Data processing method and device, equipment and storage medium

Abstract

The embodiment of the invention provides a data processing method and device, equipment and a storage medium. The method comprises the steps that on the basis of corresponding object information of a group of converted objects converted by a target entity, corresponding feature representations of the group of converted objects are determined, and the object information of each converted object is at least related to conversion operation executed by the converted object in first-time recommendation; on the basis of object information of a candidate object to be recommended, feature representation of the candidate object is determined, and the object information of the candidate object at least comprises operation information related to conversion operation executed by the candidate object in the second-time recommendation; and generating a recommendation result of the candidate object for the target entity based on the corresponding feature representation of the group of converted objects and the feature representation of the candidate object. Therefore, the recommendation effect can be improved under the scene that the objects are frequently changed.

Description

Technical Field

[0001] Example embodiments of the present disclosure generally relate to the field of computers, and more particularly, to methods, apparatuses, devices, and computer-readable storage media for data processing. Background Art

[0002] The Internet provides access to a wide variety of objects. For example, various applications, products, audio and video content, etc. can be accessed through the Internet. As the number and types of objects grow rapidly, the recommendation system recommends resources that meet user needs to the audience. However, in scenarios where the objects to be recommended change frequently, the life cycle of each object is very short, perhaps only a few days. Therefore, for the above scenarios, it is expected to achieve improved personalized recommendation effects. Summary of the invention

[0003] In a first aspect of the present disclosure, a method for data processing is provided. The method includes: determining a group of corresponding feature representations of the transformed objects based on corresponding object information of a group of transformed objects transformed by a target entity, wherein the object information of each transformed object is at least related to the transformation operation performed on the transformed object in the first number of recommendations; determining the feature representation of the candidate object based on the object information of the candidate object to be recommended, wherein the object information of the candidate object at least includes operation information related to the transformation operation performed on the candidate object in the second number of recommendations; and generating a recommendation result of the candidate object for the target entity based on the corresponding feature representations of the group of transformed objects and the feature representations of the candidate object.

[0004] In a second aspect of the present disclosure, a device for data processing is provided. The device includes: a first feature representation determination module, configured to determine a set of corresponding feature representations of converted objects based on corresponding object information of a set of converted objects converted by a target entity, wherein the object information of each converted object is at least related to the conversion operation performed on the converted object in the first number of recommendations; a second feature representation determination module, configured to determine the feature representation of a candidate object based on the object information of a candidate object to be recommended, wherein the object information of the candidate object at least includes operation information related to the conversion operation performed on the candidate object in the second number of recommendations; and a recommendation result generation module, configured to generate a recommendation result of the candidate object for the target entity based on the corresponding feature representation of the set of converted objects and the feature representation of the candidate object.

[0005] In a third aspect of the present disclosure, an electronic device is provided. The device includes at least one processing unit; and at least one memory, the at least one memory is coupled to the at least one processing unit and stores instructions for execution by the at least one processing unit. When the instructions are executed by the at least one processing unit, the device executes the method of the first aspect.

[0006] In a fourth aspect of the present disclosure, a computer-readable storage medium is provided, wherein a computer program is stored on the computer-readable storage medium, and the computer program can be executed by a processor to implement the method of the first aspect.

[0007] It should be understood that the contents described in this content section are not intended to limit the key features or important features of the embodiments of the present disclosure, nor are they 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

[0008] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein:

[0009] Figure 1 A schematic diagram showing an example environment in which embodiments of the present disclosure can be implemented;

[0010] Figure 2 A flowchart showing an example process for training an object recommendation model according to some embodiments of the present disclosure is shown;

[0011] Figure 3 A flowchart showing an example process for training an object representation model according to some embodiments of the present disclosure is shown;

[0012] Figure 4 A flowchart showing an example process of generating recommendation results using an object recommendation model according to some embodiments of the present disclosure;

[0013] Figure 5 A flowchart showing a process for data processing according to some embodiments of the present disclosure is shown;

[0014] Figure 6 A schematic structural block diagram of a device for data processing according to some embodiments of the present disclosure is shown;

[0015] Figure 7 A block diagram of an electronic device capable of implementing one or more embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0016] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0017] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0018] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0019] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0020] It is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and relevant provisions.

[0021] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0022] It should be noted that the titles of any sections / subsections provided herein are not restrictive. Various embodiments are described throughout this article, and any type of embodiment may be included under any section / subsection. In addition, the embodiments described in any section / subsection may be combined in any manner with any other embodiments described in the same section / subsection and / or different sections / subsections.

[0023] Herein, unless explicitly stated, executing a step “in response to A” does not mean executing the step immediately after “A” but may include one or more intermediate steps.

[0024] In the description of the embodiments of the present disclosure, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions may be included below. The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may be included below.

[0025] As used herein, the term "model" can learn the association between the corresponding input and output from the training data, so that after the training is completed, the corresponding output can be generated for a given input. The generation of the model can be based on machine learning technology. Deep learning is a machine learning algorithm that processes inputs and provides corresponding outputs by using multiple layers of processing units. In this article, "model" may also be referred to as "machine learning model", "machine learning network" or "network", which are used interchangeably in this article. A model may also include different types of processing units or networks.

[0026] Generally, machine learning can be roughly divided into three stages, namely the training stage, the testing stage, and the application stage (also called the inference stage). In the training stage, a given model can be trained using a large amount of training data, and the parameter values ​​are continuously updated iteratively until the model can obtain consistent inferences that meet the expected goals from the training data. Through training, the model can be considered to be able to learn the association from input to output (also called the mapping of input to output) from the training data. The parameter values ​​of the trained model are determined. In the testing stage, the test input is applied to the trained model to test whether the model can provide the correct output, thereby determining the performance of the model. In the application stage, the model can be used to process the actual input based on the parameter values ​​obtained from the training to determine the corresponding output.

[0027] Embodiments of the present disclosure relate to recommending objects to entities. In some cases, an entity may perform a conversion or transformation operation on an object. The entity here may include, but is not limited to, an individual, an organization, an institution, a user, etc., and the object may include, but is not limited to, an item, media content (e.g., a video), etc. The conversion is a resource demand behavior that an entity may perform on an object, which may include, but is not limited to, purchase, click, browse, register, download, etc.

[0028] Example Environment

[0029] Figure 1 A schematic diagram of an example environment 100 is shown in which embodiments of the present disclosure can be implemented. Figure 1One or more entities 132 - 1 , 132 - 2 , 132 - 3 (also collectively or individually referred to as entities 132 ) are shown, such as users. Figure 1 One or more electronic devices 130-1, 130-2, 130-3 (also collectively or individually referred to as electronic devices 130) respectively associated with the entity 132 are also shown. In the environment 100, the object recommendation platform 110 recommends objects for the entity 132 based on the candidate object set 140. The candidate object set 140 includes a plurality of candidate objects, such as a candidate object 142-1, a candidate object 142-2, ..., a candidate object 142-M, which are also collectively or individually referred to as candidate objects 142. The information of the recommended objects can be presented to the entity 132 via the electronic device 130.

[0030] The object recommendation platform 110 recommends objects to the entity 132 through the electronic device 130. In some embodiments, the object recommendation platform 110 can use the object recommendation model to recommend objects to the entity 132. The recommended objects include but are not limited to items, videos, files, etc. This disclosure is not limited to this.

[0031] In the environment 100, the electronic device 130 can be any type of device with computing capabilities, including a terminal device or a server device. The terminal device can be any type of mobile terminal, fixed terminal or portable terminal, including a mobile phone 130-2, a desktop computer 130-1, a laptop computer, a notebook computer 130-3, a netbook computer, a tablet computer, a media computer, a multimedia tablet, a personal communication system (PCS) device, a personal navigation device, a personal digital assistant (PDA), an audio / video player, a digital camera / camcorder, a positioning device, a television receiver, a radio broadcast receiver, an e-book device, a gaming device, or any combination of the foregoing, including accessories and peripherals of these devices or any combination thereof. The server device can include, for example, a computing system / server, such as a mainframe, an edge computing node, an electronic device in a cloud environment, and the like.

[0032] It should be understood that the structure and functionality of environment 100 are described for exemplary purposes only and does not imply any limitation on the scope of the present disclosure.

[0033] Currently, in personalized recommendations, objects and the interactions between entities are very important features. However, in scenarios where the objects to be recommended change frequently, the life cycle of each object is very short, for example, it may only be a few days. Taking product recommendations as an example, one scenario is: when a merchant launches a promotion, various items are recommended to users every day, but some items are only different in identity, but they belong to the same category of objects.

[0034] However, in scenarios where objects change frequently, the traditional solution is to directly use the object's identity as a feature and record the entity's behavior sequence, for example, recording the entity's previous object click records. In this case, the distribution of the object's identity (ID) in historical data from different periods will vary greatly, and there may even be no intersection. This is because the object's life cycle is very short. For example, the valid objects a, b, and c a few days ago may become d, e, and f a few days later. This makes it difficult for the model to learn information at the granularity of the object's identity, reducing the recommendation effect. Therefore, directly using the object's identity as a feature will reduce the recommendation effect.

[0035] In an embodiment of the present disclosure, an improved scheme for data processing is proposed. In this scheme, based on the corresponding object information of a group of transformed objects transformed by the target entity, a group of corresponding feature representations of the transformed objects is determined, and the object information of each transformed object is at least related to the conversion operation performed on the transformed object in the first recommendation. Based on the object information of the candidate object to be recommended, the feature representation of the candidate object is determined, and the object information of the candidate object at least includes operation information related to the conversion operation performed on the candidate object in the second recommendation. Based on the corresponding feature representation of a group of transformed objects and the feature representation of the candidate object, a recommendation result of the candidate object for the target entity is generated. In this embodiment, the feature representation of the object to be recommended is used to identify the object, rather than simply using the object ID to identify it. In this way, the recommendation effect can be improved in scenarios where objects change frequently.

[0036] Some example embodiments of the present disclosure will be described below with reference to the accompanying drawings. As mentioned above, in some embodiments, an object recommendation model may be used to generate recommendation results. In order to more clearly understand the embodiments of the present disclosure, the training object recommendation model is first described below.

[0037] Figure 2 A flowchart of an example process for training an object recommendation model according to some embodiments is shown. Through process 200, the object recommendation platform 110 can implement training of an object recommendation model based on an object representation model.

[0038] Part or all of the process 200 may be implemented by the object recommendation platform 110 or may be implemented by other devices independent of the object recommendation platform 110, for example, may be implemented by other remote devices (terminal devices or service devices) with computing capabilities. In the following, for ease of discussion, the execution of the process 200 is described from the perspective of the object recommendation platform 110, but this is only exemplary.

[0039] In box 201, the object recommendation platform 110 obtains reference object information of the reference object. The reference object information includes at least operation information related to the conversion operation performed on the reference object in the recommendation of a predetermined number of times (also referred to as the third number). The reference object can be any suitable object, and the operation information can, for example, indicate which entities have performed conversion operations on the reference object (such as, including attribute information of the entity group that performed the conversion operation on the reference object), the scene where the conversion operation was performed (such as, including attribute information of the scene), and the like. For example, cold start information of the reference object can be collected as operation information. In some embodiments, the reference object information can also include attribute information of the reference object, such as various static attributes of the reference object, including but not limited to type, price, manufacturer, etc.

[0040] In block 202, the object recommendation platform 110 obtains reference entity information of a reference entity to which a reference object is recommended. The reference entity may be an entity for which the object recommendation platform 110 recommends a reference object. The reference entity information may include, for example, various suitable attribute information of the reference entity.

[0041] For example, if the object recommendation platform 110 allocates 10,000 exposure opportunities to each reference object, the reference object information of the reference object can be obtained from the previous 5,000 exposures. The reference object information includes operation information related to the conversion operations performed in the previous 5,000 recommendations, for example, what kind of reference entity group clicked, purchased, etc. In addition, the information of the entity group of the recommended reference object in the previous 5,000 exposures can be used as the reference entity information.

[0042] In block 205, the object recommendation platform 110 trains the object representation model 210. Thus, a trained object representation model 210 can be obtained. In some embodiments, the object recommendation platform 110 can train the object representation model 210 using the reference object information obtained in block 201 and the reference entity information obtained in block 202. For example, if the object recommendation platform 110 allocates 10,000 exposure opportunities to each object, the last 5,000 exposures can be used as labels for training, and samples can be constructed based on the feature representations generated by the first 5,000 exposures to train the object representation model 210. The following will refer to Figure 3 An example embodiment of training the object representation model 210 is described.

[0043] Continuing with process 200, after obtaining the trained object representation model 210, the object representation model 210 may be further utilized to train an object recommendation model 220. Figure 2As shown, in block 203, the object recommendation platform 110 collects operation information for the sample object to input into the object representation model 210. For example, the operation information may indicate whether the sample object has been transformed by an entity and by what entity group in a certain number of previous recommendations (such as 5,000 times). In some embodiments, if the reference object information includes attribute information of the reference object, the attribute information of the sample object is also input into the object representation model 210.

[0044] The object recommendation platform 110 obtains sample object information of the sample object, which may include operation information and optional attribute information. After inputting the object representation model 210, the object representation model 210 outputs a feature representation of the sample object.

[0045] In some embodiments, at block 204, the object recommendation platform 110 may train an object recommendation model 220 based on the feature representation of the sample object. For example, the feature representation of the sample object may be used as part of the object recommendation model 220, such as replacing the ID of the sample object or as an addition to the ID of the sample object. In addition, the input data may also add similarity information of the feature representation. For example, the similarity between the feature representation of the sample object and the feature representation of other objects that the entity has transformed may be determined as part of the input data. In this way, a trained object recommendation model 204 may be obtained for use by the object recommendation platform 110.

[0046] An example embodiment of training an object representation model using reference object information and reference entity information is described below. Figure 3 A flow chart of an example process 300 for training an object representation model according to some embodiments is shown. Through the process 300, training the object representation model 210 using reference object information and reference entity information may be achieved.

[0047] like Figure 3 As shown, the object recommendation platform 110 can generate a feature representation of the reference object based on the reference object information 302 using the object representation model 210. The reference object information 302 is obtained in block 201, for example.

[0048] In some embodiments, the reference object information 302 may include attribute information 303 (eg, static attribute features) of the reference object and operation information 304 (eg, features acquired by cold start) of the reference object.

[0049] Figure 3The example structure of the object representation model 210 is shown. In this example, the object representation model 210 includes: an embedding layer 314 and a multilayer perceptron (MLP) 315. The MLP 315 may include: a first layer neural network 311, a second layer neural network 312, and a third layer neural network 313.

[0050] In some embodiments, the object recommendation platform 110 may generate a feature representation of the reference entity using an entity representation model based on the reference entity information 301 . The reference entity information 301 is obtained in block 202 , for example. Figure 3 An example structure of an entity representation model 350 is shown. In this example, the entity representation model 350 includes a box embedding layer 354 and an MLP 355. The MLP 355 may include a first layer neural network 351, a second layer neural network 352, and a third layer neural network 353.

[0051] In block 360, the object recommendation platform 110 generates an estimated result 360. The estimated result 360 related to the conversion operation performed by the reference entity on the reference object may be generated based on the feature representation of the reference object output by the object representation model 210 and the feature representation of the reference entity output by the entity representation model 350. For example, the object recommendation platform 110 may perform an inner product of the representation feature representation of the reference entity and the feature representation of the reference object, and then output the estimated result. The estimated result 360 may include, for example, an estimated conversion rate (e.g., click-through rate) or a prediction of whether the reference entity performs a conversion operation, etc.

[0052] Furthermore, the object recommendation platform 110 may update the parameter values ​​of the object representation model based on the difference between the estimated result and the actual processing of the reference object by the reference entity.

[0053] Figure 3 A binary classification model of a double tower structure of entities and objects is shown. However, it should be understood that this is merely exemplary and not intended to be limiting. In addition, Figure 3 The structure of the object representation model shown is also exemplary and is not intended to be limiting. In the embodiments of the present disclosure, any suitable structure may be used to implement the object representation model 210.

[0054] The following describes an example embodiment in which the object recommendation platform 110 performs object recommendation using an object recommendation model.

[0055] Figure 4 A flowchart of an example process 400 for generating recommendation results using an object recommendation model according to some embodiments is shown. Through the process 400, the object recommendation platform 110 can generate recommendation results based on the corresponding object information of the converted object and the corresponding object information of the candidate objects.

[0056] In some embodiments, the object recommendation platform 110 determines a set of corresponding feature representations 420 of the converted objects based on the corresponding object information 410 of the converted objects converted by the target entity. The object information of each converted object is at least related to the conversion operation performed on the converted object in the recommendation of a predetermined number of times (also referred to as the first number). For example, the object information of each converted object may include operation information related to the conversion operation performed on the converted object in the first number of recommendations, such as the features collected when the object is cold started. For example, the operation information may include information about the entity group that performs the conversion operation on the converted object in the first number of recommendations.

[0057] In some embodiments, the object recommendation platform 110 may generate a set of corresponding feature representations 420 of the transformed objects based on the corresponding object information 410 of the transformed objects using the object representation model 210. Figure 4 As shown, the object recommendation platform 110 may input corresponding object information 410 of the transformed object into the object representation model 210. The object representation model 210 generates a feature representation 420 of the transformed object based on the input corresponding object information 410 of the transformed object.

[0058] The object recommendation platform 110 may determine a feature representation 425 of the candidate object based on the object information 415 of the candidate object to be recommended. The object information 415 of the candidate object includes at least operation information related to the conversion operation performed on the candidate object in the recommendation of a predetermined number of times (also referred to as the second number). For example, the operation information may indicate whether the candidate object was converted by an entity and what entity group was converted by the candidate object in the previous 5,000 recommendations.

[0059] In some embodiments, the object recommendation platform 110 may generate a feature representation 425 of the candidate object based on the object information 415 of the candidate object using the object representation model 210. Figure 4 As shown, the object recommendation platform 110 inputs the object information 415 of the candidate object into the object representation model 210. The object representation model 210 generates a feature representation 425 of the candidate object based on the input object information 425 of the candidate object.

[0060] Furthermore, the object recommendation platform 110 generates a recommendation result 430 of the candidate object for the target entity based on a set of corresponding feature representations 420 of the converted object and the feature representations 425 of the candidate object. The object recommendation model 220 generates a recommendation result 430 based on the input feature representations 420 of the converted object and the feature representations 425 of the candidate object. For example, the recommendation result 430 may include or indicate whether the candidate object is recommended to the target entity, if recommended, what is the predicted conversion rate (e.g., click-through rate, purchase rate), and if recommended, where to put it in the recommendation list.

[0061] In some embodiments, the object recommendation platform 110 determines the corresponding similarity between the candidate object and the set of transformed objects based on the corresponding feature representations 420 of the set of transformed objects and the feature representations 425 of the candidate objects. The object recommendation platform 110 generates a recommendation result 430 based on the corresponding feature representations and the corresponding similarities of the set of transformed objects using the object recommendation model 220.

[0062] For example, if objects e, f, and g (which are converted objects) have been clicked by the target entity in the past, the object recommendation platform 110 can compare the feature representation of the candidate object to be recommended (output by the object representation model 210) with the feature representation of objects e, f, and g to calculate the similarity. Then, the calculated similarity can be used as part of the input of the object recommendation model 220.

[0063] In some embodiments, the object recommendation platform 110 determines the recommendation result according to the recommendation metric. The recommendation metric refers to the degree to which the candidate object is recommended to the target entity, that is, to what extent the candidate object is recommended to the target entity. Examples of the recommendation metric may include click-through rate, recommendation score, etc.

[0064] In some embodiments, the object information of each converted object also includes the attribute information of the converted object, such as the static attribute characteristics of the converted object. In some embodiments, the object information 415 of the candidate object also includes the attribute information of the candidate object, such as the static attribute characteristics of the candidate object.

[0065] In some embodiments, the first number, the second number, and the third number may be equal. In this way, the consistency of object information collection can be guaranteed, thereby improving the effect of object representation to obtain a better object recommendation effect.

[0066] In an embodiment of the present disclosure, the object recommendation platform 110 inputs the operation information collected for the sample object into the object representation model, and the object representation model outputs the feature representation of the sample object. The object recommendation platform 110 trains the object recommendation model based on the feature representation of the sample object. Therefore, in a scenario where objects are frequently changed, the object recommendation platform 110 can improve the recommendation effect based on the object recommendation model.

[0067] Example Process

[0068] Figure 5 FIG. 5 is a flowchart of a process 500 for data processing according to some embodiments of the present disclosure. The process 500 may be implemented at the object recommendation platform 110. Figure 1 Process 500 is described.

[0069] In block 510, the object recommendation platform 110 determines a set of corresponding feature representations of the transformed objects based on the corresponding object information of the set of transformed objects transformed by the target entity, wherein the object information of each transformed object is at least related to the conversion operation performed on the transformed object in the first number of recommendations.

[0070] In block 520 , the object recommendation platform 110 determines a feature representation of the candidate object based on the object information of the candidate object to be recommended. The object information of the candidate object includes at least operation information related to the conversion operation performed on the candidate object in the second number of recommendations.

[0071] In block 530 , the object recommendation platform 110 generates a recommendation result of the candidate object for the target entity based on the corresponding feature representations of the set of converted objects and the feature representations of the candidate object.

[0072] In some embodiments, generating a recommendation result of a candidate object for a target entity includes: determining a corresponding similarity between the candidate object and a group of transformed objects based on corresponding feature representations of a group of transformed objects and feature representations of the candidate object; and generating a recommendation result using an object recommendation model based on the corresponding feature representations and corresponding similarities of a group of transformed objects.

[0073] In some embodiments, the object information of each converted object further includes attribute information of the converted object, and wherein the object information of the candidate object further includes attribute information of the candidate object.

[0074] In some embodiments, determining corresponding feature representations of a set of transformed objects includes: generating corresponding feature representations using an object representation model based on corresponding object information of a set of transformed objects, and wherein determining feature representations of candidate objects includes: generating feature representations of candidate objects using an object representation model based on object information of the candidate objects.

[0075] In some embodiments, the object representation model is trained in the following manner: obtaining reference object information of a reference object, the reference object information including at least operation information related to a conversion operation performed on the reference object in a third number of recommendations; obtaining reference entity information of a reference entity of the recommended reference object; and training the object representation model using the reference object information and the reference entity information.

[0076] In some embodiments, training the object representation model includes: generating a feature representation of the reference object using the object representation model based on reference object information; generating a feature representation of the reference entity using the entity representation model based on reference entity information; generating an estimated result related to the transformation operation performed by the reference entity on the reference object based on the feature representation of the reference object and the feature representation of the reference entity; and updating the parameter value of the object representation model based on the difference between the estimated result and the actual processing of the reference object by the reference entity.

[0077] In some embodiments, the reference object information of the reference object also includes attribute information of the reference object.

[0078] In some embodiments, the first number, the second number, and the third number have the same value.

[0079] Example devices and equipment

[0080] Figure 6 A schematic structural block diagram of an apparatus 600 for data processing according to some embodiments of the present disclosure is shown. The apparatus 600 may be implemented as or included in the object recommendation platform 110. Each module / component in the apparatus 600 may be implemented by hardware, software, firmware or any combination thereof.

[0081] As shown in the figure, the apparatus 600 includes a first feature representation determination module 610, which is configured to determine a set of corresponding feature representations of the converted objects based on the corresponding object information of the set of converted objects converted by the target entity. The object information of each converted object is at least related to the conversion operation performed on the converted object in the first number of recommendations.

[0082] The device 600 further includes a first feature representation determination module 620 configured to determine a feature representation of the candidate object based on object information of the candidate object to be recommended. The object information of the candidate object at least includes operation information related to a conversion operation performed on the candidate object in the second recommendation.

[0083] The apparatus 600 further includes a recommendation result generating module 630 configured to generate a recommendation result of the candidate object for the target entity based on a set of corresponding feature representations of the converted objects and feature representations of the candidate objects.

[0084] In some embodiments, the recommendation result generation module 630 is also configured to determine the corresponding similarities between the candidate objects and a group of transformed objects based on the corresponding feature representations of the group of transformed objects and the feature representations of the candidate objects; and to generate recommendation results using the object recommendation model based on the corresponding feature representations and corresponding similarities of the group of transformed objects.

[0085] In some embodiments, the object information of each converted object further includes attribute information of the converted object, and wherein the object information of the candidate object further includes attribute information of the candidate object.

[0086] In some embodiments, the first feature representation determination module 610 is further configured to generate corresponding feature representations based on corresponding object information of a group of converted objects using an object representation model.

[0087] In some embodiments, the second feature representation determination module 620 is further configured to generate a feature representation of the candidate object based on the object information of the candidate object and using the object representation model.

[0088] In some embodiments, the object representation model is trained in the following manner: obtaining reference object information of a reference object, the reference object information including at least operation information related to a conversion operation performed on the reference object in a third number of recommendations; obtaining reference entity information of a reference entity of the recommended reference object; and training the object representation model using the reference object information and the reference entity information.

[0089] In some embodiments, training the object representation model includes: generating a feature representation of the reference object using the object representation model based on reference object information; generating a feature representation of the reference entity using the entity representation model based on reference entity information; generating an estimated result related to the transformation operation performed by the reference entity on the reference object based on the feature representation of the reference object and the feature representation of the reference entity; and updating the parameter value of the object representation model based on the difference between the estimated result and the actual processing of the reference object by the reference entity.

[0090] In some embodiments, the reference object information of the reference object also includes attribute information of the reference object.

[0091] In some embodiments, the first number, the second number, and the third number have the same value.

[0092] Figure 7 A block diagram of an electronic device 700 is shown in which one or more embodiments of the present disclosure may be implemented. It should be understood that Figure 7 The electronic device 700 shown is merely exemplary and should not constitute any limitation on the functionality and scope of the embodiments described herein. Figure 7The electronic device 700 shown can be used to implement Figure 1 An electronic device 110.

[0093] like Figure 7 As shown, the electronic device 700 is in the form of a general electronic device. The components of the electronic device 700 may include, but are not limited to, one or more processors or processing units 710, a memory 720, a storage device 730, one or more communication units 740, one or more input devices 750, and one or more output devices 760. The processing unit 710 may be an actual or virtual processor and is capable of performing various processes according to a program stored in the memory 720. In a multi-processor system, multiple processing units execute computer executable instructions in parallel to improve the parallel processing capability of the electronic device 700.

[0094] The electronic device 700 typically includes a plurality of computer storage media. Such media may be any accessible media that is accessible to the electronic device 700, including but not limited to volatile and non-volatile media, removable and non-removable media. The memory 720 may be a volatile memory (e.g., registers, caches, random access memory (RAM)), a non-volatile memory (e.g., a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof. The storage device 730 may be a removable or non-removable medium, and may include a machine-readable medium, such as a flash drive, a disk, or any other medium, which may be capable of being used to store information and / or data and may be accessed within the electronic device 700.

[0095] The electronic device 700 may further include additional removable / non-removable, volatile / non-volatile storage media. Figure 7 As shown in , a disk drive for reading or writing from a removable, non-volatile disk (e.g., a "floppy disk") and an optical drive for reading or writing from a removable, non-volatile optical disk may be provided. In these cases, each drive may be connected to the bus (not shown) by one or more data media interfaces. The memory 720 may include a computer program product 725 having one or more program modules that are configured to perform various methods or actions of various embodiments of the present disclosure.

[0096] The communication unit 740 implements communication with other electronic devices through a communication medium. Additionally, the functions of the components of the electronic device 700 can be implemented with a single computing cluster or multiple computing machines that can communicate through a communication connection. Therefore, the electronic device 700 can operate in a networked environment using a logical connection with one or more other servers, a network personal computer (PC), or another network node.

[0097] The input device 750 may be one or more input devices, such as a mouse, a keyboard, a tracking ball, etc. The output device 760 may be one or more output devices, such as a display, a speaker, a printer, etc. The electronic device 700 may also communicate with one or more external devices (not shown) through the communication unit 740 as needed, such as a storage device, a display device, etc., communicate with one or more devices that allow a user to interact with the electronic device 700, or communicate with any device that allows the electronic device 700 to communicate with one or more other electronic devices (e.g., a network card, a modem, etc.). Such communication may be performed via an input / output (I / O) interface (not shown).

[0098] According to an exemplary implementation of the present disclosure, a computer-readable storage medium is provided, on which computer-executable instructions are stored, wherein the computer-executable instructions are executed by a processor to implement the method described above. According to an exemplary implementation of the present disclosure, a computer program product is also provided, which is tangibly stored on a non-transitory computer-readable medium and includes computer-executable instructions, and the computer-executable instructions are executed by a processor to implement the method described above.

[0099] Various aspects of the present disclosure are described herein with reference to the flowcharts and / or block diagrams of the methods, devices, equipment, and computer program products implemented according to the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0100] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processing unit of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0101] Computer-readable program instructions can be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, so that the instructions executed on the computer, other programmable data processing apparatus, or other device implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0102] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple implementations of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of a module, program segment or instruction includes one or more executable instructions for realizing the logical function of the specification. In some implementations as replacements, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be realized by a special hardware-based system that performs the function or action of the specification, or can be realized by a combination of special hardware and computer instructions.

[0103] The above descriptions of various implementations of the present disclosure are exemplary, non-exhaustive, and not limited to the disclosed implementations. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described implementations. The selection of terms used herein is intended to best explain the principles of the implementations, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the various implementations disclosed herein.

Claims

1. A data processing method, comprising: Determining corresponding feature representations of a group of transformed objects based on corresponding object information of a group of transformed objects transformed by the target entity, wherein the object information of each transformed object is at least related to a conversion operation performed on the transformed object in a first number of recommendations; Determining a feature representation of the candidate object based on object information of the candidate object to be recommended, wherein the object information of the candidate object at least includes operation information related to a conversion operation performed on the candidate object in a second number of recommendations; as well as Based on the corresponding feature representations of the group of converted objects and the feature representations of the candidate objects, a recommendation result of the candidate objects for the target entity is generated.

2. The method according to claim 1, wherein generating a recommendation result of the candidate object for the target entity comprises: Determining, based on the corresponding feature representations of the set of transformed objects and the feature representations of the candidate objects, corresponding similarities between the candidate objects and the set of transformed objects; and Based on the corresponding feature representations of the group of transformed objects and the corresponding similarities, the recommendation result is generated using an object recommendation model.

3. The method according to claim 2, wherein the object information of each transformed object further includes attribute information of the transformed object, and The object information of the candidate object also includes attribute information of the candidate object.

4. The method of claim 1 , wherein determining corresponding feature representations of the set of transformed objects comprises: Based on the corresponding object information of the set of transformed objects, using the object representation model, generating the corresponding feature representation, and Determining the feature representation of the candidate object includes: Based on the object information of the candidate object, a feature representation of the candidate object is generated using the object representation model.

5. The method according to claim 4, wherein the object representation model is trained by: Acquire reference object information of the reference object, the reference object information at least including operation information related to a conversion operation performed on the reference object in the third number of recommendations; Acquire reference entity information of a reference entity of the recommended reference object; as well as The object representation model is trained using the reference object information and the reference entity information.

6. The method of claim 5, wherein training the object representation model comprises: Based on the reference object information, using the object representation model, generating a feature representation of the reference object; Based on the reference entity information, using an entity representation model, generating a feature representation of the reference entity; generating, based on the feature representation of the reference object and the feature representation of the reference entity, an estimated result related to a transformation operation performed by the reference entity on the reference object; as well as Based on the difference between the estimated result and the actual processing of the reference object by the reference entity, the parameter value of the object representation model is updated. The method according to claim 5 , wherein the reference object information of the reference object further includes attribute information of the reference object.

8. The method of claim 5, wherein the first number, the second number, and the third number have the same value.

9. A data processing device, comprising: A first feature representation determination module is configured to determine corresponding feature representations of a group of transformed objects based on corresponding object information of the group of transformed objects transformed by the target entity, wherein the object information of each transformed object is at least related to a conversion operation performed on the transformed object in a first number of recommendations; A second feature representation determination module is configured to determine a feature representation of a candidate object to be recommended based on object information of the candidate object, wherein the object information of the candidate object at least includes operation information related to a conversion operation performed on the candidate object in a second number of recommendations; as well as The recommendation result generating module is configured to generate a recommendation result of the candidate object for the target entity based on the corresponding feature representations of the group of converted objects and the feature representations of the candidate object.

10. An electronic device comprising: at least one processing unit; as well as At least one memory, the at least one memory being coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, the instructions causing the electronic device to perform the method according to any one of claims 1 to 8 when executed by the at least one processing unit.

11. A computer-readable storage medium having a computer program stored thereon, wherein the computer program can be executed by a processor to implement the method according to any one of claims 1 to 8.

Images