target period matching the candidate resource is determined from at least two candidate periods.

[0033] For example, the interest of the subject can be divided into short-term and long-term, or short-term, medium-term and long-term according to the length. For example, long-term can mean more than one year, short-term can mean within 7 days, and medium-term can mean from 7 days to one year. For example, if the subject has been browsing comedy resources continuously in the past two years, has browsed more martial arts resources in the past six months, and has browsed more food resources in the past week, then comedy is the subject's long-term interest, martial arts is the subject's medium-term interest, and food is the subject's short-term interest. The subject can be a user.

[0034] For example, the similarity between resources in each candidate period and the candidate resources can be determined, and the target period that matches the candidate resources can be determined based on the similarity. For example, the first similarity between resources that match the object's long-term interests and the candidate resources can be calculated, the second similarity between resources that match the object's medium-term interests and the candidate resources can be calculated, and the third similarity between resources that match the object's short-term interests and the candidate resources can be calculated. Then, the first similarity, the second similarity, and the third similarity are compared, and the candidate period corresponding to the maximum similarity is used as the target period.

[0035] In operation S220 , a period preference value of the subject for the target period is determined.

[0036] For example, the period preference values ​​of each candidate period may be preconfigured, for example, the period preference values ​​of long-term, medium-term, and short-term interests may be preconfigured to be 0.2, 0.3, and 0.5, respectively.

[0037] For another example, the period preference value of the target period may be determined according to the target resource characteristics of the candidate resource and the target object characteristics of the object. This determination method will be described in detail below.

[0038] In operation S230, a resource preference value of the object for the candidate resource is determined according to the target resource feature of the candidate resource and the target object feature of the object.

[0039] It should be noted that users are aware of and agree to the acquisition and use of target object characteristics, and that it complies with relevant laws and regulations and does not violate public order and good morals.

[0040] For example, the target resource features and target object features can be input into a pre-trained deep learning model, and the deep learning model can be used to determine the resource preference value. The higher the resource preference value, the higher the object's satisfaction with the candidate resource.

[0041] In operation S240, the display status of the candidate resource is determined according to the resource preference value and the period preference value.

[0042] In operation S250, resource display is performed based on the display status.

[0043] For example, when the resource preference value and the period preference value are respectively higher than the corresponding thresholds, the display state is determined to be display, so that normal display is performed. Otherwise, the display state is determined to be non-display, so no display is performed.

[0044] For another example, an evaluation value may be determined according to a resource preference value and a period preference value. If the evaluation value is higher than a corresponding threshold, the display state is determined to be display; otherwise, the display state is determined to be non-display.

[0045] According to the resource display method provided by the embodiment of the present disclosure, the method stratifies the interests of the object, for example, the interests of the object are divided into long-term interests, medium-term interests and short-term interests, and then determines the target period that matches the candidate resources, and then determines the period preference value of the object for the target period, and determines the display status of the candidate resources according to the resource preference value and the period preference value. In this way, the display status of the candidate resources can be affected by the period preference value. For example, when the period preference value of the target period is high, the possibility of displaying the candidate resources can be increased. When the period preference value of the target period is low, the possibility of displaying the candidate resources is reduced. Therefore, the resources displayed to the user can be made to meet the user's preference for interests in different periods, thereby improving the user experience. In addition, when multiple candidate resources involve multiple periods, the display status of the candidate resources is affected by the period preference value of each candidate resource, so that in the actual display process, the candidate resources of each period have a certain display rate, avoiding only recommending resources of interest in the short term to the object, thereby improving the accuracy of the recommendation, and thus improving the user experience.

[0046] In some embodiments, the object characteristics of the object or the target resource characteristics of the candidate resource may be determined through the candidate resource set. Next, the candidate resource set and the process of determining the candidate resource set are described.

[0047] Exemplarily, the number of candidate resource sets is at least two, each candidate resource set is associated with a candidate period, and the candidate periods associated with the candidate resource sets are different from each other. In addition, the resources in the candidate resource sets meet the satisfactory interaction condition and are in the candidate period associated with the historical interaction time.

[0048] For example, the candidate resource set associated with the long-term interest of the object is called the long-term interest satisfactory resource set, and the resources in the long-term interest satisfactory resource set are, for example, the resources that the user has interacted with in the past year, and all the resources meet the satisfactory interaction conditions. The candidate resource set associated with the medium-term interest of the object is called the medium-term interest satisfactory resource set, and the resources in the medium-term interest satisfactory resource set are, for example, the resources that the user has interacted with in half a year, and all the resources meet the satisfactory interaction conditions. The candidate resource set associated with the short-term interest of the object is called the short-term interest satisfactory resource set, and the resources in the short-term interest satisfactory resource set are, for example, the resources that the user has interacted with in the past week, and all the resources meet the satisfactory interaction conditions.

[0049] In the process of determining the candidate resource set associated with the candidate period, the resources that have interacted with the object and whose interaction time is in the associated candidate period can be grouped according to the category of the resources to obtain multiple resource groups. Then, for each resource group, according to the number of resources in the resource group that meet the satisfactory interaction condition and the total number of interactions between each resource in the resource group and the object, it is determined whether the resource group meets the interest condition. If the resource group meets the interest condition, it can be indicated that the object is interested in the resource group, and the resources in the resource group that meet the satisfactory interaction condition can be added to the candidate resource set associated with the candidate period.

[0050] Taking the long-term interest satisfaction resource set as an example, you can first filter out several resources that the object has interacted with within a year from the database, and then group them according to the category of the resources, for example, one resource group is martial arts resources, and the other resource group is comedy resources. Then determine whether the object is satisfied with the resource group. If satisfied, add the resources that meet the satisfactory interaction conditions to the long-term interest satisfaction resource set. The method of determining the medium-term interest satisfaction resource set and the short-term interest satisfaction resource set is similar to the method of determining the long-term interest satisfaction resource set. The difference is that the time range of the several resources used is different. For example, in the process of determining the medium-term interest satisfaction resource set, you can filter out several resources that the object has interacted with within half a year. In the process of determining the short-term interest satisfaction resource set, you can filter out several resources that the object has interacted with within a week.

[0051] This embodiment groups the resources interacted by the object in different time periods by category, then determines the resource group that the user is interested in, and filters the resources that the object is satisfied with from the resource group of interest and adds them to the corresponding candidate resource set. In this way, the object's interest preferences on different time scales can be accurately determined, thereby constructing a candidate resource set that meets the object's preferences, improving the accuracy of resource recommendations and enhancing the user experience.

[0052] According to another embodiment of the present disclosure, whether a resource group meets the interest condition can be determined in the following manner: if the ratio between the number of resources that meet the satisfactory interaction condition and the total number of resources in the resource group in a resource group is greater than or equal to a first ratio threshold, it can be determined that the resource group meets the interest condition. It can be seen that in this embodiment, if the object is relatively satisfied with more resources in the resource group, it can be considered that the object is interested in the resource group.

[0053] According to another embodiment of the present disclosure, whether a resource group meets the interest condition can be determined in the following manner: if the ratio between the number of resources that meet the satisfactory interaction condition and the total number of resources in the resource group in a certain resource group is greater than or equal to the first ratio threshold, and the total number of interactions is greater than or equal to the number threshold, then it can be determined that the resource group meets the interest condition. It should be noted that the number of resources in some resource groups is small. For example, a certain resource group is an animation resource group. The resource group only includes 3 resources. The object has completed the broadcast or other positive behaviors for the two resources. At the same time, the number of interactions between the object and the resources in the resource group is 3. Although the object has a high satisfaction rate with the resources in the animation resource, due to the small number of interactions, the confidence that the object is interested in the resource group is low, and the resource group is not suitable as a resource group of interest to the object. In this embodiment, when the object is relatively satisfied with more resources in the resource group and the number of interactions between the object and the resources in the resource group is large, it is determined that the object is interested in the resource group. Therefore, it is possible to more accurately determine whether the object is interested in the resource group, avoid mistakenly determining a resource group with low confidence as the resource group of interest, and thus improve the accuracy of determining the candidate resource group based on the resource group of interest.

[0054] In one example, a satisfactory interaction condition includes: the ratio between the browsing time of the resource and the total time of the resource is greater than or equal to a second ratio threshold. For example, taking the resource as a video resource, the higher the ratio between the browsing time of the resource and the total time of the resource, the closer the resource is to completion, and it can be considered that the object is relatively satisfied with the resource.

[0055] In another example, a satisfactory interaction condition includes: the browsing time of the resource is greater than or equal to the time threshold. For example, if the subject spends a long time browsing the resource, it can be considered that the subject is relatively satisfied with the resource.

[0056] In another example, a satisfactory interaction condition includes: the object inputs a predetermined category of operation data for a resource, and the predetermined category of operation data includes, for example, likes, favorites, shares, etc. If the object performs this type of operation on a resource, it can be considered that the object is relatively satisfied with the resource.

[0057] The candidate resource set and the process of determining the candidate resource set are described above.

[0058] Next, the process of determining the target object features of an object is described.

[0059] In one example, the attribute characteristics of an object may be used as target object characteristics. The attribute characteristics may include, for example, age, region, resource tags of interest, etc. It should be noted that the acquisition and use of information such as age, region, and tags are known and agreed to by the user, and are in compliance with relevant laws and regulations and do not violate public order and good morals.

[0060] In another example, the target object characteristics can be determined based on the attribute characteristics of the object and the characteristics of each resource in at least two candidate resource sets. For example, the attribute characteristics of the object, the long-term interest satisfaction resource set, the medium-term interest satisfaction resource set, and the short-term interest satisfaction resource set can be input into a pre-trained encoder to obtain the target object characteristics. In this embodiment, the target object characteristics are determined by two aspects, one is the basic attributes of the object, and the other is the object preference reflected by the historically satisfactory resources. In this way, the target object characteristics can be more accurately characterized by comprehensively considering the basic attributes of the user and the interest preferences in different time dimensions, and then more accurately determine whether to display the candidate resources, thereby improving the accuracy of the recommendation, making the recommended resources more in line with the actual needs of the object, and improving the object's satisfaction with the recommended resources.

[0061] The above describes the process of determining the target object features of an object.

[0062] Next, the process of determining target resource characteristics of candidate resources is described.

[0063] In one example, the features of the candidate resources may be used as the target resource features. For example, feature extraction may be performed based on the category, content, etc. of the candidate resources to obtain the features of the candidate resources.

[0064] In another example, a target resource set can be determined from at least two candidate resource sets according to the target period, and the resources in the target resource set meet the satisfactory interaction conditions and are in the target period with the historical interaction time. Then, the target resource characteristics are determined according to the characteristics of the candidate resources and the characteristics of each resource in the target resource set. It should be noted that in the process of determining whether the object is satisfied with the candidate resources, some resources related to the candidate resources can be screened first. In this embodiment, the screened related resources refer to the target resource set consistent with the candidate resource period. The object's satisfaction with these resources is equivalent to a posteriori, which can provide a reference for the resource preference value of the candidate resources, so that the object's resource preference value for the candidate resources can be more accurately determined based on the posteriori.

[0065] In the process of determining the characteristics of the target resource, for example, the target expert network can be determined from multiple candidate expert networks according to the target period, and then the characteristics of the candidate resources and the characteristics of each resource in the target resource set are input into the target expert network to obtain the characteristics of the target resource. For example, when the target period is long-term, the target resource set is a long-term interest satisfaction resource set, and the target expert network is a long-term expert network that is good at processing long-term interests. Compared with a general neural network to process the characteristics of the candidate resources and the characteristics of each resource in the target resource set, this embodiment sets multiple candidate expert networks, and selects the corresponding target expert network from multiple candidate expert networks according to different target periods, and then uses the target expert network to process data. This embodiment can use a target expert network that is more adept at processing the interest characteristics of the period for processing according to different periods such as long-term, medium-term or short-term, improve the professionalism of determining the characteristics of the target resource in long-term, medium-term or short-term interests, and then more finely characterize the characteristics of the target resource, so that the recommendation results are more in line with the needs of the object, and the accuracy of resource recommendation is improved.

[0066] According to another embodiment of the present disclosure, the process of determining the display status of the candidate resource according to the resource preference value and the period preference value may include: determining the target value of the candidate resource according to the resource preference value and the period preference value, and then determining the display status according to the target value. In this embodiment, the target value of the candidate resource is determined by the resource preference value and the period preference value, so that it can accurately determine whether to display the candidate resource, thereby improving the object's satisfaction with the recommendation result.

[0067] In one example, if the period preference value is greater than the first threshold and less than the second threshold, the initial value of the candidate resource is adjusted according to the resource preference value and the period preference value to obtain the target value. For example, the product of the initial value, the resource preference value and the period preference value can be determined as the target value.

[0068] In one example, if the period preference value is greater than or equal to the second threshold, the initial value is adjusted according to the resource preference value and the period preference value after the weighting process to obtain the target value; wherein the period preference value after the weighting process is greater than the period preference value. For example, the period preference value can be increased according to a predetermined ratio to obtain the period preference value after the weighting process, or the sum of the period preference value and the first predetermined value can be used as the period preference value after the weighting process. For example, the period preference value is 0.8, and the period preference value after the weighting process is 0.88. Then the product of the initial value, the resource preference value and the period preference value after the weighting process can be used as the target value. In this embodiment, when the period preference value of the target period is large, it means that the object is more likely to browse the resources that meet the interests of the target period. Therefore, the period preference value of the target period can be appropriately increased, thereby increasing the target value of the candidate resources that meet the interests of the target period, so that the candidate resources that meet the interests of the target period have a greater probability of being displayed to the object.

[0069] In another example, if the period preference value is less than or equal to the first threshold, the initial value is adjusted according to the resource preference value and the period preference value after the weight reduction process to obtain the target value; wherein the period preference value after the weight reduction process is less than the period preference value. For example, the period preference value can be reduced according to a predetermined ratio to obtain the period preference value after the weight reduction process, or the difference between the period preference value and the second predetermined value is used as the period preference value after the weight reduction process, for example, the period preference value is 0.15, and the period preference value after the weight increase process is 0.1. Then the product of the initial value, the resource preference value and the period preference value after the weight reduction process can be used as the target value. In this embodiment, when the period preference value of the target period is small, it means that the object has a low intention to browse resources that meet the interests of the target period. Therefore, the period preference value of the target period can be appropriately reduced, thereby reducing the target value of the resources that meet the interests of the target period, so as to reduce the possibility of the resources that meet the interests of the target period being recommended to the object.

[0070] It is understandable that in other examples, the period preference value may not be weighted up or down based on the size of the period preference value.

[0071] It is understandable that the number of candidate resources can be multiple, and the target value of each candidate resource can be determined in the above manner. Then, the multiple candidate resources can be sorted according to their respective target values. Then, among the multiple candidate resources, the candidate resources whose order satisfies the predetermined order condition are displayed. For example, the candidate resources ranked first or in the first several order are displayed to the object, so that the candidate resources with higher target values ​​are preferentially displayed to the object, thereby improving the user experience.

[0072] Figure 3 It is a schematic diagram of a resource display method according to an embodiment of the present disclosure.

[0073] In this embodiment, the model may include a routing network Router, multiple candidate expert networks Expert, an encoder Encoder, a first deep learning sub-model DNN1 and a second deep learning sub-model DNN2.

[0074] The target resource feature F_1 of the candidate resource may be determined first. For example, the feature F_3 of the candidate resource may be input into the routing network Router, and the routing network Router determines the correlation between the candidate resource and each candidate period, and determines the candidate period with the highest correlation as the target period.

[0075] The candidate expert network Expert includes, for example, a long-term expert network Expert1, a mid-term expert network Expert2, and a short-term expert network Expert3. The candidate period corresponds one-to-one to the candidate expert network Expert. After the target period is determined, the candidate expert network corresponding to the target period can be determined as the target expert network and activated. The candidate period also corresponds one-to-one to the candidate resource set. After the target period is determined, the candidate resource set associated with the target period can also be determined as the target resource set. For example, if the candidate resource has the highest correlation with the long-term interest of the object, the long-term interest satisfaction resource set can be determined as the target resource set, and the long-term expert network Expert1 can be activated. Next, the feature F_3 of the candidate resource and the feature F_4 of each resource in the long-term interest satisfaction resource set can be spliced ​​and input into the long-term expert network Expert1 to obtain the target resource feature F_1 of the candidate resource.

[0076] The target object feature F_2 of the object can be determined. For example, the attribute feature Xuser of the object and the features of each resource in each candidate resource set can be input into the encoder Encoder to obtain the target object feature F_2. During the processing, each candidate resource set includes, for example, a long-term interest satisfaction resource set, a medium-term interest satisfaction resource set, and a long-term interest satisfaction resource set. The three candidate resource sets have a total of m resources, namely nid1~nidm. The period marking token includes long-term bias, medium-term bias, and short-term bias. All three are pre-set tokens for marking the period in which the resource is located. Each resource corresponds to a period marking token, so there are a total of m period marking tokens. In addition, the period marking tokens corresponding to the resources of the same candidate period are the same, for example, the long-term bias corresponding to the resources of long-term interest are the same. Xseq is a pre-set sequence token. After being processed by the encoder, we can obtain the hidden features Hseq of the sequence, the hidden features H1~Hm of each resource, and the hidden features Huser of the object. Among them, the hidden features Hseq of the sequence have a global vision, and the global vision of other hidden features is limited. Therefore, the hidden features Hseq of the sequence can be used as the target object features F_2 of the object, or other network structures can be used to map the hidden features Hseq of the sequence to the target object features F_2 of the object.

[0077] Next, the target resource feature F_1 of the candidate resource and the target object feature F_2 of the object may be input into the first deep learning sub-model DNN1 to obtain the resource preference value V_p1.

[0078] The target resource feature F_1 of the candidate resource and the target object feature F_2 of the object can be input into the second deep learning sub-model DNN2, so as to obtain the preference ratios of the multiple candidate periods, and the preference ratio of the target period among the multiple preference ratios can be determined as the period preference value V_p2. It should be noted that the second deep learning sub-model DNN2 can output the object's preference ratio for long-term interests, medium-term interests, and short-term interests, but when the target period is long-term, only the preference ratio of long-term interests can be used later, without using the other two preference ratios. The reason for determining the preference ratios of long-term, medium-term, and short-term interests by the second deep learning sub-model DNN2 here is that different candidate resources may involve different periods, so it is necessary to model the preference ratios of long-term, medium-term, and short-term interests, instead of modeling the preference ratio of a single period. It should be noted that with the browsing behavior of the object, the object's preference for long-term, medium-term, and short-term interests changes dynamically, rather than being fixed. For example, after the object browses multiple resources of short-term interests continuously, it may no longer want to continue to look at resources of short-term interests. This embodiment determines the period preference value V_p2 of the target period according to the feature F_3 of the candidate resource. Compared with the method of pre-configuring the period preference value of each candidate period, this embodiment can dynamically adjust the period preference value V_p2, thereby improving the accuracy of recommendation.

[0079] After obtaining the resource preference value V_p1 and the period preference value V_p2, the target value V_tar of the candidate resource can be determined. For example, the initial value V_ini can be obtained first, and this embodiment does not limit the method for determining the initial value V_ini. Then, the initial value V_ini, the resource preference value V_p1 and the period preference value V_p2 are multiplied or weighted or calculated in other ways, and the calculation result is used as the target value V_tar.

[0080] Next, it can be determined whether to display the candidate resources according to the target value V_tar. For example, multiple candidate resources can be sorted according to their respective target values ​​V_tar, and then the first or first several candidate resources can be displayed to the object.

[0081] It should be noted that the above model is pre-trained, and the training process is similar to the inference process, except that the training samples correspond to labels, and the labels may include a first label representing the true resource preference value and a second label representing the true period preference value. During the training process, after obtaining the resource preference value and the period preference value output by the model, the first loss value can be calculated based on the first loss function according to the difference between the resource preference value and the first label, and the first loss function can be a cross entropy loss. The second loss value can also be calculated based on the second loss function according to the difference between the period preference value and the second label, and the second loss function can be a KL divergence. Then, the total loss value can be determined based on the first loss value and the second loss value, and the parameters of the routing network Router, multiple candidate expert networks Expert, encoder Encoder, the first deep learning sub-model DNN1 and the second deep learning sub-model DNN2 can be adjusted based on the total loss value until the model converges.

[0082] Figure 4 It is a schematic structural block diagram of a resource display device according to an embodiment of the present disclosure.

[0083] like Figure 4 As shown, the resource display device 400 may include a target period determination module 410 , a first determination module 420 , a second determination module 430 , a state determination module 440 and a display module 450 .

[0084] The target period determination module 410 is used to determine a target period that matches the candidate resource from at least two candidate periods.

[0085] The first determination module 420 is used to determine the subject's period preference value for the target period.

[0086] The second determination module 430 is used to determine the resource preference value of the object for the candidate resource according to the target resource feature of the candidate resource and the target object feature of the object.

[0087] The state determination module 440 is used to determine the display state of the candidate resource according to the resource preference value and the period preference value.

[0088] The display module 450 is used to display resources based on the display status.

[0089] According to another embodiment of the present disclosure, the state determination module includes: a target value determination submodule and a state determination submodule. The target value determination submodule is used to determine the target value of the candidate resource according to the resource preference value and the period preference value. The state determination submodule is used to determine the display state according to the target value.

[0090] According to another embodiment of the present disclosure, the target value determination submodule includes: a first adjustment unit, a second adjustment unit, and a third adjustment unit. The first adjustment unit is used to adjust the initial value of the candidate resource according to the resource preference value and the period preference value in response to detecting that the period preference value is greater than the first threshold and less than the second threshold, so as to obtain the target value. The second adjustment unit is used to adjust the initial value according to the resource preference value and the period preference value after the weight increase processing in response to detecting that the period preference value is greater than or equal to the second threshold, so as to obtain the target value; wherein the period preference value after the weight increase processing is greater than the period preference value. The third adjustment unit is used to adjust the initial value according to the resource preference value and the period preference value after the weight decrease processing in response to detecting that the period preference value is less than or equal to the first threshold, so as to obtain the target value; wherein the period preference value after the weight decrease processing is less than the period preference value.

[0091] According to another embodiment of the present disclosure, the target resource characteristics are obtained by the following modules: a target resource set determination module and a target resource characteristic determination module. The target resource set determination module is used to determine the target resource set from at least two candidate resource sets according to the target period. The target resource characteristic determination module is used to determine the target resource characteristics according to the characteristics of the candidate resources and the characteristics of each resource in the target resource set; wherein the resources in the target resource set meet the satisfactory interaction conditions and are in the target period with the historical interaction time.

[0092] According to another embodiment of the present disclosure, the target resource feature determination module includes: an expert network determination submodule and a target resource feature determination submodule. The expert network determination submodule is used to determine the target expert network from multiple candidate expert networks according to the target period. The target resource feature determination submodule is used to input the features of the candidate resources and the features of each resource in the target resource set into the target expert network to obtain the target resource features.

[0093] According to another embodiment of the present disclosure, the target object characteristics are obtained through the following modules: a target object characteristic determination module, used to determine the target object characteristics based on the attribute characteristics of the object and the characteristics of each resource in at least two candidate resource sets; wherein each candidate resource set is associated with a candidate period, and at least two candidate periods associated with at least two candidate resource sets are different from each other; the resources in the candidate resource sets meet the satisfactory interaction conditions and are in the candidate period associated with the historical interaction time.

[0094] According to another embodiment of the present disclosure, the candidate resource set associated with the candidate period is determined by the following modules: a grouping module, a determination module, and an adding module. The grouping module is used to group the resources that have interacted with the object and whose interaction time is in the associated candidate period according to the category of the resources to obtain multiple resource groups. The determination module is used to determine, for each resource group, whether the resource group meets the interest condition based on the number of resources in the resource group that meet the satisfactory interaction condition and the total number of interactions between each resource in the resource group and the object. The adding module is used to add the resources in the resource group that meet the satisfactory interaction condition to the candidate resource set associated with the candidate period in response to determining that the resource group meets the interest condition.

[0095] According to another embodiment of the present disclosure, the device also includes a determination module for determining that the resource group meets the interest condition in response to detecting that the ratio between the number of resources that meet the satisfactory interaction condition and the total number of resources of the resource group is greater than or equal to a first ratio threshold, and the total number of interactions is greater than or equal to a number threshold.

[0096] According to another embodiment of the present disclosure, satisfactory interaction conditions include at least one of the following: the ratio between the browsing time of the resource and the total browsing time of the resource is greater than or equal to a second ratio threshold; the browsing time of the resource is greater than or equal to the time threshold; the object inputs operation data of a predetermined category for the resource.

[0097] According to another embodiment of the present disclosure, the first determination module includes: a preference ratio determination submodule and a period preference value determination submodule. The preference ratio determination submodule is used to determine the preference ratio of each of multiple candidate periods according to the target resource characteristics and the target object characteristics. The period preference value determination submodule is used to determine the preference ratio of the target period among the multiple preference ratios as the period preference value.

[0098] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, including at least one processor; and a memory communicatively connected to the at least one processor; the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned resource display method.

[0099] According to an embodiment of the present disclosure, the present disclosure further provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute the above-mentioned resource display method.

[0100] According to an embodiment of the present disclosure, the present disclosure further provides a computer program product, including a computer program, and the computer program implements the above-mentioned resource display method when executed by a processor.

[0101] Figure 51 is a block diagram of an electronic device for implementing the resource display method of an embodiment of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0102] like Figure 5 As shown, the device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0103] A number of components in the device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0104] The computing unit 501 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 501 performs the various methods and processes described above, such as the resource display method. For example, in some embodiments, the resource display method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the resource display method described above may be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to perform the resource display method in any other appropriate manner (e.g., by means of firmware).

[0105] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0106] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0107] 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 device, 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.

[0108] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0109] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0110] A computer system may include clients and servers. Clients and servers are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship to each other.

[0111] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0112] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A resource display method, comprising: Determine a target period that matches the candidate resource from at least two candidate periods; determining a period preference value of the subject for the target period; Determining a resource preference value of the object for the candidate resource according to a target resource feature of the candidate resource and a target object feature of the object; Determining the display status of the candidate resource according to the resource preference value and the period preference value; The resource is displayed based on the display status.

2. The method according to claim 1, wherein: Determining the display status of the candidate resource according to the resource preference value and the period preference value includes: Determining a target value of the candidate resource according to the resource preference value and the period preference value; The display status is determined according to the target value.

3. The method according to claim 2, wherein: Determining the target value of the candidate resource according to the resource preference value and the period preference value includes: In response to detecting that the period preference value is greater than a first threshold and less than a second threshold, adjusting the initial value of the candidate resource according to the resource preference value and the period preference value to obtain the target value; In response to detecting that the period preference value is greater than or equal to the second threshold, adjusting the initial value according to the resource preference value and the period preference value after the weighting process to obtain the target value; wherein the period preference value after the weighting process is greater than the period preference value; In response to detecting that the period preference value is less than or equal to the first threshold, the initial value is adjusted according to the resource preference value and the period preference value after demotion to obtain the target value; wherein the period preference value after demotion is less than the period preference value.

4. The method according to claim 1, wherein: The target resource characteristics are obtained by the following method: Determining a target resource set from at least two candidate resource sets according to the target period; Determining the target resource characteristics according to the characteristics of the candidate resources and the characteristics of each resource in the target resource set; The resources in the target resource set meet satisfactory interaction conditions and are within the target period with respect to historical interaction time.

5. The method according to claim 4, wherein: The determining the target resource characteristics according to the characteristics of the candidate resources and the characteristics of each resource in the target resource set includes: determining a target expert network from a plurality of candidate expert networks according to the target period; The characteristics of the candidate resources and the characteristics of each resource in the target resource set are input into the target expert network to obtain the characteristics of the target resource.

6. The method according to claim 1, wherein: The target object features are obtained by the following method: Determining the target object characteristics according to the attribute characteristics of the object and the characteristics of each resource in at least two candidate resource sets; Each of the candidate resource sets is associated with a candidate period, and at least two of the candidate periods associated with at least two of the candidate resource sets are different from each other; the resources in the candidate resource sets meet satisfactory interaction conditions and are in the candidate period associated with the historical interaction time.

7. The method according to any one of claims 4 to 6, wherein: The candidate resource set associated with the candidate period is determined by the following method: For the resources that the object has interacted with and whose interaction time is in the associated candidate period, grouping them according to resource categories to obtain multiple resource groups; For each resource group, determining whether the resource group meets the interest condition according to the number of resources in the resource group that meet the satisfactory interaction condition and the total number of interactions between each resource in the resource group and the object; In response to determining that the resource group satisfies the interest condition, resources in the resource group that satisfy the satisfactory interaction condition are added to the candidate resource set associated with the candidate period.

8. The method according to claim 7, wherein: The method further comprises: In response to detecting that the ratio between the number of resources satisfying the satisfactory interaction condition and the total number of resources of the resource group is greater than or equal to a first ratio threshold, and the total number of interactions is greater than or equal to a number threshold, it is determined that the resource group satisfies the interest condition.

9. The method according to any one of claims 4 to 8, wherein: The satisfactory interaction condition includes at least one of the following: The ratio between the browsing time of the resource and the total browsing time of the resource is greater than or equal to a second ratio threshold; The browsing time of the resource is greater than or equal to the time threshold; The object enters a predetermined category of operational data for the resource.

10. The method according to claim 1, wherein: The determining the object's period preference value for the target period includes: Determining the preference ratio of each of the plurality of candidate periods according to the target resource characteristics and the target object characteristics; The preference proportion of the target period among the plurality of preference proportions is determined as the period preference value.

11. A resource display device, comprising: A target period determination module, used to determine a target period matching the candidate resource from at least two candidate periods; A first determination module, used to determine the subject's period preference value for the target period; A second determination module, configured to determine a resource preference value of the object for the candidate resource according to a target resource feature of the candidate resource and a target object feature of the object; A state determination module, used to determine the display state of the candidate resource according to the resource preference value and the period preference value; The display module is used to display resources based on the display status.

12. The device according to claim 11, wherein The state determination module comprises: A target value determination submodule, used to determine the target value of the candidate resource according to the resource preference value and the period preference value; The state determination submodule is used to determine the display state according to the target value.

13. The device according to claim 12, wherein: The target value determination submodule includes: A first adjustment unit, configured to adjust the initial value of the candidate resource according to the resource preference value and the period preference value to obtain the target value in response to detecting that the period preference value is greater than a first threshold and less than a second threshold; a second adjustment unit, configured to adjust the initial value according to the resource preference value and the period preference value after the weighting process to obtain the target value in response to detecting that the period preference value is greater than or equal to the second threshold; wherein the period preference value after the weighting process is greater than the period preference value; A third adjustment unit is used to adjust the initial value according to the resource preference value and the period preference value after demotion to obtain the target value in response to detecting that the period preference value is less than or equal to the first threshold; wherein the period preference value after demotion is less than the period preference value.

14. The device according to claim 11, wherein: The target resource characteristics are obtained through the following modules: A target resource set determination module, configured to determine a target resource set from at least two candidate resource sets according to the target period; A target resource feature determination module, used to determine the target resource feature according to the feature of the candidate resource and the feature of each resource in the target resource set; The resources in the target resource set meet satisfactory interaction conditions and are within the target period with respect to historical interaction time.

15. The device according to claim 14, wherein: The target resource feature determination module includes: An expert network determination submodule, used to determine a target expert network from a plurality of candidate expert networks according to the target period; The target resource feature determination submodule is used to input the features of the candidate resources and the features of each resource in the target resource set into the target expert network to obtain the target resource features.

16. The device according to claim 11, wherein The target object features are obtained through the following modules: A target object feature determination module, used to determine the target object feature according to the attribute feature of the object and the features of each resource in at least two candidate resource sets; Each of the candidate resource sets is associated with a candidate period, and at least two of the candidate periods associated with at least two of the candidate resource sets are different from each other; the resources in the candidate resource sets meet satisfactory interaction conditions and are in the candidate period associated with the historical interaction time.

17. The device according to any one of claims 14 to 16, wherein: The candidate resource set associated with the candidate period is determined by the following modules: A grouping module, configured to group the resources that the object has interacted with and whose interaction time is in the associated candidate period according to resource categories to obtain a plurality of resource groups; A determination module, configured to determine, for each resource group, whether the resource group meets the interest condition according to the number of resources in the resource group that meet the satisfactory interaction condition and the total number of interactions between each resource in the resource group and the object; An adding module is used for adding the resources in the resource group that meet the satisfactory interaction condition to the candidate resource set associated with the candidate period in response to determining that the resource group meets the interest condition.

18. The apparatus according to claim 17, further comprising: A determination module is used to determine that the resource group meets the interest condition in response to detecting that the ratio between the number of resources that meet the satisfactory interaction condition and the total number of resources in the resource group is greater than or equal to a first ratio threshold, and the total number of interactions is greater than or equal to a number threshold.

19. The device according to any one of claims 14 to 18, wherein: The satisfactory interaction condition includes at least one of the following: The ratio between the browsing time of the resource and the total browsing time of the resource is greater than or equal to a second ratio threshold; The browsing time of the resource is greater than or equal to the time threshold; The object enters a predetermined category of operational data for the resource.

20. The device according to claim 11, wherein The first determining module comprises: A preference ratio determination submodule, used to determine the preference ratios of the plurality of candidate periods according to the target resource characteristics and the target object characteristics; The period preference value determination submodule is used to determine the preference proportion of the target period among the plurality of preference proportions as the period preference value.

21. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 10.

22. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 10.

23. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 10.

Images