Object matching method and device, readable medium, electronic equipment and program product

Abstract

An object matching method and apparatus, a readable medium, an electronic device and a program product, the method comprising: determining a first object and a second object, the first object being an object of target content to be published, the second object being an object of watching the target content, the target content comprising promotion content provided by a third object; predicting conversion information of the second object after the first object publishes the target content through a deep learning model; determining a target object for publishing the target content from the first object according to the conversion information; and matching the target object with the third object. The conversion information for different first objects and second objects can be predicted through the deep learning model, and then the target object is screened out according to the conversion information to be matched with the third object, so that the matching accuracy of the target object and the third object can be improved. For example, in a content promotion scene, waste of network resources and display resources in a content promotion process can be reduced.

Description

Technical Field

[0001] The present disclosure relates to the fields of deep learning models and computer technologies, and in particular, to an object matching method, apparatus, readable medium, electronic device, and program product. Background Art

[0002] Merchants cooperate with content authors. The merchants provide products, and the content authors publish content related to the products to promote the products for the merchants. Users select suitable products by browsing the content published by the content authors and the product information provided by the merchants.

[0003] In related technologies, in order to facilitate the cooperation between merchants and content authors, the user groups of merchants and the user groups of content authors are usually determined respectively, and matching is performed according to the overlap degree between the two user groups, and then matchmaking is performed according to the matching results. However, for different content authors, the conversion probability of the same user is different. Therefore, the accuracy of matching based on user overlap is not high, the content promotion effect after the cooperation between merchants and content authors cannot be guaranteed, and it is easy to cause waste of network resources and display resources during the content promotion process. Summary of the Invention

[0004] This summary of the invention is provided to introduce concepts in a brief form, which will be described in detail in the following detailed implementation section. This summary of the invention is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to be used to limit the scope of the claimed technical solution.

[0005] In a first aspect, the present disclosure provides an object matching method, and the object matching method includes: Determine a first object and a second object, where the first object is an object to publish target content, the second object is an object to view the target content, and the target content includes promotion content provided by a third object; Predict conversion information of the second object after the first object publishes the target content through a deep learning model; Determine a target object that publishes the target content from the first objects according to the conversion information; Match the target object with the third object.

[0006] In a second aspect, the present disclosure provides an object matching apparatus, and the object matching apparatus includes: A first determination module, configured to determine a first object and a second object, where the first object is an object to publish target content, the second object is an object to view the target content, and the target content includes promotion content provided by a third object; A model prediction module, configured to predict conversion information of the second object after the first object publishes the target content through a deep learning model; A second determination module, configured to determine a target object that publishes the target content from the first objects according to the conversion information; A matching module, configured to match the target object with the third object.

[0007] In a third aspect, the present disclosure provides a computer-readable medium, on which a computer program is stored, and when the program is executed by a processing device, the steps of the method described in the first aspect are implemented.

[0008] In a fourth aspect, the present disclosure provides an electronic device, including: A storage device, on which a computer program is stored; A processing device, configured to execute the computer program in the storage device to implement the steps of the method described in the first aspect.

[0009] In a fifth aspect, the present disclosure provides a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of the method described in the first aspect are implemented.

[0010] In the above technical solution, the conversion information of the second object after the first object publishes the target content is predicted through a deep learning model, and then the target object that publishes the target content is determined from the first objects according to the conversion information, and the target object is matched with the third object. By adopting the above method, the conversion information of the second object can be predicted for different first objects through the deep learning model, and then the target object is selected according to the conversion information to be matched with the third object, which can improve the matching accuracy between the target object and the third object. For example, in the content promotion scenario, the matching accuracy between the merchant and the content author can be improved, so that the merchant can select a content author with a high matching degree for cooperation, and further improve the content promotion effect after the cooperation between the merchant and the content author, thereby reducing the waste of network resources and display resources in the content promotion process.

[0011] Other features and advantages of the present disclosure will be described in detail in the subsequent specific implementation part. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Combined with the drawings and referring to the following specific implementation manners, the above and other features, advantages and aspects of the embodiments of the present disclosure will become more obvious. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic, and the original components and elements are not necessarily drawn to scale. In the drawings: Figure 1 is a flowchart of an object matching method shown according to an exemplary embodiment of the present disclosure; Figure 2 It is a schematic diagram showing the probability of multi - user prediction conversion according to an exemplary embodiment of the present disclosure; Figure 3 It is a schematic structural diagram of a deep - learning model according to an exemplary embodiment of the present disclosure; Figure 4 It is a schematic structural diagram of a deep - learning model according to an exemplary embodiment of the present disclosure; Figure 5 It is a schematic structural diagram of a deep - learning model according to an exemplary embodiment of the present disclosure; Figure 6 It is a block diagram of the structure of an object matching device according to an exemplary embodiment of the present disclosure; Figure 7 It is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present disclosure. Detailed implementation manners

[0013] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0014] It should be understood that the steps recited in the method embodiments of the present disclosure can be executed in a different order and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this regard.

[0015] As used herein, the term "including" and its variants are open - ended, that is, "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0016] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependent relationships.

[0017] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0018] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are for illustrative purposes only and are not used to limit the scope of these messages or information.

[0019] It can be understood that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, usage scopes, 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 through appropriate means in accordance with relevant laws and regulations.

[0020] For example, when responding 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 by the user 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 program, server, or storage medium that performs the operations of the technical solutions of the present disclosure based on the prompt message.

[0021] As an optional but non-limiting implementation manner, the manner of sending a prompt message to the user in response to receiving an active request from the user can be, for example, in the form of a pop-up window, and the prompt message can be presented in text in the pop-up window. In addition, the pop-up window can also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0022] It can be understood that the above process of notifying and obtaining the user's authorization is only illustrative and does not limit the implementation manners of the present disclosure. Other manners that meet relevant laws and regulations can also be applied to the implementation manners of the present disclosure.

[0023] At the same time, it can be understood that the data involved in the technical solution (including but not limited to the data itself, the acquisition or use of the data) should comply with the requirements of the corresponding laws, regulations and related provisions.

[0024] Taking a content publishing platform as an example, a merchant is a customer who pays to place advertisements on the platform, a content author is a user who creates and publishes content, and can also be called a talent, blogger, anchor, etc., and a user is a user who browses content on the content publishing platform. The content can be long video content, short video content, live broadcast content, graphic content, text content, etc., and the present disclosure does not limit this.

[0025] A merchant can cooperate with a content author. The merchant pays the content author, and the content author publishes content related to the merchant's products to promote the products for the merchant. The user selects suitable products by browsing the content published by the content author and the product information provided by the merchant.

[0026] In the related art, in order to facilitate the cooperation between merchants and content authors, the user groups of merchants and content authors are usually determined separately, and matching is performed according to the overlap between the two user groups, and then matchmaking is performed according to the matching results. For example, the target users of a merchant of a clothing brand are young women, and the content published by a certain content author is mainly dressing videos, and most of the users who watch the dressing videos are young women, then it is considered that the merchant and the content author are more matched, and the cooperation between the two can be facilitated. However, for different content authors, the conversion probability of the same user is different. For example, both content authors A and B publish video content related to product C. User D did not purchase product C after watching the video content published by content author A, but purchased product C after watching the video content published by content author B.

[0027] Therefore, the accuracy of matching based on user overlap is not high, and it is impossible to guarantee the content promotion effect after the cooperation between merchants and content authors, which in turn leads to the waste of network resources and display resources in the content promotion process.

[0028] In view of this, the embodiments of the present disclosure provide an object matching method, device, readable medium, electronic device and program product to solve the above technical problems.

[0029] The following further explains the embodiments of the present disclosure with reference to the accompanying drawings.

[0030] Figure 1 is a flowchart of an object matching method shown according to an exemplary embodiment of the present disclosure. Referring to Figure 1 , the object matching method may include the following steps: S101: Determine a first object and a second object.

[0031] Among them, the first object is the object to publish the target content, and the second object is the object to watch the target content. The target content includes the promotion content provided by the third object.

[0032] Exemplarily, the target content may be long video content, short video content, live content, graphic content, text content, etc. The promotion content may be product content related to the product to be promoted. The product may be a physical product or a virtual product. The product content may be content such as videos, pictures, and texts related to the product. The present disclosure does not limit this.

[0033] In possible ways, determining the second object includes: determining candidate second objects associated with a third object, and selecting a preset number of objects from the candidate second objects in descending order of priority according to the historical behaviors of the candidate second objects and a preset behavior priority as the second objects associated with the third object; or, determining similar objects of the same type as the third object, and using the second objects associated with the similar objects as the second objects associated with the third object.

[0034] Exemplarily, the behavior priorities for different merchants can be preset. For example, for game products, the behavior priorities from high to low are payment, activation, installation, download, click, etc., and for physical products, the behavior priorities from high to low are purchase, add to cart, favorite, click, etc. Specifically, it can be set according to the actual application scenario, and the present disclosure does not limit this.

[0035] Exemplarily, with the user's authorization, the historical behaviors of the user on the platform can be obtained. For example, if user X clicks on the product details page of merchant Y, then user X is determined as the associated user of merchant Y. Circle them level by level from the associated users in descending order of priority. For example, first select the users who triggered historical behaviors with higher priorities, and then select the users who triggered historical behaviors with the next level of priority until the preset number of target users is screened out. The specific number can be selected according to requirements, for example, it can be 10,000, and the present disclosure does not limit this.

[0036] Exemplarily, for new merchants or merchants with a small number of associated users, similar merchants of the same type as the merchant can be selected. For example, a representative reference merchant can be preset for different industry types, such as the clothing industry, the food industry, etc. For new merchants in this industry, the target users associated with the reference merchant can be used as the target users of this merchant. The process of the reference merchant determining the target users can refer to the above process of circling the target users, and the present disclosure will not elaborate here.

[0037] It should be understood that the higher the behavior priority, the higher the degree of user conversion, and the more representative the user group of the merchant. Based on this, predicting the conversion information of the content author to the user can better represent the matching degree between the merchant and the content author.

[0038] In other possible ways, since there are a large number of content authors on the same platform, content authors can also be selected. For example, content authors related to the merchant type can be preferentially selected for matching. For a clothing brand merchant, content authors who publish dressing content can be preferentially screened for matching. Or, content authors can also be circled level by level according to the preset matching priorities for different merchants. Taking a clothing brand merchant as an example again, the matching priorities from high to low are the conversion probability of historical users, the number of fans, content authors of dressing content, etc. Specifically, it can be set according to the actual situation, and the present disclosure does not limit this.

[0039] S102: Predict the conversion information of the second object after the first object publishes the target content through a deep learning model.

[0040] Exemplarily, the conversion information is the information regarding the conversion of the second object after the first object publishes the target content. For example, after the content author publishes the target content of a certain product, it is determined that after the user sees this target content, the conversion information in aspects such as purchase behavior and attention, such as the user browsing, favoriting, and purchasing the product, etc., can be characterized by the conversion probability. The conversion information can intuitively show the product promotion ability of the content author for the merchant.

[0041] S103: Determine the target object that publishes the target content from the first object according to the conversion information.

[0042] Exemplarily, the content authors can be recalled in the order of the conversion probability from high to low. For example, the top 20 content authors are selected and recommended to the merchant. The present disclosure does not limit this.

[0043] S104: Match the target object with the third object.

[0044] Exemplarily, the selected content authors can be shown to the merchant for the merchant to choose. Further, the conversion information and recommended reasons corresponding to different content authors can also be shown to the merchant, so that the merchant can more intuitively understand the differences between different content authors and facilitate the merchant to select suitable content authors for cooperation.

[0045] By adopting the above method, through the deep learning model, the conversion information of the second object for different first objects can be predicted, and then the target object can be screened out according to the conversion information and matched with the third object, which can improve the matching accuracy between the target object and the third object. For example, in the content promotion scenario, the matching accuracy between the merchant and the content author can be improved, so that the merchant can select a content author with a high matching degree for cooperation, and then improve the content promotion effect after the merchant and the content author cooperate, thereby reducing the waste of network resources and display resources in the content promotion process.

[0046] In a possible manner, the number of second objects is multiple. Predicting the conversion information of the second object after the first object publishes the target content through a deep learning model includes: for each target second object among the multiple second objects, predicting the conversion probability of the target second object after the first object publishes the target content through the deep learning model; predicting the conversion information of the multiple second objects after the first object publishes the target content according to the cumulative value of the conversion probabilities of each second object.

[0047] Exemplarily, the number of target users circled may be multiple, such as Figure 2As shown, the conversion probability ctr of all target users of merchant B by content author A can be predicted separately through a deep learning model first, and then the ctr of the target users of merchant B can be summed up through the following calculation formula. The goods-carrying ability of content author A for merchant B:

[0048] Among them, TA represents the target user. By analogy, the goods-carrying abilities of multiple content authors for merchant B can be obtained, so as to select merchants and content authors for matching according to requirements. For example, match content authors with high goods-carrying abilities with merchants for cooperation.

[0049] In a possible way, the conversion information of the second object after the first object publishes the target content is predicted through a deep learning model, including: determining a first feature based on the first object, determining a second feature based on the second object, and determining a third feature based on the third object and / or the promotional content provided by the third object; inputting the first feature, the second feature, and the third feature into the deep learning model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model.

[0050] Exemplarily, the first feature may be basic features such as the region where the content author is located and the vertical field of the published content obtained with the authorization of the content author, as well as statistical features such as historical goods-carrying statistical information. The second feature may be basic features such as the region where the user is located, the subscribed topics, the liked content, the favorited content, and the commented content obtained with the authorization of the user. The third feature may be features such as the product identifier and product type of the merchant obtained with the authorization of the merchant. Specifically, it can be determined according to the actual scenario, and the present disclosure does not limit this.

[0051] Exemplarily, inputting the first feature, the second feature, and the third feature into the deep learning model can obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model, such as the predicted conversion probability of the user after browsing the target content.

[0052] Predicting the goods-carrying ability of content authors for merchants through a deep learning model based on the three-party matching of content authors, merchants, and users not only bases on the number of target users, but also calculates the conversion probability of each user under the goods-carrying of different content authors respectively. The result is more accurate, the matching accuracy is higher, it can effectively predict the content promotion effect after the merchant cooperates with different content authors, and is convenient for the merchant to select a suitable content author, thereby reducing the waste of network resources and display resources in the content promotion process.

[0053] Among possible ways, the deep learning model is a single-tower model. The first feature, the second feature, and the third feature are input into the deep learning model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model, including: concatenating the first feature, the second feature, and the third feature to obtain a first concatenated feature; inputting the first concatenated feature into the single-tower model to obtain a first feature vector, and performing activation processing on the first feature vector according to the activation function of the single-tower model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model.

[0054] Exemplarily, the deep learning model can be a single-tower model as Figure 3 shown. After concatenating the first feature, the second feature, and the third feature and inputting them into a multi-layer neural network, a first feature vector is obtained. Then, activation processing is performed on the first feature vector according to the activation function of the single-tower model, such as the sigmoid function, to obtain the conversion probability of the second object after the first object publishes the target content predicted by the deep learning model.

[0055] Among possible ways, the deep learning model is a two-tower model. The first feature, the second feature, and the third feature are input into the deep learning model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model, including: concatenating the first feature and the third feature to obtain a second concatenated feature; inputting the second feature and the second concatenated feature into the two-tower model to obtain a second feature vector corresponding to the second feature and a third feature vector corresponding to the second concatenated feature, performing an inner product calculation on the second feature vector and the third feature vector, and performing activation processing on the result of the inner product calculation according to the activation function of the two-tower model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model.

[0056] Exemplarily, the deep learning model can be a two-tower model as Figure 4 shown. Inputting the second feature into a multi-layer neural network to obtain a second feature vector, inputting the concatenated first feature and third feature into another multi-layer neural network to obtain a third feature vector. Then, performing an inner product calculation on the second feature vector and the third feature vector, and performing activation processing on the result of the inner product calculation according to the activation function of the two-tower model to obtain the conversion probability of the second object after the first object publishes the target content predicted by the deep learning model.

[0057] Among possible ways, the deep learning model is a three-tower model. The first feature, the second feature, and the third feature are input into the deep learning model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model, including: inputting the first feature, the second feature, and the third feature into the three-tower model to obtain the fourth feature vector corresponding to the first feature, the fifth feature vector corresponding to the second feature, and the sixth feature vector corresponding to the third feature, performing an inner product calculation on the fourth feature vector, the fifth feature vector, and the sixth feature vector, and performing an activation process on the inner product calculation result according to the activation function of the three-tower model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model.

[0058] Exemplarily, the deep learning model can be a three-tower model as Figure 5 shown. Inputting the first feature into a multi-layer neural network to obtain the fourth feature vector, inputting the second feature into a multi-layer neural network to obtain the fifth feature vector, and inputting the third feature into a multi-layer neural network to obtain the sixth feature vector. Then, performing an inner product calculation on the fourth feature vector, the fifth feature vector, and the sixth feature vector, and performing an activation process on the inner product calculation result according to the activation function of the three-tower model to obtain the conversion probability of the second object after the first object publishes the target content predicted by the deep learning model.

[0059] In practical applications, any of the above deep learning model structures can be selected for use according to requirements, and the present disclosure places no restrictions thereon.

[0060] Among possible ways, the deep learning model is trained in the following manner: determining positive samples, negative samples, and an initial deep learning model, where the positive samples include the features of the first positive sample object, the features of the second positive sample object, and the features of the third positive sample object, and the second positive sample object is the object that is converted after viewing the sample content published by the first positive sample object and including the promotion content provided by the third positive sample object; the negative samples include the features of the first negative sample object, the features of the second negative sample object, and the features of the third negative sample object, and the second negative sample object is the object that is not converted after viewing the sample content published by the first negative sample object and including the promotion content provided by the third negative sample object; obtaining the augmented negative samples by randomly replacing the features of the first negative sample object in the negative samples; performing model training on the initial deep learning model based on the positive samples and the augmented negative samples, and obtaining the trained deep learning model when the preset model training completion condition is satisfied.

[0061] Exemplarily, links pointing to specific pages, such as product pages, mini-programs, event pages, etc., are added to the platform page. Whether a user is converted can be determined based on whether the user triggers these link behaviors. During the model training phase, positive samples of converted users and negative samples of unconverted users can be obtained for model training. For example, if user U1 is converted after browsing the content of merchant B1 published by content author A1, then (U1, A1, B1) is used as a set of positive samples. If user U2 is not converted after browsing the content of merchant B2 published by content author A2, then (U2, A2, B2) is used as a set of negative samples.

[0062] Exemplarily, in order to improve the generalization ability and discrimination ability of the model, negative samples can be augmented. For example, for negative samples (U2, A2, B2) and (U3, A3, B3), by randomly replacing the features of the content authors, the augmented negative samples (U2, A2, B2), (U3, A3, B3), (U2, A3, B2), and (U3, A2, B3) can be obtained. Of course, when the number of samples is large enough, the negative samples may not be augmented, and the present disclosure does not limit this.

[0063] Furthermore, based on the positive samples and the augmented negative samples, the initial deep learning model is trained. When the preset model training completion conditions are met, such as the number of model training iterations reaches the preset number, the accuracy of model prediction is greater than the preset threshold, etc., a trained deep learning model is obtained.

[0064] It should be noted that during the training process, feature preprocessing can be performed on any one of the first feature, the second feature, and the third feature. For example, feature values with a large data range are bucketed and normalized to obtain enumerated values of the features, and the preprocessed features are input into the deep learning model to improve the training efficiency of the deep learning model.

[0065] Based on the same concept, an object matching device is provided in an embodiment of the present disclosure, as Figure 6 shown. The object matching device 600 includes: A first determination module 601, configured to determine a first object and a second object. The first object is the object for which the target content is to be published, and the second object is the object that views the target content. The target content includes promotional content provided by a third object; A model prediction module 602, configured to predict the conversion information of the second object after the first object publishes the target content through a deep learning model; A second determination module 603, configured to determine a target object for publishing the target content from the first objects according to the conversion information; A matching module 604, configured to match the target object with the third object.

[0066] Optionally, the second determination module 603 is configured to: Determine a first feature based on the first object, and determine a second feature based on the second object, Determine a third feature based on the third object and / or the promotion content provided by the third object; Input the first feature, the second feature, and the third feature into a deep learning model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model.

[0067] Optionally, the deep learning model is a single-tower model, and the second determination module 603 is configured to: Concatenate the first feature, the second feature, and the third feature to obtain a first concatenated feature; Input the first concatenated feature into the single-tower model to obtain a first feature vector, and perform activation processing on the first feature vector according to the activation function of the single-tower model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model.

[0068] Optionally, the deep learning model is a two-tower model, and the second determination module 603 is configured to: Concatenate the first feature and the third feature to obtain a second concatenated feature; Input the second feature and the second concatenated feature into the two-tower model to obtain a second feature vector corresponding to the second feature and a third feature vector corresponding to the second concatenated feature, perform an inner product calculation on the second feature vector and the third feature vector, and perform activation processing on the inner product calculation result according to the activation function of the two-tower model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model.

[0069] Optionally, the deep learning model is a three-tower model, and the second determination module 603 is configured to: Input the first feature, the second feature, and the third feature into the three-tower model to obtain a fourth feature vector corresponding to the first feature, a fifth feature vector corresponding to the second feature, and a sixth feature vector corresponding to the third feature, perform an inner product calculation on the fourth feature vector, the fifth feature vector, and the sixth feature vector, and perform activation processing on the inner product calculation result according to the activation function of the three-tower model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model.

[0070] Optionally, the deep learning model is trained in the following manner: Determine positive samples, negative samples, and an initial deep learning model. Among them, the positive samples include the features of the first positive sample object, the features of the second positive sample object, and the features of the third positive sample object. The second positive sample object is an object that is converted after viewing the sample content published by the first positive sample object and including the promotional content provided by the third positive sample object; the negative samples include the features of the first negative sample object, the features of the second negative sample object, and the features of the third negative sample object. The second negative sample object is an object that is not converted after viewing the sample content published by the first negative sample object and including the promotional content provided by the third negative sample object; By randomly replacing the features of the first negative sample object in the negative samples, obtain the augmented negative samples; Based on the positive samples and the augmented negative samples, train the initial deep learning model. When the preset model training completion condition is met, obtain the trained deep learning model.

[0071] Optionally, the number of the second objects is multiple. The model prediction module 602 is used for: For each target second object among the multiple second objects, predict the conversion probability of the target second object after the first object publishes the target content through the deep learning model; According to the cumulative value of the conversion probabilities of each second object, predict the conversion information of the multiple second objects after the first object publishes the target content.

[0072] Optionally, the first determination module 601 is used for: Determine the candidate second objects associated with the third object. According to the historical behaviors of the candidate second objects and the preset behavior priorities, select a preset number of objects from the candidate second objects in the order of decreasing priority as the second objects associated with the third object; or, Determine the similar objects of the same type as the third object, and use the second objects associated with the similar objects as the second objects associated with the third object.

[0073] Based on the same concept, an embodiment of the present disclosure further provides a computer-readable medium, on which a computer program is stored. When the program is executed by a processing device, the steps of the above object matching method are implemented.

[0074] Based on the same concept, an embodiment of the present disclosure further provides an electronic device, which may include: A storage device, on which a computer program is stored; A processing device, configured to execute the computer program in the storage device to implement the steps of the above object matching method.

[0075] Based on the same concept, an embodiment of the present disclosure also provides a computer program product, including a computer program, which when executed by a processor, implements the steps of the above object matching method.

[0076] Reference is made below to Figure 7 , which shows a schematic structural diagram of an electronic device 700 suitable for implementing an embodiment of the present disclosure. The terminal device in the embodiment of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), in-vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 7 The electronic device shown is only an example and should not impose any limitations on the functions and usage scope of the embodiment of the present disclosure.

[0077] As Figure 7 shown, the electronic device 700 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 701, which may perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage device 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the electronic device 700 are also stored. The processing device 701, the ROM 702, and the RAM 703 are connected to each other through a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0078] Generally, the following devices may be connected to the I / O interface 705: an input device 706 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 707 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 708 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 709. The communication device 709 may allow the electronic device 700 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 7 the electronic device 700 with various devices is shown, it should be understood that it is not required to implement or include all the shown devices. More or fewer devices may be alternatively implemented or included.

[0079] In particular, according to embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product that includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes program code for performing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device 709, or installed from a storage device 708, or installed from a ROM 702. When the computer program is executed by a processing device 701, the above-described functions defined in the methods of the embodiments of the present disclosure are performed.

[0080] It should be noted that the above-mentioned computer-readable medium in the present disclosure can be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of a computer-readable storage medium can 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 can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present disclosure, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, and the computer-readable signal medium can 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 can be transmitted by any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0081] In some embodiments, any currently known or future-developed network protocol such as HTTP (HyperText Transfer Protocol) can be used for communication, and can be interconnected with digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include local area networks ("LANs"), wide area networks ("WANs"), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed networks.

[0082] The above computer-readable medium can be included in the above electronic device; or can exist separately without being assembled into the electronic device.

[0083] The above computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: determine a first object and a second object, where the first object is the object for which the target content is to be published, the second object is the object for viewing the target content, and the target content includes promotional content provided by a third object; predict the conversion information of the second object after the first object publishes the target content through a deep learning model; determine, based on the conversion information, the target object for publishing the target content from the first objects; and match the target object with the third object.

[0084] Computer program code for performing the operations of the present disclosure can be written in one or more programming languages or combinations thereof. The programming languages include, but are not limited to, object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., by using an Internet service provider to connect through the Internet).

[0085] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a portion of code that contains one or more executable instructions for implementing the specified logical function. It should also be noted that, in some alternative implementations, the functions noted in the blocks may occur in a different order than noted in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or by a combination of dedicated hardware and computer instructions.

[0086] The modules described in the embodiments of the present disclosure can be implemented in software or in hardware. In some cases, the name of a module does not constitute a limitation on the module itself.

[0087] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. By way of example, and without limitation, the 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), system-on-a-chip (SOCs), complex programmable logic devices (CPLDs), and the like.

[0088] In the context of the present disclosure, a machine-readable medium may be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. 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, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0089] The above description is only a preferred embodiment of the present disclosure and an explanation of the applied technical principles. 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 the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosure concept. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) having similar functions disclosed in the present disclosure.

[0090] In addition, although the operations are depicted in a particular order, this should not be understood as requiring that the operations be performed in the particular order shown or in sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the present disclosure. Certain features described in the context of separate embodiments may also be implemented in combination in a single embodiment. Conversely, the various features described in the context of a single embodiment may also be implemented separately or in any suitable sub-combination in multiple embodiments.

[0091] Although the subject matter has been described in language specific to structural features and / or methodological logical acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. On the contrary, the specific features and acts described above are merely example forms for implementing the claims. Regarding the apparatus in the above embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments related to the method, and will not be elaborated here.

Claims

1. An object matching method, characterized in that: The object matching method comprises: Determine a first object and a second object, wherein the first object is an object to which target content is to be published, and the second object is an object to view the target content, wherein the target content includes promotional content provided by a third object; Predicting, by a deep learning model, conversion information of the second object after the first object publishes the target content; Determining, from the first object, a target object for publishing the target content according to the conversion information; The target object is matched with the third object.

2. The object matching method according to claim 1, characterized in that: The predicting, by using a deep learning model, the conversion information of the second object after the first object publishes the target content includes: Based on the first object, determining a first feature, and based on the second object, determining a second feature, determining a third feature based on the third object and / or the promotional content provided by the third object; The first feature, the second feature, and the third feature are input into a deep learning model to obtain conversion information of the second object after the first object publishes the target content as predicted by the deep learning model.

3. The object matching method according to claim 2, characterized in that: The deep learning model is a single-tower model, and the first feature, the second feature, and the third feature are input into the deep learning model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model, including: Splicing the first feature, the second feature and the third feature to obtain a first splicing feature; The first concatenated feature is input into the single-tower model to obtain a first feature vector, and the first feature vector is activated according to the activation function of the single-tower model to obtain conversion information of the second object after the first object publishes the target content predicted by the deep learning model.

4. The object matching method according to claim 2, characterized in that: The deep learning model is a double-tower model, and the first feature, the second feature, and the third feature are input into the deep learning model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model, including: Splicing the first feature and the third feature to obtain a second splicing feature; The second feature and the second splicing feature are input into the double-tower model to obtain a second feature vector corresponding to the second feature and a third feature vector corresponding to the second splicing feature, an inner product calculation is performed on the second feature vector and the third feature vector, and the inner product calculation result is activated according to the activation function of the double-tower model to obtain conversion information of the second object after the first object publishes the target content as predicted by the deep learning model.

5. The object matching method according to claim 2, characterized in that: The deep learning model is a three-tower model, and the first feature, the second feature, and the third feature are input into the deep learning model to obtain the conversion information of the second object after the first object publishes the target content predicted by the deep learning model, including: The first feature, the second feature and the third feature are input into the three-tower model to obtain a fourth feature vector corresponding to the first feature, a fifth feature vector corresponding to the second feature and a sixth feature vector corresponding to the third feature, inner product calculation is performed on the fourth feature vector, the fifth feature vector and the sixth feature vector, and the inner product calculation result is activated according to the activation function of the three-tower model to obtain conversion information of the second object after the first object publishes the target content predicted by the deep learning model.

6. The object matching method according to any one of claims 1 to 5, characterized in that: The deep learning model is trained in the following way: Determine positive samples, negative samples, and an initial deep learning model, wherein the positive samples include features of a first positive sample object, features of a second positive sample object, and features of a third positive sample object, and the second positive sample object is an object that is converted after viewing sample content published by the first positive sample object and including promotional content provided by the third positive sample object; the negative samples include features of the first negative sample object, features of the second negative sample object, and features of the third negative sample object, and the second negative sample object is an object that is not converted after viewing sample content published by the first negative sample object and including promotional content provided by the third negative sample object; Obtaining an expanded negative sample by randomly replacing the features of the first negative sample object in the negative sample; The initial deep learning model is trained based on the positive samples and the expanded negative samples, and a trained deep learning model is obtained when a preset model training completion condition is met.

7. The object matching method according to any one of claims 1 to 5, characterized in that: The number of the second objects is multiple, and the predicting, by using the deep learning model, the conversion information of the second objects after the first object publishes the target content includes: For each target second object among the multiple second objects, predicting, by a deep learning model, a probability of the target second object being converted after the first object publishes the target content; The conversion information of the plurality of second objects after the first object publishes the target content is predicted according to the accumulated value of the conversion probability of each second object.

8. The object matching method according to any one of claims 1 to 5, characterized in that: Determining a second object includes: Determine candidate second objects associated with the third object, and select a preset number of objects from the candidate second objects in descending order of priority according to historical behaviors and preset behavior priorities of the candidate second objects as second objects associated with the third object; or, A similar object of the same type as the third object is determined, and a second object associated with the similar object is used as the second object associated with the third object.

9. An object matching device, characterized in that: The object matching device comprises: A first determination module is used to determine a first object and a second object, wherein the first object is an object to which target content is to be published, and the second object is an object to view the target content, wherein the target content includes promotional content provided by a third object; A model prediction module, configured to predict, through a deep learning model, conversion information of the second object after the first object publishes the target content; A second determination module, configured to determine, from the first object, a target object for publishing the target content according to the conversion information; A matching module is used to match the target object with the third object.

10. A computer readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processing device, the steps of the method according to any one of claims 1 to 8 are implemented.

11. An electronic device, characterized in that: include: a storage device having a computer program stored thereon; A processing device, configured to execute the computer program in the storage device to implement the steps of the method according to any one of claims 1 to 8.

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

Images