Object processing method and device, equipment and storage medium

Abstract

The embodiment of the invention provides an object processing method and device, equipment and a storage medium. The method comprises the following steps: based on task requirements of a predetermined task, selecting a plurality of objects associated with the predetermined task from a first object set to form a second object set; determining a first aggregation relationship of the plurality of objects in the first object set, the first aggregation relationship indicating that the plurality of objects are aggregated into a plurality of clusters based on the similarity, and each cluster comprising at least one object in the plurality of objects; based on a predetermined aggregation strategy associated with the predetermined task and the first aggregation relationship, determining a second aggregation relationship of the plurality of objects in a second object set; based on the first object set and the second aggregation relation, a target object set matched with the preset task is generated, the target object set comprises a plurality of objects, and the objects at least have the second aggregation relation in the target object set.

Description

Technical Field

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

[0002] With the development of information technology, e-commerce platforms have gradually become an indispensable part of people's daily lives. The relationship between similar product objects belongs to the basic data capabilities of e-commerce platforms. In the process of performing various tasks such as product search and recommendation on e-commerce platforms, they rely on the relationship between similar product objects. However, as the business models of e-commerce platforms become more and more diversified, how to ensure the stability of the relationship between similar product objects while meeting the diversified needs of diversified business models for the relationship between similar product objects has become a problem. Summary of the invention

[0003] In a first aspect of the present disclosure, a method for object processing is provided. The method includes: based on the task requirements of the predetermined task, selecting multiple objects associated with the predetermined task from a first object set to form a second object set; determining a first aggregation relationship of the multiple objects in the first object set, the first aggregation relationship indicating that the multiple objects are aggregated into multiple clusters based on similarity, each cluster including at least one object of the multiple objects; based on a predetermined aggregation strategy associated with the predetermined task and the first aggregation relationship, determining a second aggregation relationship of the multiple objects in the second object set; and based on the first object set and the second aggregation relationship, generating a target object set matching the predetermined task, wherein the target object set includes multiple objects, and the multiple objects have at least a second aggregation relationship in the target object set.

[0004] In a second aspect of the present disclosure, a device for object processing is provided. The device includes: a selection module configured to select multiple objects associated with a predetermined task from a first object set based on the task requirements of the predetermined task to form a second object set; a first determination module configured to determine a first aggregation relationship of multiple objects in the first object set, the first aggregation relationship indicating that the multiple objects are aggregated into multiple clusters based on similarity, each cluster including at least one object from the multiple objects; a second determination module configured to determine a second aggregation relationship of the multiple objects in the second object set based on a predetermined aggregation strategy associated with the predetermined task and the first aggregation relationship; and a generation module configured to generate a target object set matching the predetermined task based on the first object set and the second aggregation relationship, wherein the target object set includes multiple objects, and the multiple objects have at least a second aggregation relationship in the target object set.

[0005] In a third aspect of the present disclosure, an electronic device is provided. The device includes at least one processor; and at least one memory, the at least one memory is coupled to the at least one processor and stores instructions for execution by the at least one processor. When the instructions are executed by the at least one processor, 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 computer-executable instructions are stored on the computer-readable storage medium, and the computer-executable instructions can be executed by a processor to implement the method of the first aspect.

[0007] In a fifth aspect of the present disclosure, a computer executable instruction product is provided, comprising computer executable instructions, wherein when the computer executable instructions are executed by a processor, the method according to the first aspect of the present disclosure is implemented.

[0008] 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

[0009] 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:

[0010] Figure 1 A schematic diagram showing an example environment in which embodiments according to the present disclosure may be implemented;

[0011] Figure 2 A flowchart showing a process for object processing according to some embodiments of the present disclosure is shown;

[0012] Figure 3 A schematic diagram showing an example architecture for object processing according to some embodiments of the present disclosure;

[0013] Figure 4 A schematic diagram showing an example architecture for object processing according to other embodiments of the present disclosure;

[0014] Figure 5 A flowchart illustrating an example process for object processing according to some embodiments of the present disclosure;

[0015] Figure 6 A schematic diagram showing an example scenario for object processing according to some embodiments of the present disclosure;

[0016] Figure 7A schematic structural block diagram showing an example apparatus for object processing according to some embodiments of the present disclosure; and

[0017] Figure 8 A block diagram of an electronic device capable of implementing various embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0018] 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.

[0019] 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 also be included below.

[0020] 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.

[0021] 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.

[0022] 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.

[0023] 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, so that the user can independently choose whether to provide personal information to software or hardware such as electronic devices, applications, servers or storage media that execute operations of the technical solution of the present disclosure based on the prompt message.

[0024] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information is sent to the user in a manner such as a pop-up window, in which the prompt information can be presented in text form. 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.

[0025] 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 relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0026] 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 multi-layer processing units. A neural network model is an example of a model based on deep learning. In this article, "model" may also be referred to as "machine learning model", "learning model", "machine learning network" or "learning network", and these terms are used interchangeably in this article.

[0027] A "neural network" is a machine learning network based on deep learning. A neural network is capable of processing inputs and providing corresponding outputs, and typically includes an input layer and an output layer and one or more hidden layers between the input layer and the output layer. Neural networks used in deep learning applications typically include many hidden layers, thereby increasing the depth of the network. The layers of a neural network are connected in sequence so that the output of the previous layer is provided as input to the next layer, where the input layer receives the input of the neural network and the output of the output layer serves as the final output of the neural network. Each layer of a neural network includes one or more nodes (also called processing nodes or neurons), each of which processes input from the previous layer.

[0028] 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.

[0029] As mentioned above, with the development of information technology, e-commerce platforms have gradually become an indispensable part of people's daily lives. The relationship between similar product objects belongs to the basic data capabilities of e-commerce platforms. In the process of performing various tasks such as product search and recommendation on e-commerce platforms, they all rely on the relationship between similar product objects.

[0030] The association relationship between traditional similar commodity objects can be represented by the aggregation relationship between commodity objects and clusters (also called groups). Specifically, similar commodity objects can be aggregated into the same cluster based on the similarity between commodity objects, forming an aggregation relationship between commodity objects and clusters. In the process of searching for similar commodity objects, the cluster associated with the commodity object being searched can be determined based on the aggregation relationship between commodity objects and clusters. Then, the commodity object associated with the determined cluster is searched.

[0031] This type of aggregation relationship usually only guarantees that the product objects under the same cluster belong to similar product objects, but cannot guarantee that all similar product objects are in one cluster, resulting in the recall rate of similar product objects to be improved. Moreover, the formed aggregation relationship is associated with the tasks of the e-commerce platform. The control test of the aggregation strategy on the formed aggregation relationship is likely to cause unknown impact on the tasks of the e-commerce platform. In addition, this aggregation relationship has high stability, but the demand for diversified associations between similar product objects in different business scenarios is increasing. How to meet the needs of diversified business models for diversified associations between similar product objects on the basis of ensuring the stability of the associations between similar product objects has become a problem.

[0032] In view of this, an embodiment of the present disclosure proposes an improved scheme for object processing. In this scheme, based on the task requirements of the predetermined task, multiple objects associated with the predetermined task are selected from the first object set to form a second object set. A first aggregation relationship of the multiple objects in the first object set is determined, and the first aggregation relationship indicates that the multiple objects are aggregated into multiple clusters based on similarity, and each cluster includes at least one object of the multiple objects. Based on a predetermined aggregation strategy associated with the predetermined task and the first aggregation relationship, a second aggregation relationship of the multiple objects in the second object set is determined. Afterwards, based on the first object set and the second aggregation relationship, a target object set matching the predetermined task is generated, wherein the target object set includes multiple objects, and the multiple objects have at least a second aggregation relationship in the target object set.

[0033] The embodiments of the present disclosure can optimize and adjust the aggregation relationship between objects related to the scheduled task according to the predetermined aggregation strategy associated with the scheduled task, can determine the aggregation relationship that meets the needs of the scheduled task, and can meet the diversified needs of the scheduled tasks corresponding to different business models for the aggregation relationship between similar objects without affecting the original aggregation relationship of the objects, which is conducive to improving the stability of the system.

[0034] Various example implementations of the solution are described in detail below in conjunction with the accompanying drawings.

[0035] Example Environment

[0036] Figure 1 A schematic diagram of an example environment 100 in which an embodiment of the present disclosure can be implemented is shown. In this example environment, an application 125 is installed in a terminal device 120. A user 130 can interact with the application 125 via the terminal device 120 and / or an attached device of the terminal device 120. The application 125 can provide services related to physical objects or virtual objects to the user 130. In some examples, the application 125 can provide services related to e-commerce to the user 130, for example, the application 125 can use the user interface 140 to present the commodity objects queried by the user 130 and the commodity objects recommended to the user 130, and the application 125 can also provide the user 130 with purchase, payment, logistics, return, after-sales and other services of the commodity objects. In other examples, the application 125 can provide services related to objects such as multimedia content, books, advertisements, and travel products to the user 130.

[0037] In an embodiment of the present disclosure, the server device 110 is deployed with a service platform 115, and the service platform 115 can realize the service supply to the application 125 based on the communication connection between the server device 110 and the terminal device 120. In some examples, the service platform 115 may include an e-commerce platform, and the e-commerce platform may determine the commodity objects to be queried by the user and the commodity objects to be recommended by the application 125 based on the aggregation relationship between the commodity objects. The e-commerce platform may provide the commodity objects queried by the user or the commodity objects to be recommended to the application 125 based on the communication connection between the server device 110 and the terminal device 120. The e-commerce platform may also provide support for the purchase, payment, logistics, return, after-sales, etc. of the commodity objects. In other examples, the service platform 115 may also include a service platform that provides multimedia content (such as video, audio, etc.), a service platform that provides paper books or electronic books, a service platform that provides advertising, a service platform that provides travel products, and so on.

[0038] In some embodiments, at least some functions of the service platform 115 or at least some functions of the application 125 may be implemented based on the machine learning model 150. During the operation of the service platform 115, one or more machine learning models 150 may be called. For the convenience of description, one or more machine learning models are collectively referred to as machine learning models 150 herein. Figure 1 The machine learning model 150 is shown as being independent of the server device 110, but one or more machine learning models 150 may run on the server device 110, or other remote servers.

[0039] The machine learning model 150 can be different types of models. In some embodiments, one or more machine learning models 150 can be built based on a language model (LM). The machine learning model used is a content generation model that can generate corresponding outputs based on model inputs. In some embodiments, a machine learning model based on a language model can receive model inputs in textual modalities (e.g., natural language and / or machine language) and / or model inputs in non-textual modalities (e.g., images, voice, video, etc.), and can generate desired outputs based on the model inputs and prompt words. The prompt words here are used to guide the machine learning model to generate model outputs that can solve the user needs indicated by the model inputs.

[0040] In some embodiments, the server device 110 may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content distribution networks, and big data and artificial intelligence platforms. The server device 110 may include, for example, a computing system / server, such as a mainframe, an edge computing node, a computing device in a cloud environment, and the like.

[0041] In some embodiments, the terminal device 120 can be any type of mobile terminal, fixed terminal or portable terminal, including a mobile phone, a desktop computer, a laptop computer, a notebook computer, 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.

[0042] It should be understood that the structure and functionality of environment 100 are described for exemplary purposes only and do not imply any limitation on the scope of the present disclosure.

[0043] Example Process

[0044] Some example embodiments of the present disclosure will be described below with continued reference to the accompanying drawings. Figure 2 A flow chart of a process 200 for object processing according to some embodiments of the present disclosure is shown. Part or all of the process 200 may be implemented by the server device 110 or may be implemented by other devices independent of the server device 110, for example, may be implemented by other remote devices (in a terminal device or a service device) with computing capabilities. In the following, for ease of discussion, the execution of the process 200 is described from the perspective of the server device 110, but this is only exemplary.

[0045] In block 210 , the server device 110 selects a plurality of objects associated with the predetermined task from the first object set based on the task requirement of the predetermined task to form a second object set.

[0046] The object here may include a data object indicating or describing an entity, and the entity may include services, goods, multimedia content, paper books, electronic books, advertisements, travel products, etc. provided by the service platform 115 or the application 125. The data object may include tags, attributes, images, etc. related to the entity. For example, the service platform 115 may be an e-commerce platform, the entity may include goods provided by the e-commerce platform, and the object may include data objects related to the goods provided by the e-commerce platform. The data object may include attribute information of the goods, images of the goods, etc. The attribute information of the goods may include basic attributes of the goods (such as name, number, brand, model, specification, function, origin, etc.), sales attributes, logistics attributes, service attributes (such as after-sales service, installation service, instructions for use), etc.

[0047] The first object set may include objects related to the services or goods provided by the service platform or application. The server-side device 110 may aggregate the objects in the first object set into multiple clusters based on similarity in advance, and obtain relationship data indicating the aggregation relationship between the objects in the first object set and the clusters. Each cluster here may include one object or multiple objects. In one case, the aggregation relationship between the objects in the first object set and the clusters may be referred to as the primary aggregation relationship, and the relationship data indicating the primary aggregation relationship may be referred to as the primary relationship data. Generally speaking, the primary aggregation relationship is usually related to all or most of the objects provided by the service platform 115, has a wider range of influence, and is closely related to the stable operation of the service platform 115 and the application 125.

[0048] As an example, Figure 3 Schematic diagram of an example architecture 300 for object processing according to some embodiments of the present disclosure is shown. Figure 3 As shown, the service platform 115 can be implemented as an e-commerce platform, and the e-commerce platform can include a storage platform 310, a test platform 320, a mining platform 330, a data platform 340, and a task platform 350. The first object set 311 and the main relationship data indicating the aggregation relationship between the objects and clusters in the first object set 311 can be stored in the storage platform 310. The first object set 311 may include objects associated with all or part of the goods provided by the e-commerce platform. The e-commerce platform can pre-group the objects associated with the goods in the first object set 311 into several clusters based on similarity, and obtain the main relationship data indicating the aggregation relationship between the objects and clusters in the first object set 311. The main relationship data can indicate the association relationship between the object number (PID) and the cluster number (CID).

[0049] The scheduled tasks may be tasks related to the entity. In some examples, the scheduled tasks may include tasks related to the goods provided by the e-commerce platform, for example, the scheduled tasks may include recommendation tasks, search tasks, and the like for goods. Task requirements may include requirements for all or part of the information in the data objects related to the entity, for example, the task requirements may include requirements for at least one of the attribute parameters such as the type, brand, model, function, and origin of the goods. It should be noted that the above-mentioned scheduled tasks are only exemplary. In actual applications, the scheduled tasks may include various tasks associated with the services or goods provided by the service platform 115 or the application 125. The embodiments of the present disclosure do not limit the specific content of the task types and task requirements of the scheduled tasks. The second object set can be understood as a set consisting of multiple objects associated with the scheduled tasks.

[0050] Regarding selecting a plurality of objects associated with a predetermined task from a first set of objects according to task requirements, in some examples, Figure 4 Schematic diagram of an example architecture 400 for object processing according to some embodiments of the present disclosure is shown. Figure 4 As shown, the server device 110 selects multiple objects matching the task requirements of the predetermined task from the first object set 311 stored in the storage platform 310 based on the task requirements of the predetermined task to form a second object set 410. While retaining the first object set 311 in the storage platform 310 and the aggregation relationship between the objects and clusters in the first object set 311, the server device 110 can provide the second object set 410 to the transfer service module 331 of the mining platform 330. In this way, the impact on the main aggregation relationship in the storage platform 310 can be reduced, which is conducive to improving system stability.

[0051] Reference Figure 2After determining the second object set, for the multiple objects in the second object set, in block 220, the server device 110 determines a first aggregation relationship of the multiple objects in the first object set. The first aggregation relationship indicates that the multiple objects are aggregated into multiple clusters based on similarity in the original first object set, and each cluster includes at least one object from the multiple objects. The first aggregation relationship here can also be understood as the original aggregation relationship of the multiple objects in the multiple objects (i.e., the second object set) determined from the first object set.

[0052] Regarding the determination of the first aggregation relationship, in some examples, Figure 5 A schematic diagram of an example process 500 for object processing according to some embodiments of the present disclosure is shown. In process 500, the server device 110 may select a group of objects 502 that meet the task requirements from a first object set based on the task requirements of a predetermined task. The server device 110 may select a group of objects 504 from the first object set whose similarity with each object in the group of objects 502 meets a similarity threshold based on, for example, a similarity algorithm. The server device 110 may merge the group of objects 502 and the group of objects 504 to form a second object set 410. In box 506, the server device 110 may determine a first aggregation relationship of multiple objects in the second object set in the first object set. As an example, the server device 110 may determine the object number (PID) of multiple objects in the second object set. The server device 110 may determine the first aggregation relationship of multiple objects in the first object set based on the association relationship between the object number and the cluster number indicated by the primary relationship data and the object number of multiple objects in the second object set. Of course, the server device 110 may also be based on the first aggregation relationship of multiple objects in the first object set, such as object name, object number, object description information, etc., and the embodiments of the present disclosure are not limited to this.

[0053] It is understandable that the above process of selecting multiple objects and determining the first aggregation relationship is only exemplary, and any other appropriate method can be selected according to actual needs to select objects associated with the predetermined task from the first object set. The embodiments of the present disclosure are not limited to this.

[0054] Continue to refer back Figure 2In box 230, the server device 110 determines a second aggregation relationship of multiple objects in a second object set based on a predetermined aggregation strategy associated with a predetermined task and a first aggregation relationship of multiple objects in a first object set. In some embodiments, the predetermined aggregation strategy may include a strategy for determining an aggregation relationship between multiple objects in a second object set. The server device 110 may adjust the aggregation relationship between at least some of the objects indicated by the first aggregation relationship and the cluster based on the predetermined aggregation strategy to form a second aggregation relationship. In other words, the second aggregation relationship may be understood as an aggregation relationship formed after adjusting the aggregation relationship between at least some of the objects indicated by the first aggregation relationship and the cluster.

[0055] In some embodiments, the predetermined aggregation strategy may include a selection strategy for a target cluster and a non-target cluster, wherein the target cluster is a cluster into which an object needs to be transferred, and the non-target cluster is a cluster from which an object needs to be transferred out. The predetermined aggregation strategy may also include a determination strategy for the first similarity between an object and a cluster, and a first predetermined condition indicating whether to aggregate an object to a cluster. The server device 110 determines a target cluster and a non-target cluster from a plurality of clusters indicated by the first aggregation relationship based on the predetermined aggregation strategy. Determine the first similarity between each object in the non-target cluster and the target cluster. Thereafter, determine whether the first similarity meets the first predetermined condition. If it is determined that the first similarity between an object in the non-target cluster and the target cluster does not meet the first predetermined condition, the server device 110 may not change the aggregation relationship between the object and the non-target cluster. If it is determined that the first similarity between an object in the non-target cluster and the target cluster meets the first predetermined condition, the server device 110 may transfer the object from the non-target cluster to the target cluster to form a second aggregation relationship. In some examples, the first predetermined condition may include a first threshold set for the first similarity, and the first threshold indicates whether to aggregate the object to the target cluster. Of course, the first predetermined condition may also include any other appropriate conditions, which is not limited in the embodiments of the present disclosure.

[0056] In some embodiments, the server device 110 sorts the multiple clusters based at least on the number of objects contained in each of the multiple clusters indicated by the first aggregation relationship. The server device 110 may select the first cluster with a higher ranking from the sorting as the target cluster, and select the second cluster with a lower ranking from the sorting as the non-target cluster. In some examples, the server device 110 may sort the multiple clusters indicated by the first aggregation relationship in a manner from large to small in terms of the number of objects contained in the clusters. In this way, a cluster with a relatively large number of objects may be selected as the target cluster, and a cluster with a relatively small number of objects may be selected as the non-target cluster. Objects may be transferred from a cluster with a relatively small number of objects to a cluster with a relatively large number of objects, which is beneficial to improving the aggregation degree of similar objects, and is beneficial to reducing duplicate clusters in the second object set, thereby facilitating improving indicators such as the recall rate and accuracy rate of objects. The duplicate clusters here may be understood as at least some of the objects in the two clusters being similar objects or the same objects.

[0057] Continue to refer Figure 5 In process 500, in box 508, the server device 110 can sort the multiple clusters according to the number of objects contained in each of the multiple clusters and the construction time of each of the multiple clusters. As an example, the server device 110 can sort the multiple clusters according to the number of objects contained in the multiple clusters. In the case where there are clusters with the same number of objects in the multiple clusters, the server device 110 can sort the clusters with the same number of objects according to the construction time of the clusters with the same number of objects. Of course, the above sorting method is only exemplary, and any other appropriate method can be selected to sort the clusters according to actual needs, and the embodiments of the present disclosure are not limited to this.

[0058] Regarding the selection of target clusters and non-target clusters, in some examples, such as Figure 5As shown, in box 510, the server device 110 can determine the first cluster in the sort as the target cluster based on the first selection strategy, and determine one or more clusters located after the first cluster in the sort as non-target clusters. In other examples, in box 512, the server device 110 can determine the cluster at the end of the sort as a non-target cluster based on the second selection strategy, and determine one or more clusters before the cluster at the end of the sort as the target cluster. For example, the sorting may include cluster C1, cluster C2, cluster C3, ..., cluster CN, where N is a positive integer. The server device 110 can divide clusters C1, cluster C2, cluster C3, ..., cluster CN into two groups according to the sorting, one group includes cluster C1 to cluster CK, and the other group includes cluster CK+1 to cluster CN, where K<N, and K is a positive integer. The server device 110 can determine one or more clusters from cluster C1 to cluster CK as the target cluster. The server device can determine cluster CN as a non-target cluster. It can be understood that the first selection strategy and the second selection strategy are different strategies. The server device 110 may select to perform the operation of box 510 or the operation of box 512 according to the configuration.

[0059] After the server device 110 determines the target cluster and the non-target cluster based on the first selection strategy or the second selection strategy, the server device 110 continues to perform operations from box 514 to box 520. In box 514, the server device 110 determines the first similarity between the object in the non-target cluster and the target cluster. As an example, the server device 110 can determine the first similarity between the object in each non-target cluster and the target cluster in reverse order of the sorting of the multiple non-target clusters. As another example, the server device 110 can determine the first similarity between the object in the non-target cluster and each target cluster according to the sorting of the target cluster.

[0060] In box 516, the server device 110 determines whether the first similarity exceeds the first threshold. If it is determined in box 516 that the first similarity does not exceed the first threshold, the process 500 proceeds to box 518. In box 518, the server device 110 does not change the aggregation relationship between the corresponding object and the non-target cluster. As an example, in the case of determining multiple target clusters, the server device 110 can determine whether the first similarity between the object in the non-target cluster and the first target cluster in the sorting exceeds the first threshold according to the sorting of the target clusters. If it is determined in box 516 that the first similarity does not exceed the first threshold, the server device 110 determines whether the first similarity between the object in the non-target cluster and the second target cluster in the sorting exceeds the first threshold according to the sorting of the target clusters, and so on. If the first similarity between an object in the non-target cluster and all the target clusters does not exceed the first threshold, the process 500 proceeds to box 518.

[0061] If it is determined in block 516 that the first similarity exceeds the first threshold, process 500 proceeds to block 520. In block 520, the server device 110 transfers the corresponding object from the non-target cluster to the corresponding target cluster, for example, if it is determined that the first similarity between the object in the non-target cluster and the second target cluster in the sorting exceeds the first threshold, the server device 110 may transfer the object from the non-target cluster to the second target cluster.

[0062] In some embodiments, the server device 110 may determine object feature information of objects in the non-target cluster and cluster feature information of the target cluster according to a predetermined aggregation strategy. Based on the object feature information and the cluster feature information, a first similarity between the objects in the non-target cluster and the target cluster is determined. Specifically, the predetermined aggregation strategy may indicate a feature extraction strategy for objects in the non-target cluster and the target cluster. The server device 110 may perform feature extraction on objects in the non-target cluster according to the feature extraction strategy to obtain object feature information. The server device 110 may also perform feature extraction on objects in the non-target cluster according to the feature extraction strategy to obtain cluster feature information.

[0063] As an example, the service platform 115 may be an e-commerce platform, and the object may be a commodity provided by the e-commerce platform. The server device 110 may perform feature extraction on multimodal information such as text, images, videos, and audio related to the commodity in the object according to the feature extraction strategy to obtain object feature information. The server device 110 may perform feature extraction on the target cluster to obtain cluster feature information of the target cluster. The server device 110 may also perform feature extraction on each object in the target cluster, and use the extracted features as a component of the cluster feature information of the target cluster.

[0064] As another example, Figure 3 As shown, the test platform 320 may include a similarity service module 321. The server device 110 may call the similarity service module 321 using the transfer service module 331. Feature extraction is performed on the objects in the non-target cluster and the target cluster via the similarity service module 321 to obtain object feature information and cluster feature information. Afterwards, the similarity service module 321 may be used to determine the first similarity between the objects in the non-target cluster and the target cluster based on the object feature information and the cluster feature information.

[0065] In some embodiments, the server device 110 generates prompt information for the machine learning model 150 based on the object feature information and the cluster feature information. The prompt information is input into the machine learning model 150 to obtain the model output of the machine learning model 150. Afterwards, the server device 110 may determine the first similarity between the object in the non-target cluster and the target cluster based on the model output. As an example, the predetermined aggregation strategy may indicate a generation strategy for the prompt information of the machine learning model 150. Figure 4 As shown, the server device 110 can use the similarity service module 321 to generate prompt information for the machine learning model 150 according to a predetermined aggregation strategy, object feature information and cluster feature information. The prompt information is provided to the machine learning model 150 using the similarity service module 321 to obtain the model output of the machine learning model 150. The similarity transfer module 321 is used to determine the first similarity between the object in the non-target cluster and the target cluster based on the model output. If it is determined that the first similarity exceeds the first threshold, the transfer service module 331 is used to migrate the corresponding object from the non-target cluster to the target cluster to form a second aggregation relationship. The use of the machine learning model 150 can not only quickly determine the similarity between objects and clusters, but also improve the accuracy of the similarity determination with the help of the understanding ability of the machine learning model 150.

[0066] In some embodiments, the server device 110 determines a second similarity between an object in a non-target cluster and an object in a target cluster. A group of objects whose second similarity meets a second predetermined condition is selected from the non-target cluster, and object feature information of each object in the group of objects is determined. Afterwards, the server device 110 may generate prompt information for the machine learning model 150 based on the object feature information of each object in the group of objects and the cluster feature information of the target cluster. In some examples, the second predetermined condition may include a second threshold value set for the second similarity. Of course, the second predetermined condition may also include any other appropriate condition, and should not be understood as being limited to the second threshold value.

[0067] As an example, Figure 6 Schematic diagram of an example scene 600 for object processing according to some embodiments of the present disclosure is shown. Figure 6 As shown, in box 610, when the non-target cluster and the target cluster have been determined, the server device 110 can select object P3 from the target cluster C2. The server device 110 can determine the second similarity between object P3 and each object in the second object set, and the server device 110 can select objects P1, P2, P4, and P5 whose second similarity meets the second threshold from the second object set.

[0068] In box 620, the server device 110 can determine that objects P4, P5 and object P3 all belong to the target cluster C2, and objects P1 and P2 belong to the non-target cluster C1 based on the first aggregation relationship. The server device 110 can perform feature extraction on objects P1 and P2 to obtain object feature information associated with objects P1 and P2. The server device 110 can also perform feature extraction on the target cluster C2 to obtain cluster feature information associated with the target cluster C2. Based on the object feature information associated with objects P1 and P2 and the cluster feature information associated with the target cluster C2, prompt information for the machine learning model 150 is generated. The prompt information is provided to the machine learning model 150 through the similarity service module 321 to obtain the model output of the machine learning model 150.

[0069] In block 630, the similarity service module 321 is used to determine the first similarity between the objects P1, P2 and the target cluster C2 based on the model output. If it is determined that the first similarity between the object P1 and the target cluster C2 does not exceed the first threshold, the server device 110 may not change the aggregation relationship between the object P1 and the non-target cluster C1. If it is determined that the first similarity between the object P2 and the target cluster C2 exceeds the first threshold, the transfer module 331 may be used to migrate the object P2 to the target cluster C2.

[0070] In block 640, the server device 110 uses the transfer service module 331 to obtain the updated aggregation relationship between the objects and the clusters, that is, the second aggregation relationship. Specifically, object P1 is aggregated to cluster C1, and objects P2, P3, P4, and P5 are aggregated to cluster C2. This is conducive to improving the aggregation degree between similar objects in the second object set, thereby improving the recall rate and accuracy of similar objects, and also reducing the task load of the machine learning model 150 to reduce the consumption of system resources.

[0071] It is understandable that the server device 110 may repeatedly execute the process shown in box 610 to box 640 for the same target cluster or different target clusters to gradually migrate objects in the non-target cluster to the target cluster, thereby improving the degree of aggregation between similar objects in the second object set. It should also be noted that the above method for determining the first similarity is only exemplary. In actual application, any other appropriate method can be selected according to actual needs to determine the first similarity between the objects in the non-target cluster and the target cluster. The embodiments of the present disclosure are not limited to this.

[0072] Returning to process 200, in box 240, the server device 110 generates a target object set that matches the scheduled task based on the first object set and the second aggregation relationship. The target object set includes multiple objects associated with the scheduled task. In some examples, the target object set may also include other objects in the first object set, for example, the target object set may include all objects in the first object set. Here, the multiple objects associated with the scheduled task have at least the second aggregation relationship in the target object set.

[0073] In some embodiments, Figure 4 As shown, in box 430, the server device 110 can perform a JOIN operation on the first object set and the second object set to generate a target object set 341 that matches the predetermined task. In some examples, when performing the JOIN operation, the server device 110 can use the second aggregation relationship to cover the first aggregation relationship of the multiple objects associated with the predetermined task in the first object set to generate a target object set that matches the predetermined task. In this way, the multiple objects associated with the predetermined task have the second aggregation relationship in the target object set, but do not have the first aggregation relationship.

[0074] In other examples, the server device 110 may retain the first aggregation relationship and the second aggregation relationship during the connection operation. In this way, multiple objects associated with the predetermined task have not only the second aggregation relationship but also the first aggregation relationship in the target object set. It should be understood that the above method for generating a target object set is only exemplary. In actual application, other appropriate methods can be selected to generate a target object set. For example, the server device 110 can also select some objects from the first object set to generate a target object set. The embodiments of the present disclosure are not limited to this.

[0075] In some embodiments, after the target object set, the server device 110 can match the target object set to a predetermined task. The server device 110 can perform the predetermined task based on the target object set. Figure 3 As shown, the server device 110 can save the target object set 341 to the data platform 340 without modifying the main relationship data in the storage platform 310. The task platform 350 can use the corresponding target object set 341 in the data platform 340 to perform a predetermined task, including but not limited to a commodity recommendation task, a search task, a deduplication task, etc. In this way, the aggregation relationship between objects and clusters can be adjusted and tested for the predetermined task without affecting the main aggregation relationship.

[0076] As an example, the predetermined task may include a recommendation task. The application 125 may present the target product using the user interface 140, and the user interface 140 may include a first area for presenting products similar to the target product. The server device 110 may determine, based on the target object set, multiple products that are aggregated into the same cluster as the target product using the task module 350, and present recommendation information of the multiple products that are aggregated into the same cluster as the target product through the first area.

[0077] As another example, the predetermined task may include a search task. The user may input a search term into the application 125 using the terminal device 120 or an attached device of the terminal device 120. The server device 110 retrieves at least one target commodity matching the search term from the target object set, and the server device 110 may also determine commodities aggregated into the same cluster as the at least one target commodity. The server device 110 may present the at least one target commodity and commodities aggregated into the same cluster as the at least one target commodity to the user through the application 125.

[0078] It is understandable that the above-mentioned scheduled tasks are only exemplary. In actual application, the scheduled tasks may include various tasks related to the services or commodities provided by the service platform 115 or the application 125, and the embodiments of the present disclosure do not limit the types of scheduled tasks.

[0079] In some examples, such as Figure 3 and Figure 4 As shown, the server device 110 can generate multiple associated target object sets 341-1, 341-2, ..., 341-M for multiple scheduled tasks 351-1, 351-2, ..., 351-M, where M is a positive integer. The server device 110 can save the multiple target object sets 341-1, 341-2, ..., 341-M to the data platform 340. The task platform 350 can use the corresponding target object set 341 in the data platform 340 to perform the corresponding scheduled task. For example, the task platform 350 can use the target object set 341-1 to perform a commodity search task, can use the target object set 341-2 to perform a commodity recommendation task, and so on.

[0080] In some embodiments, before generating the target object set, the server device 110 performs an evaluation operation on the aggregation effect of the second aggregation relationship to obtain an evaluation result of the second aggregation relationship. The server device 110 determines whether the evaluation result of the second aggregation relationship meets the predetermined aggregation requirements. If it is determined that the evaluation result of the second aggregation relationship meets the predetermined aggregation requirements, the server device 110 generates a target object set that matches the predetermined task based on the first object set and the second aggregation relationship. If it is determined that the evaluation result of the second aggregation relationship does not meet the predetermined aggregation requirements, the server device 110 can adjust the predetermined aggregation strategy associated with the predetermined task and repeat at least part of the process in box 210 to box 230 to generate a second aggregation relationship whose evaluation result meets the predetermined aggregation requirements.

[0081] In some examples, at least some objects may be selected from the second object set in advance, and the predetermined aggregation relationship of at least some objects may be predetermined. For example, the aggregation relationship between at least some objects and clusters may be manually annotated to form a predetermined aggregation relationship. The server device 110 may determine at least one evaluation index of the recall rate, accuracy rate, macro-average, or micro-average of the second aggregation relationship according to the predetermined aggregation relationship to generate an evaluation result for the second aggregation relationship. Correspondingly, the predetermined aggregation requirement may include a requirement for at least one evaluation index of the recall rate, accuracy rate, macro-average, or micro-average, for example, the predetermined aggregation requirement may include an index threshold for the at least one evaluation index. The server device 110 may determine whether the recall rate, accuracy rate, macro-average, or micro-average of the second aggregation relationship exceeds the corresponding index threshold. If the at least one evaluation index exceeds the corresponding index threshold, the server device 110 determines that the evaluation result of the second aggregation relationship meets the predetermined aggregation requirement. If the at least one evaluation index does not exceed the corresponding index threshold, the server device 110 determines that the evaluation result of the second aggregation relationship does not meet the predetermined aggregation requirement.

[0082] In other examples, the server device 110 may use the second object set to perform a predetermined task for a small sample size of users to evaluate the aggregation effect of the second aggregation relationship. Specifically, the server device 110 may perform a predetermined task based on the second object set to obtain the execution result of the predetermined task. According to the execution result of the predetermined task, it is determined whether the second aggregation relationship meets the predetermined aggregation requirement. As an example, the service platform 115 may be an e-commerce platform, and the multiple objects in the second object set may be objects associated with the goods provided by the e-commerce platform. The server device 110 may perform predetermined tasks such as recommendation, search, and deduplication of goods based on the second object set, and determine task indicators such as the click-through rate and conversion rate of the goods. Correspondingly, the predetermined aggregation requirement may include task requirements for task indicators such as the click-through rate and conversion rate of the goods. The server device 110 may determine whether the task indicator meets the task requirement, for example, determining whether the click-through rate and conversion rate of the goods exceed the corresponding threshold. If it is determined that the task indicator meets the task requirement, the server device 110 may determine that the second aggregation relationship meets the predetermined aggregation requirement. If it is determined that the task indicator does not meet the task requirement, the server device 110 may determine that the second aggregation relationship part meets the predetermined aggregation requirement. It should be noted that the above method for evaluating the second aggregation relationship is only exemplary, and any other appropriate method may be selected to evaluate the second aggregation relationship according to actual needs, and the embodiments of the present disclosure are not limited in this regard.

[0083] In some embodiments, the server device 110 may determine to perform an evaluation operation on the aggregation effect of a predetermined aggregation strategy and obtain an evaluation result of the predetermined aggregation strategy. If the evaluation result of the predetermined aggregation strategy meets the predetermined policy requirements, the server device 110 may update the aggregation relationship between the objects and clusters included in the first object set based on the predetermined aggregation strategy. In some examples, the server device 110 may determine the evaluation result of the predetermined aggregation strategy based on the second object set or the target object set. As an example, the server device 110 may use a pre-built predetermined aggregation relationship to determine at least one evaluation indicator of the recall rate, accuracy rate, macro average, and micro average of the second object set or the target object set. If the at least one evaluation indicator exceeds the corresponding indicator threshold, the server device 110 determines that the evaluation result of the predetermined aggregation strategy meets the predetermined policy requirements.

[0084] In some examples, the server device 110 can use the target object set to perform a predetermined task and obtain the execution result of the predetermined task. According to the execution result of the predetermined task, the evaluation result of the predetermined aggregation strategy is determined. As an example, the server device 110 can use the target object set to perform predetermined tasks such as recommendation and search for the goods provided by the e-commerce platform, and determine task indicators such as the click-through rate, conversion rate, and repurchase rate of the goods. If the task indicator meets the task requirements, the server device 110 can determine that the evaluation result of the predetermined aggregation strategy meets the predetermined strategy requirements.

[0085] In some examples, the server device 110 can respectively use the first object set and the target object set to perform the predetermined task and obtain the first execution result and the second execution result. The server device 110 can use the first execution result to determine whether the second execution result meets the task requirements of the predetermined task. If the second execution result meets the task requirements of the predetermined task, the server device 110 can determine that the predetermined aggregation strategy meets the predetermined strategy requirements. In this way, a comparative test of the predetermined aggregation strategy can be implemented.

[0086] As an example, the server device 110 can use the first object set to perform predetermined tasks such as searching, recommending, and deduplicating products, and obtain task indicators such as the click-through rate, conversion rate, and repurchase rate of the products as the first execution result. The server device 110 can also use the target object set to perform predetermined tasks such as searching, recommending, and deduplicating products, and obtain task indicators such as the click-through rate, conversion rate, and repurchase rate of the products as the second execution result. The server device 110 can determine whether the task indicators such as the click-through rate, conversion rate, and repurchase rate indicated by the second execution result exceed the task indicators indicated by the first execution result. If the task indicators indicated by the second execution result exceed the task indicators indicated by the first execution result, the server device 110 can determine that the predetermined aggregation strategy meets the predetermined strategy requirements.

[0087] As another example, the server device 110 determines the difference between the task indicator indicated by the second execution result and the task indicator indicated by the first execution result, for example, determining that the difference between the task indicators such as click-through rate, conversion rate, and repurchase rate exceeds the corresponding difference threshold. If the difference exceeds the difference threshold, the server device 110 can determine that the predetermined aggregation strategy meets the predetermined strategy requirements.

[0088] In some examples, multiple predetermined aggregation strategies can be pre-built for a predetermined task. The server device 110 can determine multiple target object sets according to the multiple predetermined aggregation strategies. The server device 110 uses the multiple target object sets to respectively perform the predetermined tasks and obtain the evaluation results corresponding to each of the multiple predetermined aggregation strategies. The server device 110 can determine the predetermined aggregation strategy whose evaluation results meet the requirements of the predetermined strategy according to the evaluation results of the multiple predetermined aggregation strategies, so as to implement a comparative test of the predetermined aggregation strategy. As an example, the server device 110 can use multiple target object sets to perform tasks related to commodities provided by an e-commerce platform, for example, and respectively determine the task indicators such as the click-through rate, conversion rate, and repurchase rate of the commodities. The server device 110 can determine that the predetermined aggregation strategy corresponding to a certain highest single task indicator is determined as the predetermined aggregation strategy that meets the requirements of the predetermined strategy. The server device 110 can also determine the comprehensive scores of the multiple predetermined aggregation strategies based on multiple task indicators. The server device 110 can determine the predetermined aggregation strategy corresponding to the highest comprehensive score as the predetermined aggregation strategy that meets the requirements of the predetermined strategy.

[0089] In this way, the embodiments of the present disclosure can optimize and adjust the aggregation relationship between objects related to the scheduled task according to the predetermined aggregation strategy associated with the scheduled task, can produce an aggregation relationship that meets the requirements of the scheduled task, and can meet the diversified requirements of the scheduled tasks corresponding to different business models for the aggregation relationship between similar commodity objects.

[0090] Example devices and equipment

[0091] The embodiments of the present disclosure also provide corresponding devices for implementing the above methods or processes. Figure 7 A schematic structural block diagram of an example apparatus 700 for object processing according to some embodiments of the present disclosure is shown. The apparatus 700 may be implemented as or included in the server device 110. Each module / component in the apparatus 700 may be implemented by hardware, software, firmware, or any combination thereof.

[0092] like Figure 7As shown, the device 700 includes: a selection module 710, configured to select multiple objects associated with a predetermined task from a first object set based on the task requirements of the predetermined task to form a second object set; a first determination module 720, configured to determine a first aggregation relationship of the multiple objects in the first object set, the first aggregation relationship indicating that the multiple objects are aggregated into multiple clusters based on similarity, and each cluster includes at least one object of the multiple objects; a second determination module 730, configured to determine a second aggregation relationship of the multiple objects in the second object set based on a predetermined aggregation strategy associated with the predetermined task and the first aggregation relationship; and a generation module 740, configured to generate a target object set matching the predetermined task based on the first object set and the second aggregation relationship, wherein the target object set includes multiple objects, and the multiple objects have at least a second aggregation relationship in the target object set.

[0093] In some embodiments, the second determination module 730 is further configured to: determine a target cluster and a non-target cluster from multiple clusters indicated by the first aggregation relationship; determine a first similarity between objects in the non-target cluster and the target cluster based on a predetermined aggregation strategy; and in response to the first similarity meeting a first predetermined condition, transfer at least one object in the non-target cluster to the target cluster to form a second aggregation relationship.

[0094] In some embodiments, the second determination module 730 is further configured to: sort the multiple clusters based at least on the number of objects contained in each of the multiple clusters indicated by the first aggregation relationship; select the first cluster with the highest ranking from the sorting as the target cluster; and select the second cluster with the lowest ranking from the sorting as the non-target cluster.

[0095] In some embodiments, the second relationship determination module 730 is further configured to sort the multiple clusters based on the number of objects contained in each of the multiple clusters and the construction time of each of the multiple clusters.

[0096] In some embodiments, the second determination module 730 is further configured to: determine the object feature information of the objects in the non-target cluster and the cluster feature information of the target cluster according to a predetermined aggregation strategy; and determine the first similarity between the objects in the non-target cluster and the target cluster based on the object feature information and the cluster feature information.

[0097] In some embodiments, the second determination module 730 is further configured to: generate prompt information for the machine learning model based on object feature information and cluster feature information; input the prompt information into the machine learning model to obtain a model output of the machine learning model; and determine the first similarity between the objects in the non-target cluster and the target cluster based on the model output.

[0098] In some embodiments, the second determination module 730 is further configured to: determine a second similarity between objects in the non-target cluster and objects in the target cluster; select a group of objects from the non-target cluster whose second similarity meets a second predetermined condition; and determine object feature information of each object in a group of objects.

[0099] In some embodiments, the second determination module 730 is further configured to: determine whether at least one object in the non-target cluster includes a seed object of the non-target cluster; based on determining that at least one object includes a seed object, remove the non-target cluster from the second object set; and based on determining that at least one object does not include a seed object, retain the non-target cluster in the second object set.

[0100] In some embodiments, the generation module 740 is further configured to: in response to the evaluation result of the second aggregation relationship meeting the predetermined aggregation requirement, generate a target object set matching the predetermined task based on the first object set and the second aggregation relationship.

[0101] In some embodiments, the plurality of objects further have a first aggregation relationship in the target object set.

[0102] In some embodiments, the apparatus 700 further includes: an updating module configured to update the aggregation relationship between the objects and clusters included in the first object set based on the predetermined aggregation strategy in response to an evaluation result of the predetermined aggregation strategy meeting the predetermined strategy requirement.

[0103] The units and / or modules included in the device 700 can be implemented in various ways, including software, hardware, firmware, or any combination thereof. In some embodiments, one or more units and / or modules can be implemented using software and / or firmware, such as machine executable instructions stored on a storage medium. In addition to or as an alternative to machine executable instructions, some or all of the units and / or modules in the device 700 can be implemented at least in part by one or more hardware logic components. As an example and not limitation, exemplary types of hardware logic components that can be used include field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0104] Figure 8 8 is a block diagram of an electronic device 800 in which one or more embodiments of the present disclosure may be implemented. It should be understood that Figure 8 The electronic device 800 shown is merely exemplary and should not constitute any limitation on the functionality and scope of the embodiments described herein. Figure 8 The electronic device 800 shown may include or be implemented as Figure 1 The server device 110, or Figure 7 Device 700.

[0105] like Figure 8 As shown, the electronic device 800 is in the form of a general electronic device. The components of the electronic device 800 may include, but are not limited to, one or more processors 810, a memory 820, a storage device 830, one or more communication units 840, one or more input devices 850, and one or more output devices 860. The processor 810 may be an actual or virtual processor and is capable of performing various processes according to executable instructions stored in the memory 820. In a multi-processor system, multiple processors execute computer executable instructions in parallel to improve the parallel processing capability of the electronic device 800.

[0106] The electronic device 800 typically includes a plurality of computer storage media. Such media can be any accessible media that can be obtained by the electronic device 800, including but not limited to volatile and non-volatile media, removable and non-removable media. The memory 820 can be a volatile memory (e.g., a register, a cache, a 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 830 can be a removable or non-removable medium, and can include a machine-readable medium, such as a flash drive, a disk, or any other medium, which can be used to store information and / or data and can be accessed within the electronic device 800.

[0107] The electronic device 800 may further include additional removable / non-removable, volatile / non-volatile storage media. Figure 8 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 a bus (not shown) by one or more data media interfaces. The memory 820 may include a computer executable instruction product 825 having one or more executable instruction modules that are configured to perform various methods or actions of various embodiments of the present disclosure.

[0108] The communication unit 840 implements communication with other electronic devices through a communication medium. Additionally, the functions of the components of the electronic device 800 can be implemented with a single computing cluster or multiple computing machines that can communicate through a communication connection. Therefore, the electronic device 800 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.

[0109] The input device 850 may be one or more input devices, such as a mouse, a keyboard, a tracking ball, etc. The output device 860 may be one or more output devices, such as a display, a speaker, a printer, etc. The electronic device 800 may also communicate with one or more external devices (not shown) through the communication unit 840 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 800, or communicate with any device that allows the electronic device 800 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).

[0110] 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-executable instruction 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.

[0111] Various aspects of the present disclosure are described herein with reference to flowcharts and / or block diagrams of methods, devices, equipment, and computer-executable instruction 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 boxes in the flowchart and / or block diagram can be implemented by computer-readable executable instructions.

[0112] These computer executable instructions can be provided to a processor 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 processor 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 executable 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.

[0113] Computer-executable 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.

[0114] The flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the systems, methods and computer executable instruction products according to multiple implementations of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, an executable instruction or a part of an instruction, and a module, an executable instruction or a part of an instruction contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0115] 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 method for object processing, comprising: Based on the task requirement of the predetermined task, selecting a plurality of objects associated with the predetermined task from the first object set to form a second object set; Determine a first aggregation relationship of the plurality of objects in the first object set, the first aggregation relationship indicating that the plurality of objects are aggregated into a plurality of clusters based on similarity, each cluster including at least one object among the plurality of objects; Determining a second aggregation relationship of the plurality of objects in the second object set based on a predetermined aggregation strategy associated with the predetermined task and the first aggregation relationship; as well as Based on the first object set and the second aggregation relationship, a target object set matching the predetermined task is generated, wherein the target object set includes the plurality of objects, and the plurality of objects have at least the second aggregation relationship in the target object set.

2. The method according to claim 1, wherein determining the second aggregation relationship comprises: Determine a target cluster and a non-target cluster from the multiple clusters indicated by the first aggregation relationship; Based on the predetermined aggregation strategy, determining a first similarity between the objects in the non-target cluster and the target cluster; as well as In response to the first similarity meeting a first predetermined condition, at least one object in the non-target cluster is transferred to the target cluster to form the second aggregation relationship.

3. The method according to claim 2, wherein determining the target cluster and the non-target cluster comprises: sorting the multiple clusters based at least on the number of objects contained in each of the multiple clusters indicated by the first aggregation relationship; Selecting the first cluster with the highest ranking from the ranking as the target cluster; as well as A second cluster ranked later in the ranking is selected as the non-target cluster.

4. The method of claim 3, wherein sorting the plurality of clusters comprises: The plurality of clusters are sorted based on the number of objects contained in each cluster and the construction time of each cluster in the plurality of clusters.

5. The method according to claim 2, wherein determining a first similarity between the objects in the non-target cluster and the target cluster comprises: Determining object feature information of objects in the non-target cluster and cluster feature information of the target cluster according to the predetermined aggregation strategy; as well as Based on the object feature information and the cluster feature information, a first similarity between the objects in the non-target cluster and the target cluster is determined.

6. The method according to claim 5, wherein determining the first similarity between the objects in the non-target cluster and the target cluster based on the object feature information and the cluster feature information comprises: Based on the object feature information and the cluster feature information, generating prompt information for a machine learning model; Inputting the prompt information into the machine learning model to obtain a model output of the machine learning model; and Based on the model output, a first similarity between the objects in the non-target cluster and the target cluster is determined.

7. The method according to claim 5, wherein determining the object feature information of the objects in the non-target cluster comprises: Determining a second similarity between the objects in the non-target cluster and the objects in the target cluster; Selecting a group of objects whose second similarity meets a second predetermined condition from the non-target cluster; and Object feature information of each object in the set of objects is determined.

8. The method according to claim 2, wherein determining the second aggregation relationship further comprises: determining whether the at least one object in the non-target cluster includes a seed object of the non-target cluster; removing the non-target cluster from the second set of objects based on determining that the at least one object includes the seed object; as well as Based on determining that the at least one object does not include the seed object, retaining the non-target cluster in the second set of objects.

9. The method according to claim 1, wherein generating a set of target objects matching the predetermined task comprises: In response to the evaluation result of the second aggregation relationship meeting the predetermined aggregation requirement, the target object set matching the predetermined task is generated based on the first object set and the second aggregation relationship. 10 . The method according to claim 1 , wherein the plurality of objects also have the first aggregation relationship in the target object set.

11. The method according to claim 1, further comprising: In response to an evaluation result of the predetermined aggregation strategy meeting a predetermined strategy requirement, the aggregation relationship between the objects and clusters included in the first object set is updated based on the predetermined aggregation strategy.

12. An apparatus for object processing, comprising: A selection module configured to select a plurality of objects associated with a predetermined task from the first object set based on a task requirement of the predetermined task to form a second object set; A first determining module is configured to determine a first aggregation relationship of the plurality of objects in the first object set, wherein the first aggregation relationship indicates that the plurality of objects are aggregated into a plurality of clusters based on similarity, and each cluster includes at least one object among the plurality of objects; a second determining module, configured to determine a second aggregation relationship of the plurality of objects in the second object set based on a predetermined aggregation strategy associated with the predetermined task and the first aggregation relationship; as well as A generation module is configured to generate a target object set matching the predetermined task based on the first object set and the second aggregation relationship, wherein the target object set includes the multiple objects, and the multiple objects have at least the second aggregation relationship in the target object set.

13. An electronic device, comprising: at least one processor; as well as At least one memory, the at least one memory is coupled to the at least one processor and stores instructions for execution by the at least one processor, the instructions, when executed by the at least one processor, causing the electronic device to perform the method according to any one of claims 1 to 11.

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

15. A computer executable instruction product, comprising computer executable instructions, wherein the computer executable instructions, when executed by a processor, implement the method according to any one of claims 1 to 11.

Images