Object modification method and device, electronic equipment and storage medium

Abstract

The invention provides an object modification method and device, electronic equipment and a storage medium, and relates to the technical field of large models, in particular to the technical field of artificial intelligence such as natural language processing. Determining target abnormal request ends with abnormal return quantity from the candidate object request end set according to the return object quantity; obtaining a target filtering strategy corresponding to the target exception request end, and determining a target modification strategy corresponding to the filtering strategy; and based on the target modification strategy, obtaining a target return object of the target exception request end, which is not filtered out by the target filtering strategy.

Description

Technical Field

[0001] This disclosure relates to the field of large model technology, and in particular to artificial intelligence fields such as natural language processing. Background Technology

[0002] With the development of technology, more and more user terminals can realize the function of displaying advertisements to deliver relevant information to users. In the process of advertising placement and display, relevant filtering operations may be required to improve the matching degree between the advertisements displayed on the user terminal and user needs. In this scenario, abnormal advertisement filtering may occur, affecting the user's advertisement reading experience. Summary of the Invention

[0003] This disclosure presents a method, apparatus, electronic device, and storage medium for modifying objects.

[0004] According to a first aspect of this disclosure, an object modification method is proposed, the method comprising: obtaining the number of returned objects received by each of a set of candidate object requesters, and determining a target abnormal requester with an abnormal number of returned objects from the set of candidate object requesters based on the number of returned objects; obtaining a target filtering policy corresponding to the target abnormal requester, and determining a target modification policy corresponding to the filtering policy; and obtaining a target returned object of the target abnormal requester that is not filtered out by the target filtering policy based on the target modification policy.

[0005] According to a second aspect of this disclosure, an object modification apparatus is proposed, the apparatus comprising: an acquisition module, configured to acquire the number of returned objects received by each of a set of candidate object requesters, and determine, based on the number of returned objects, a target abnormal requester with an abnormal number of returned objects from the set of candidate object requesters; a determination module, configured to acquire a target filtering policy corresponding to the target abnormal requester, and determine a target modification policy corresponding to the filtering policy; and a modification module, configured to, based on the target modification policy, obtain target returned objects of the target abnormal requester that are not filtered out by the target filtering policy.

[0006] According to a third aspect of this disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the object modification method proposed in the first aspect above.

[0007] According to a fourth aspect of this disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to perform the object modification method proposed in the first aspect above.

[0008] According to a fifth aspect of this disclosure, a computer program product is proposed, comprising a computer program that, when executed by a processor, implements the object modification method proposed in the first aspect above.

[0009] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0010] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0011] Figure 1 This is a flowchart illustrating an object modification method according to an embodiment of the present disclosure;

[0012] Figure 2 This is a flowchart illustrating an object modification method according to another embodiment of the present disclosure;

[0013] Figure 3 This is a schematic diagram of a target directed acyclic graph according to an embodiment of the present disclosure;

[0014] Figure 4 This is a flowchart illustrating an object modification method according to another embodiment of the present disclosure;

[0015] Figure 5 This is a flowchart illustrating an object modification method according to another embodiment of the present disclosure;

[0016] Figure 6 This is a flowchart illustrating an object modification method according to another embodiment of the present disclosure;

[0017] Figure 7 This is a flowchart illustrating an object modification method according to another embodiment of the present disclosure;

[0018] Figure 8 This is a schematic diagram of the structure of an object modification device according to an embodiment of the present disclosure;

[0019] Figure 9 This is a schematic block diagram of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0020] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0021] Large Language Models (LLMs) are deep learning models trained on massive amounts of text data, enabling them to generate natural language text or understand the meaning of language text. These models can provide in-depth knowledge and language production on a wide range of topics through training on large datasets. Their core idea is to learn patterns and structures of natural language through large-scale unsupervised training, thus mimicking the human language cognition and generation process to some extent.

[0022] Natural Language Processing (NLP) is an important field within computer science and artificial intelligence. It studies the theories and methods that enable effective communication between humans and computers using natural language. NLP is a science that integrates linguistics, computer science, and mathematics. Its main applications include machine translation, public opinion monitoring, automatic summarization, opinion extraction, text classification, question answering, text semantic comparison, speech recognition, and Chinese OCR.

[0023] Artificial Intelligence (AI) is a key driving force behind the new round of technological revolution and industrial transformation. It is a new technological science that studies and develops theories, methods, technologies, and application systems to simulate, extend, and expand human intelligence. AI is an important component of the discipline of intelligence; it attempts to understand the essence of intelligence and produce a new kind of intelligent machine capable of reacting in a manner similar to human intelligence. AI is a very broad science, encompassing robotics, speech recognition, image recognition, natural language processing, expert systems, machine learning, computer vision, and more.

[0024] It should be noted that the acquisition, storage, use, and processing of data in this disclosed technical solution comply with the relevant provisions of national laws and regulations and do not violate public order and good morals.

[0025] Figure 1 This is a flowchart illustrating an object modification method according to an embodiment of the present disclosure, as follows: Figure 1 As shown, the method includes:

[0026] S101, obtain the number of returned objects received by each of the candidate object requesting ends, and determine the target abnormal requesting end with an abnormal number of returned objects from the candidate object requesting ends based on the number of returned objects.

[0027] In the daily delivery and display of advertisements, it is necessary to select the advertisements that match the user's needs from the multiple advertisements delivered, based on the user's relevant needs for advertisement display, and filter the other advertisements, and then display the filtered and retained advertisements on the user's end.

[0028] In this scenario, based on the ad delivery request initiated by the user, the ad filtering algorithm deployed in the ad filtering system filters the ad content delivered to the user, and returns the filtered and retained ads to the user who initiated the request for ad display.

[0029] In this embodiment of the disclosure, the requesting end of the ad delivery can be determined as the candidate object requesting end, and a corresponding dataset can be constructed based on each candidate object requesting end as the candidate object requesting end set. In this scenario, the ad delivery end can transmit the corresponding ad to the ad retrieval end based on the ad delivery request received from any candidate object requesting end. The ad retrieval end filters and selects the ads based on the relevant demand information carried in the request initiated by the candidate object requesting end, and then transmits the filtered and retained ads to the candidate object requesting end for display.

[0030] Among them, the advertisements that are filtered and retained by the advertisement retrieval end and transmitted to the candidate object request end can be identified as the return object corresponding to the delivery request initiated by the candidate object request end.

[0031] In this embodiment of the disclosure, the number of returned objects received by each candidate object requesting end within a set time range can be counted to obtain the number of returned objects for each candidate object requesting end. For any candidate object requesting end, the number of returned objects can be used to identify whether the advertising filtering strategy used by the candidate object requesting end can normally meet the advertising display requirements of the candidate object requesting end.

[0032] Optionally, the number of any returned object can be compared with a preset abnormal quantity judgment condition. When the comparison result indicates that the two match, it can be determined that the number of returned objects may be abnormal. In addition, it can be determined that the advertising filtering strategy used by the candidate object request end may not be able to achieve its required advertising display needs. In this scenario, the candidate object request end can be identified as the target abnormal request end with abnormal returned quantity.

[0033] S102, obtain the target filtering policy corresponding to the target abnormal request end, and determine the target modification policy corresponding to the filtering policy.

[0034] In this embodiment of the disclosure, the advertising filtering strategy used by the advertising filtering system to filter the advertisements to be displayed on the target abnormal request terminal can be obtained, and the strategy can be determined as the target filtering strategy corresponding to the target abnormal request terminal.

[0035] For any field, the target filtering strategy includes a risk description judgment condition corresponding to the character in that field. This condition can identify whether the character in that field is a risk description character. When the character in that field is identified as a risk description character, the advertisement content can be filtered and deleted based on the filtering and deletion method carried in the target filtering strategy.

[0036] In this scenario, the target filtering strategy can be parsed using a strategy parsing algorithm from relevant technologies. The result of the algorithm processing can then be used to obtain a corresponding modification strategy, which can be used as the target modification strategy. Advertisements modified based on the target modification strategy can be removed by filtering operations that use that target filtering strategy.

[0037] In other words, when a target-based filtering strategy is used to identify and filter the modified advertisement based on that target-based modification strategy, the ad filtering tool using that target-based filtering strategy will not identify the advertisement as one that needs to be deleted. Instead, it will retain the advertisement and send it as a return object to the corresponding candidate object requesting end for display.

[0038] S103, based on the target modification strategy, obtain the target return object of the target exception request that is not filtered out by the target filtering strategy.

[0039] In this embodiment of the disclosure, based on the target modification strategy, advertisements delivered to the target abnormal request terminal within a set time range can be identified and modified respectively. That is, based on the target modification strategy, it is possible to identify whether there are risky characters in each advertisement, and modify the identified risky characters accordingly to obtain the modified advertisement. Here, risky characters can be understood as characters carried in the advertisement that may be judged as non-compliant expressions by the target filtering strategy.

[0040] In this scenario, the modified advertisement will not carry any relevant risky characters. In other words, the modified advertisement will not be filtered and deleted by the target filtering strategy. Therefore, the modified advertisement can be identified as the target return object on the target abnormal request end based on the target modification strategy.

[0041] The object modification method proposed in this disclosure obtains the number of returned objects for each candidate object request set to identify the target abnormal request with an abnormal number of returned objects. Based on the target modification strategy corresponding to the target abnormal request, the relevant ad delivery objects are modified to obtain target returned objects that are not filtered out by the target filtering strategy on the target abnormal request. In this disclosure, identifying the target abnormal request based on the number of returned objects improves the efficiency and accuracy of locating the target abnormal request with an abnormal number of ad returned objects, thereby improving the efficiency and accuracy of obtaining the corresponding target filtering strategy. Determining the target modification strategy based on the target filtering strategy improves the matching degree between the target modification strategy and the target filtering strategy. Obtaining the target returned objects based on the target modification strategy reduces the possibility that the modified target returned objects will be filtered and deleted by the target filtering strategy, thereby reducing the possibility of abnormal ad returned numbers on the target abnormal request, improving the ad returned number on the target abnormal request, improving the stability of ad display on the target abnormal request, optimizing the ad filtering method, and thus optimizing the user experience.

[0042] In the above embodiments, the acquisition of the target returned object can also be combined with... Figure 2 To understand further, Figure 2 This is a flowchart illustrating an object modification method according to another embodiment of the present disclosure, as shown below. Figure 2 As shown, the method includes:

[0043] S20 1, obtain the dependency relationships between the requesters of each candidate object, and construct the corresponding target directed acyclic graph based on the requesters of each candidate object, the dependency relationships between the requesters of each candidate object, and the number of returned objects of each requester of each candidate object.

[0044] Optionally, each candidate object requester is treated as a node, and edges are constructed between the nodes based on the dependencies between the candidate object requesters to obtain a candidate directed acyclic graph.

[0045] In this embodiment of the disclosure, there may be a dependency relationship between the requesting ends of each candidate object, and the upstream and downstream ends of each requesting end of the candidate object can be determined based on the dependency relationship.

[0046] Specifically, for any candidate object requesting end, the dependent object requesting end of the candidate object requesting end can be obtained as the downstream end of the candidate object requesting end, and the corresponding end that needs to depend on the candidate object requesting end can be obtained as the upstream end of the candidate object requesting end.

[0047] In this scenario, any candidate object requesting end is taken as the parent node, and its corresponding downstream end is taken as the child node. A directed edge is constructed between the parent node and the child node based on the direction from the candidate object requesting end to the downstream end it depends on. Then, a directed acyclic graph is constructed based on the directed acyclic graph construction algorithm in related technologies, and the constructed directed acyclic graph is determined as the candidate directed acyclic graph.

[0048] Optionally, the number of returned objects from each candidate object request is used to mark each node in the candidate directed acyclic graph to obtain the target directed acyclic graph.

[0049] In this embodiment of the disclosure, the number of returned objects of any candidate object requesting end can be marked to the corresponding node of the candidate requesting object in the candidate directed acyclic graph based on the data marking method in the related technology, thereby realizing the marking of the number of returned objects in the candidate directed acyclic graph, and determining the marked directed acyclic graph as the target directed acyclic graph.

[0050] As an example, such as Figure 3 As shown, if a dependency relationship exists between candidate object requester A and candidate object requester B, where candidate object requester B is a dependent of candidate object requester A, then candidate object requester A can be designated as the parent node, candidate object requester B as the child node, and the directed edge from candidate object requester A to candidate object requester B can be designated as the edge between them, thus constructing a directed acyclic graph. Figure 3 The diagram shows a directed acyclic graph.

[0051] It should be noted that the construction of the directed acyclic graphs corresponding to the candidate object request terminals C, D, E, and F can be understood by referring to the construction method of the directed acyclic graphs corresponding to the candidate object request terminals A and B mentioned above, and will not be repeated here.

[0052] As another example, such as Figure 4 As shown, it can be based on Figure 4 The link information acquisition module shown acquires the requesting end of each candidate object and parses the log data of each candidate object requesting end to obtain the number of returned objects on each candidate object requesting end.

[0053] like Figure 4 As shown, through Figure 4 The directed acyclic graph construction module shown obtains the dependency relationships between the request ends of each candidate object, takes each request end of the candidate object as a node, and obtains the directed edges between each node based on the dependency relationships, thereby constructing the corresponding target directed acyclic graph.

[0054] S202, traverse the target directed acyclic graph to determine the target return exception node from each node, and determine the target exception request end based on the target return exception node.

[0055] In this embodiment of the disclosure, the target directed acyclic graph can be traversed based on the directed acyclic graph traversal method in the related technology to obtain the number of returned objects marked on each node, and then the nodes with abnormal number of returned objects can be identified from each node and determined as the target returned abnormal nodes.

[0056] Among them, the traversal method of the target directed acyclic graph can be constructed based on the depth-first search (DFS) in related technologies, or it can be based on other methods to implement the traversal of the target directed acyclic graph. No specific limitation is made here.

[0057] As an example, such as Figure 4 As shown, it can be done through Figure 4 The filtering plugin positioning module shown traverses the target directed acyclic graph, obtains the number of returned objects on the current node, identifies whether the node is a target return abnormal node, and thus realizes the positioning of the corresponding target return abnormal request end.

[0058] Optionally, for any node in the target directed acyclic graph, in response to the number of returned objects on the node matching a preset return number anomaly judgment condition, the node is determined to be a target return anomaly node.

[0059] In this embodiment of the disclosure, for any node in the target directed acyclic graph, the number of returned objects on that node can be compared with the corresponding return number anomaly judgment condition to obtain the corresponding comparison result.

[0060] Specifically, for any node in the target directed acyclic graph, in response to the number of returned objects corresponding to the node being greater than or equal to a preset first threshold for the number of returned objects, it is determined that the number of returned objects corresponding to the node does not match the abnormal return count judgment condition.

[0061] In other words, when the number of returned objects is greater than or equal to the first threshold for the number of returned objects, it can be determined that the number of returned objects can meet the advertising display requirements of the corresponding candidate object requesting end. That is, the advertising filtering strategy used when the number of returned objects is obtained meets the advertising filtering requirements of the corresponding candidate object requesting end. In this scenario, it can be determined that the number of returned objects does not match the return number abnormality judgment.

[0062] This can be understood as the number of returned objects from the candidate object requesting end corresponding to this node being normal.

[0063] Optionally, in response to the number of returned objects corresponding to a node being less than or equal to a preset second threshold for the number of returned objects, and the number of returned objects of the next adjacent node of the node being greater than or equal to a first threshold for the number of returned objects, the number of returned objects corresponding to the node is determined to match the return number anomaly judgment condition, and the node is determined to be the target return anomaly node.

[0064] In this embodiment of the disclosure, there is a dependency relationship between any node and its next adjacent node. That is, there may be a certain degree of correlation between the advertising display needs of the two nodes. In this scenario, the number of returned objects received by the two nodes may be relatively close. In other words, the number deviation between the number of returned objects received by the two nodes may be less than a set range.

[0065] In this scenario, if the number of returned objects is less than or equal to the second threshold for the number of returned objects and the number of returned objects of the corresponding next adjacent node is greater than or equal to the first threshold for the number of returned objects, it can be understood that the number of returned objects received by the next adjacent node of the current node is normal. As can be seen from the above, in this scenario, the number of returned objects received by the current node may be close to the number of returned objects received by its corresponding next adjacent node.

[0066] Therefore, when the number of returned objects received by the current node is less than or equal to the second threshold for the number of returned objects, it can be determined that the number of returned objects received by the current node is abnormal. In this case, the number of returned objects on the current node matches the abnormal judgment condition, and the node can be identified as the target returned abnormal node with an abnormal number of returned objects.

[0067] Optionally, the associated object request of the target returned abnormal node in the candidate object request set is determined as the target abnormal request.

[0068] It should be noted that the threshold for the number of first returned objects can be set to 0 or other values; no specific restrictions are imposed here.

[0069] Furthermore, the second return object quantity threshold can be set to 0. When the second return object quantity threshold is 0, if the number of return objects of the current node is equal to 0 and the number of return objects on the next adjacent node is greater than 0, the node can be identified as the target return abnormal node.

[0070] In this embodiment of the disclosure, the associated end corresponding to the target returned abnormal node in each candidate object request end can be obtained as the corresponding associated object request end, wherein the associated object request end is the target abnormal request end corresponding to the target returned abnormal node.

[0071] As an example, such as Figure 5 As shown, it can be based on Figure 5 The method shown traverses the target directed acyclic graph, where, for Figure 5 The current node is shown. If the current node has been visited, it is skipped. If the current node has not been visited, it is marked as visited and the number of returned objects on the node is obtained.

[0072] like Figure 5 As shown, when the number of objects returned by a node is greater than 0, its filtering is considered normal. When the number of objects returned is equal to 0, the list of child nodes under that node is obtained and each child node in the list is traversed. If the number of objects returned by the current child node is greater than 0, the filtering of that child node is considered normal. If the number of objects returned by the current child node is equal to 0 and the number of objects returned by its next adjacent child node is greater than 0, the filtering strategy of that child node is considered abnormal, and that child node is considered a target return abnormal node. The remaining part of the directed acyclic graph is then traversed until all target return abnormal nodes are obtained.

[0073] S203, obtain the candidate filtering strategies on each candidate filtering plugin, and construct the corresponding target knowledge graph based on each candidate filtering strategy.

[0074] Optionally, the candidate filtering strategies on each candidate filtering plugin are obtained, wherein the candidate filtering strategy includes at least the strategy filtering configuration information, strategy code snippets, and strategy code logic corresponding to the candidate filtering plugin.

[0075] As an example, such as Figure 6 As shown, it can be seen from Figure 6 In the module codebase shown, filter plugin files and filter plugin functions are selected from any candidate filter plugin. Figure 6 The rule mining algorithm shown performs rule mining on the filter plugin file and filter plugin function to obtain... Figure 6 The code snippet shown is used by the filtering plugin to filter ads, and the filtering logic carried in the code snippet is the strategy code snippet corresponding to the candidate filtering plugin. The filtering logic in the code can be identified as the corresponding strategy code logic.

[0076] And, from Figure 6 Retrieve the filter plugin configuration file of any candidate filter plugin from the configuration information repository shown, and based on... Figure 6 The rule mining algorithm shown extracts the corresponding configuration information from the filter plugin configuration file, which serves as the policy filtering configuration information for the candidate filter plugin.

[0077] In this example, the strategy acquisition algorithm in the relevant technology can be used to process the strategy filtering configuration information, strategy code snippets and strategy code logic, and then the filtering strategy used by the candidate filtering plugin can be obtained based on the result of the algorithm processing, which can be used as the candidate filtering strategy.

[0078] Optionally, the model capabilities of the large model can be invoked to obtain the candidate modification strategies for each candidate filtering strategy.

[0079] In this embodiment of the disclosure, for any candidate filtering strategy, the model capabilities of a large model can be invoked to perform filtering strategy analysis on the candidate filtering strategy, and a corresponding modification strategy can be generated based on the analysis results. This modification strategy can then be determined as a candidate modification strategy for the candidate filtering strategy.

[0080] As an example, such as Figure 6 As shown, candidate filtering strategies constructed based on the obtained strategy configuration information, strategy code snippets, and strategy code logic can be input into the system. Figure 6 In the large model shown, the model capabilities of the large model are used to generate the candidate modification strategy corresponding to the candidate filtering strategy.

[0081] Optionally, a target knowledge graph is constructed based on each candidate filtering strategy and the candidate modification strategy of each candidate filtering strategy.

[0082] Among them, knowledge graph construction algorithms based on relevant technologies can be used to process each candidate filtering strategy and each candidate modification strategy, and then the corresponding knowledge graph can be constructed based on the results of the algorithm processing, which can be used as the target knowledge graph.

[0083] As an example, such as Figure 6 As shown, the acquired candidate filtering strategies and their corresponding candidate modification strategies can be associated, and then a knowledge graph can be constructed based on the associated strategy group to obtain... Figure 6 The knowledge graph shown is used as the target knowledge graph.

[0084] S204, obtain the target filtering logs from the target abnormal request end, and obtain the target modification strategy from the target knowledge graph based on the target filtering logs.

[0085] Optionally, the filter operation identifier corresponding to the filter operation executed on the target abnormal request end can be obtained from the target filter log, wherein the filter operation identifier is obtained by marking the filter operation with dots;

[0086] In this embodiment of the disclosure, log data of the advertising filtering behavior corresponding to the target abnormal request terminal can be obtained. The log can be identified as the target filtering log corresponding to the target abnormal request terminal. The target filtering log records various filtering operation data of the corresponding filtering plugin when filtering the advertisement on the target abnormal request terminal.

[0087] In this scenario, data can be read from the target filtering logs to determine the relevant identification information of the filtering plugin when performing filtering operations. This identification information can then be identified as the filtering operation identifier corresponding to the filtering operation.

[0088] The filtering operation identifier can be obtained based on dot operations in related technologies or based on other identifier construction methods; no specific limitation is made here.

[0089] Optionally, the target filtering plugin for performing the filtering operation is determined from among the candidate filtering plugins based on the filtering operation identifier.

[0090] In this embodiment of the disclosure, the plugin identifier of the corresponding filtering operation can be determined based on the filtering operation identifier. In this scenario, the candidate filtering plugin corresponding to the plugin identifier can be obtained from each candidate filtering plugin, and the candidate filtering plugin is the corresponding target filtering plugin.

[0091] Optionally, knowledge retrieval is performed on the target knowledge graph based on the target filtering plugin to obtain the target knowledge corresponding to the target filtering plugin in the target knowledge graph, and the modification strategy carried in the target knowledge is determined as the target modification strategy.

[0092] In this embodiment of the disclosure, a knowledge retrieval algorithm based on a target filtering plugin can be used to perform corresponding knowledge retrieval based on the target filtering plugin, and the retrieved knowledge related to the target filtering plugin can be identified as target knowledge.

[0093] The target knowledge includes at least the relevant knowledge of the filtering strategy of the target filtering plugin, as well as the relevant knowledge of the corresponding modification strategy. In this scenario, the modification strategy carried in the target knowledge can be obtained, and the modification strategy can be identified as the target modification strategy.

[0094] S205, based on the target modification strategy, obtain the target return object of the target exception request that is not filtered out by the target filtering strategy.

[0095] Optionally, obtain the candidate delivery object corresponding to the target abnormal request client.

[0096] In this embodiment of the disclosure, after receiving an advertising delivery request initiated by the target abnormal request terminal, the advertising delivery terminal can transmit the delivered advertisement to the advertising filtering terminal based on the received request. In this scenario, the advertisements transmitted to the advertising filtering terminal can be identified as candidate delivery objects on the target abnormal request terminal.

[0097] Optionally, based on the target filtering strategy, the root cause field corresponding to the candidate delivery object is determined, wherein the root cause field is the root cause that causes the candidate delivery object to be filtered out by the target filtering strategy.

[0098] In this embodiment of the disclosure, the candidate delivery object may include multiple levels of information. Based on the target filtering strategy used by the target abnormal request end, the fields included in each level of the candidate delivery object can be identified and judged, and risk fields that may cause the candidate delivery object to be filtered and deleted by the target filtering strategy can be screened out. The screened risk fields are then determined as the corresponding root cause fields.

[0099] Optionally, based on the target modification strategy, the update character corresponding to the root cause field is determined, and the candidate delivery object is modified based on the update field to obtain the target return object.

[0100] In this embodiment of the disclosure, the character content on the root cause field can be modified based on the target modification strategy, and the new character obtained from the modification can be determined as the updated character.

[0101] In this scenario, candidate targets can be modified based on the updated characters. Specifically, the characters in the root cause field can be deleted to obtain a blank field, and the updated characters can be filled into the blank field to modify the candidate targets.

[0102] The modified delivery object can be identified as the target return object of the target abnormal request end. The target return object can be exempted from being filtered and deleted by the filtering plugin using the target filtering strategy.

[0103] It should be noted that the root cause fields determined based on the target modification strategy may include relevant fields in the advertising request corresponding to the candidate target, relevant fields in the original material content of the candidate target, relevant fields in the configuration file and the thesaurus file, and fields under the relevant level of the candidate target. No specific limitations are made here.

[0104] Specifically, if the root cause field is a field in the corresponding request, then the next ad request will be modified; if the root cause field is a related field in the original material content, then the original material content will be modified; if the root cause field is at least one of a related field in the configuration file and a related field in the thesaurus file, then the corresponding field in the corresponding file will be modified.

[0105] In this embodiment of the disclosure, before modifying the candidate delivery targets based on the target modification strategy, it is also necessary to analyze and determine the target modification strategy, which can be understood in conjunction with the following:

[0106] Optionally, obtain the description information of the modification strategy corresponding to the target modification strategy and the execution script of the modification strategy.

[0107] In this embodiment of the disclosure, the target modification strategy has a corresponding text version, wherein the text version includes strategy description information such as the relevant modification methods, modification rules and modification logic of the target modification strategy, and the text version can be identified as the modification strategy description information corresponding to the target modification strategy.

[0108] Furthermore, the target modification strategy has a corresponding execution script, which can be identified as the modification strategy execution script corresponding to the target modification strategy.

[0109] Optionally, in response to determining that a target modification strategy is available based on the modification strategy description information, a modification strategy execution script is run to modify the candidate delivery object and obtain the target return object;

[0110] In this embodiment of the disclosure, various information in the target modification strategy can be directly read by modifying the strategy description information. Then, based on the read information, it can be identified whether the current target modification strategy can achieve the purpose of preventing the modified target return object from being filtered and deleted by the filtering plugin using the target filtering strategy. When it is identified that the target return object modified based on the target modification strategy can be prevented from being filtered and deleted by the filtering plugin using the target filtering strategy, it can be determined that the current target modification strategy is usable.

[0111] In this scenario, the modified strategy execution script can be run to modify the candidate delivery objects based on the current target modification strategy, thereby obtaining the modified target return object.

[0112] Optionally, in response to determining that the target modification strategy is unavailable based on the modification strategy description information, the target modification strategy is adjusted based on the target filtering strategy to obtain a new target modification strategy until the new target modification strategy becomes available, and the candidate delivery objects are modified based on the new target modification strategy to obtain the target return object.

[0113] In this embodiment of the disclosure, when it is determined based on the information read from the modification strategy description information that the target return object obtained after modification based on the current target modification strategy may be filtered and deleted by the filtering plugin using the target filtering strategy, it can be determined that the current target modification strategy is unusable.

[0114] In this scenario, the target modification strategy can be adjusted and optimized based on the modification strategy optimization method in related technologies. The target modification strategy after each round of adjustment can be identified and judged until the target return object obtained by modifying based on the new target modification strategy can be determined to be usable by the filtering plugin that uses the target filtering strategy.

[0115] In this scenario, the new target modification strategy execution script can be run, and the candidate delivery objects can be modified based on the new target modification strategy to obtain the corresponding target return objects.

[0116] The object modification method proposed in this disclosure determines the abnormal nodes of the target based on the traversal of the target directed acyclic graph, and then identifies the corresponding abnormal request end of the target, improving the efficiency and accuracy of the location of the abnormal request end. By obtaining the filter plugin identifier through the point-marking operation, the corresponding target filter plugin is identified, thereby obtaining the target filter strategy, which improves the accuracy of target filter strategy acquisition. Based on the model capabilities of the large model, modification strategies corresponding to each filter strategy are generated, which improves the accuracy and matching degree of the candidate modification strategies corresponding to each candidate filter strategy. The target modification strategy corresponding to the target filter strategy is obtained from the target knowledge graph, which improves the efficiency of target modification strategy acquisition, thereby improving the modification efficiency of the target returned object. The usability of the target modification strategy is judged and optimized, which improves the quality of the target modification strategy and reduces the possibility that the target returned object obtained based on the target modification strategy is filtered and deleted by the target filter strategy.

[0117] To better understand the above embodiments, it can be combined with Figure 7 , Figure 7 This is a flowchart illustrating an object modification method according to another embodiment of the present disclosure.

[0118] like Figure 7 As shown, through Figure 7 The environmental labels shown are for Figure 7 The database corresponding to the environmental information data is retrieved to determine the corresponding ad filtering system. Then, each candidate requester that initiates the relevant ad delivery request and uses the ad filtering system for ad filtering is obtained. Based on the dependency relationship between each candidate requester and the number of returned objects on each candidate requester, the corresponding target directed acyclic graph is constructed.

[0119] like Figure 7 As shown, based on the traversal of the target directed acyclic graph, the corresponding target anomaly nodes are determined, thereby locating the target return anomaly requesting an abnormal number of returned objects. Figure 7The database corresponding to the log data shown obtains the filter logs corresponding to the target's returned abnormal request, and then locates the target filtering plugin that performs ad filtering on the target's returned abnormal request based on the filtering operation identifier in the filter logs.

[0120] Optionally, obtain the target filtering policy on the target filtering plugin, and based on Figure 7 The target knowledge graph shown determines the target modification strategy corresponding to the target filtering strategy, and then modifies the candidate delivery objects delivered to the target return exception request end based on the target modification strategy to obtain the target return object.

[0121] The object modification method proposed in this disclosure obtains the number of returned objects for each candidate object request set to identify the target abnormal request with an abnormal number of returned objects. Based on the target modification strategy corresponding to the target abnormal request, the relevant ad delivery objects are modified to obtain target returned objects that are not filtered out by the target filtering strategy on the target abnormal request. In this disclosure, identifying the target abnormal request based on the number of returned objects improves the efficiency and accuracy of locating the target abnormal request with an abnormal number of ad returned objects, thereby improving the efficiency and accuracy of obtaining the corresponding target filtering strategy. Determining the target modification strategy based on the target filtering strategy improves the matching degree between the target modification strategy and the target filtering strategy. Obtaining the target returned objects based on the target modification strategy reduces the possibility that the modified target returned objects will be filtered and deleted by the target filtering strategy, thereby reducing the possibility of abnormal ad returned numbers on the target abnormal request, improving the ad returned number on the target abnormal request, improving the stability of ad display on the target abnormal request, optimizing the ad filtering method, and thus optimizing the user experience.

[0122] An embodiment of this disclosure also proposes an object modification apparatus. Since the object modification apparatus for scheduling agents proposed in this disclosure corresponds to the object modification methods for scheduling agents proposed in the above embodiments, the implementation methods for object modification methods for scheduling agents described above are also applicable to the object modification apparatus for scheduling agents proposed in this disclosure. It will not be described in detail in the following embodiments.

[0123] Figure 8 This is a schematic diagram of the structure of an object modification device according to an embodiment of the present disclosure, as shown below. Figure 8 As shown, the object modification device 800 includes an acquisition module 81, a determination module 82, and a modification module 83, wherein:

[0124] The acquisition module 81 is used to acquire the number of returned objects received by each of the candidate object requesting end sets, and to determine the target abnormal requesting end with an abnormal number of returned objects from the candidate object requesting end set based on the number of returned objects.

[0125] The determination module 82 is used to obtain the target filtering policy corresponding to the target abnormal request end, and determine the target modification policy corresponding to the filtering policy;

[0126] Modify module 83 to obtain the target return object of the target abnormal request that is not filtered out by the target filtering strategy, based on the target modification strategy.

[0127] In this embodiment of the disclosure, the acquisition module 81 is further configured to: acquire the dependency relationship between each candidate object requester, and construct a corresponding target directed acyclic graph based on each candidate object requester, the dependency relationship between each candidate object requester and the number of returned objects of each candidate object requester; traverse the target directed acyclic graph to determine the target return anomaly node from each node, and determine the target anomaly requester based on the target return anomaly node.

[0128] In this embodiment of the disclosure, the acquisition module 81 is further configured to: take each candidate object requester as a node, construct edges between each node based on the dependency relationship between each candidate object requester to obtain a candidate directed acyclic graph; mark each node in the candidate directed acyclic graph with the number of returned objects of each candidate object requester to obtain a target directed acyclic graph.

[0129] In this embodiment of the disclosure, the obtained module 81 is further configured to: for any node in the target directed acyclic graph, in response to the number of returned objects on the node matching the preset return number anomaly judgment condition, determine the node as the target return anomaly node.

[0130] The associated object request of the target returned abnormal node in the candidate object request set is identified as the target abnormal request.

[0131] In this embodiment of the disclosure, the acquisition module 81 is further configured to: for any node in the target directed acyclic graph, in response to the number of returned objects corresponding to the node being greater than or equal to a preset first threshold for the number of returned objects, determine that the number of returned objects corresponding to the node does not match the return quantity abnormality judgment condition; in response to the number of returned objects corresponding to the node being less than or equal to a preset second threshold for the number of returned objects, and the number of returned objects of the next adjacent node of the node being greater than or equal to the first threshold for the number of returned objects, determine that the number of returned objects corresponding to the node matches the return quantity abnormality judgment condition, and determine the node as the target return abnormal node.

[0132] In this embodiment of the disclosure, the determining module 82 is further configured to: obtain candidate filtering strategies on each candidate filtering plugin, and construct a corresponding target knowledge graph based on each candidate filtering strategy; obtain target filtering logs from the target abnormal request end, and obtain target modification strategies from the target knowledge graph based on the target filtering logs.

[0133] In this embodiment of the disclosure, the determining module 82 is further configured to: obtain candidate filtering strategies on each candidate filtering plugin, wherein the candidate filtering strategy includes at least the strategy filtering configuration information, strategy code snippets, and strategy code logic corresponding to the corresponding candidate filtering plugin; call the model capabilities of the large model to obtain the candidate modification strategies of each candidate filtering strategy; and construct the target knowledge graph based on each candidate filtering strategy and the candidate modification strategies of each candidate filtering strategy.

[0134] In this embodiment of the disclosure, the determining module 82 is further configured to: obtain the filter operation identifier corresponding to the filter operation executed at the target abnormal request end from the target filter log, wherein the filter operation identifier is obtained by marking the filter operation; determine the target filter plugin that performs the filter operation from each candidate filter plugin based on the filter operation identifier; perform knowledge retrieval on the target knowledge graph based on the target filter plugin to obtain the target knowledge corresponding to the target filter plugin in the target knowledge graph, and determine the modification strategy carried in the target knowledge as the target modification strategy.

[0135] In this embodiment of the disclosure, the modification module 83 is further configured to: obtain the candidate delivery object corresponding to the target abnormal request end; determine the root cause field corresponding to the candidate delivery object based on the target filtering strategy, wherein the root cause field is the root cause that causes the candidate delivery object to be filtered out by the target filtering strategy; determine the update character corresponding to the root cause field based on the target modification strategy, and modify the candidate delivery object based on the update character to obtain the target return object.

[0136] In this embodiment of the disclosure, the modification module 83 is further configured to: obtain modification strategy description information and modification strategy execution script corresponding to the target modification strategy; in response to determining that the target modification strategy is available based on the modification strategy description information, run the modification strategy execution script to modify the candidate delivery objects and obtain the target return object; in response to determining that the target modification strategy is unavailable based on the modification strategy description information, adjust the target modification strategy based on the target filtering strategy to obtain a new target modification strategy, until the new target modification strategy is available, modify the candidate delivery objects based on the new target modification strategy, and obtain the target return object.

[0137] The object modification apparatus disclosed herein obtains the number of returned objects for each candidate object request set to identify the target abnormal request set with an abnormal number of returned objects. Based on the target modification strategy corresponding to the target abnormal request set, the relevant ad delivery objects are modified to obtain target returned objects on the target abnormal request set that are not filtered out by the target filtering strategy. In this disclosure, identifying the target abnormal request set based on the number of returned objects improves the location efficiency and accuracy of the target abnormal request set with an abnormal number of returned ads, thereby improving the acquisition efficiency and accuracy of the corresponding target filtering strategy. Determining the target modification strategy based on the target filtering strategy improves the matching degree between the target modification strategy and the target filtering strategy. Obtaining the target returned objects based on the target modification strategy reduces the possibility that the modified target returned objects will be filtered and deleted by the target filtering strategy, thereby reducing the possibility of abnormal ad returned numbers on the target abnormal request set, improving the ad returned number on the target abnormal request set, improving the stability of ad display on the target abnormal request set, optimizing the ad filtering method, and thus optimizing the user experience.

[0138] According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0139] Figure 9 A schematic block diagram of an example electronic device 900 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0140] like Figure 9 As shown, device 900 includes a computing unit 901, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 902 or a computer program loaded from storage unit 908 into random access memory (RAM) 903. RAM 903 may also store various programs and data required for the operation of device 900. The computing unit 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0141] Multiple components in device 900 are connected to I / O interface 905, including: input unit 906, such as keyboard, mouse, etc.; output unit 907, such as various types of monitors, speakers, etc.; storage unit 908, such as disk, optical disk, etc.; and communication unit 909, such as network card, modem, wireless transceiver, etc. Communication unit 909 allows device 900 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0142] The computing unit 901 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 performs the various methods and processes described above, such as the object modification method. For example, in some embodiments, the object modification method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program may be loaded and / or installed on device 900 via ROM 902 and / or communication unit 909. When the computer program is loaded into RAM 903 and executed by the computing unit 901, one or more steps of the object modification method described above may be performed. Alternatively, in other embodiments, the computing unit 901 may be configured to perform the object modification method by any other suitable means (e.g., by means of firmware).

[0143] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0144] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0145] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0146] To initiate interaction with a user account, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user account; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user account can submit input to the computer. Other types of devices can also be used to initiate interaction with the user account; for example, feedback submitted to the user account can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user account can be received in any form (including voice input, speech input, or tactile input).

[0147] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user account computer with a graphical user interface or web browser through which a user account can interact with the implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0148] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0149] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0150] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. An object modification method, wherein, The method includes: Obtain the number of returned objects received by each of the candidate object requesting ends, and determine the target abnormal requesting end with an abnormal number of returned objects from the candidate object requesting ends based on the number of returned objects. Obtain the target filtering policy corresponding to the target abnormal request client, and determine the target modification policy corresponding to the filtering policy; Based on the target modification strategy, the target return object of the target abnormal request end that is not filtered out by the target filtering strategy is obtained.

2. The method according to claim 1, wherein, The step of obtaining the number of returned objects received by each of the candidate object requesting ends and determining the target abnormal requesting end with an abnormal number of returned objects from the candidate object requesting end set based on the number of returned objects includes: Obtain the dependencies between the requesters of each candidate object, and construct the corresponding target directed acyclic graph based on the requesters of each candidate object, the dependencies between the requesters of each candidate object, and the number of returned objects of each requester of each candidate object. Traverse the target directed acyclic graph to determine the target return anomaly node from each node, and determine the target anomaly requesting end based on the target return anomaly node.

3. The method according to claim 2, wherein, The process of obtaining the dependencies between each candidate object requester, and constructing the corresponding target directed acyclic graph based on each candidate object requester, the dependencies between each candidate object requester, and the number of returned objects from each candidate object requester, includes: Each candidate object requester is treated as a node, and edges are constructed between the nodes based on the dependencies between the candidate object requesters to obtain a candidate directed acyclic graph. The number of returned objects from each candidate object request is used to mark each node in the candidate directed acyclic graph to obtain the target directed acyclic graph.

4. The method according to claim 2, wherein, The step of traversing the target directed acyclic graph to determine the target return anomaly node from each node, and determining the target anomaly requesting end based on the target return anomaly node, includes: For any node in the target directed acyclic graph, in response to the number of returned objects on the node matching a preset return number anomaly judgment condition, the node is determined to be the target return anomaly node; The associated object request terminal of the target returned abnormal node in the candidate object request terminal set is determined as the target abnormal request terminal.

5. The method according to claim 4, wherein, The method further includes: For any node in the target directed acyclic graph, in response to the number of returned objects corresponding to the node being greater than or equal to a preset first threshold for the number of returned objects, it is determined that the number of returned objects corresponding to the node does not match the abnormal return number determination condition. In response to the fact that the number of returned objects corresponding to the node is less than or equal to a preset second threshold for the number of returned objects, and the number of returned objects of the next adjacent node of the node is greater than or equal to the first threshold for the number of returned objects, it is determined that the number of returned objects corresponding to the node matches the return number anomaly determination condition, and the node is determined to be the target return anomaly node.

6. The method according to claim 1, wherein, The step of obtaining the target filtering policy corresponding to the target abnormal request client and determining the target modification policy corresponding to the filtering policy includes: Obtain the candidate filtering strategies on each candidate filtering plugin, and construct the corresponding target knowledge graph based on each candidate filtering strategy; Obtain the target filtering logs from the target abnormal request end, and obtain the target modification strategy from the target knowledge graph based on the target filtering logs.

7. The method according to claim 6, wherein, The step of obtaining candidate filtering strategies from each candidate filtering plugin and constructing the corresponding target knowledge graph based on each candidate filtering strategy includes: Obtain the candidate filtering strategies on each candidate filtering plugin, wherein the candidate filtering strategy includes at least the strategy filtering configuration information, strategy code snippets, and strategy code logic corresponding to the candidate filtering plugin; By utilizing the capabilities of the large model, we can obtain the candidate modification strategies for each candidate filtering strategy. The target knowledge graph is constructed based on each candidate filtering strategy and the candidate modification strategy of each candidate filtering strategy.

8. The method according to claim 7, wherein, The step of obtaining the target filtering log of the target abnormal request terminal and obtaining the target modification strategy from the target knowledge graph based on the target filtering log includes: From the target filtering log, obtain the filtering operation identifier corresponding to the filtering operation executed on the target abnormal request end, wherein the filtering operation identifier is obtained based on the filtering operation being marked with dots; Based on the filtering operation identifier, the target filtering plugin for performing the filtering operation is determined from each candidate filtering plugin; Based on the target filtering plugin, knowledge retrieval is performed on the target knowledge graph to obtain the target knowledge corresponding to the target filtering plugin in the target knowledge graph, and the modification strategy carried in the target knowledge is determined as the target modification strategy.

9. The method according to claim 1, wherein, The step of obtaining the target return object of the target abnormal request that is not filtered out by the target filtering strategy based on the target modification strategy includes: Obtain the candidate delivery object corresponding to the target abnormal request terminal; Based on the target filtering strategy, the root cause field corresponding to the candidate target object is determined, wherein the root cause field is the root cause that causes the candidate target object to be filtered out by the target filtering strategy; Based on the target modification strategy, the update character corresponding to the root cause field is determined, and the candidate delivery object is modified based on the update character to obtain the target return object.

10. The method according to claim 9, wherein, Before obtaining the target return object of the target abnormal request that is not filtered out by the target filtering strategy based on the target modification strategy, the method further includes: Obtain the description information of the modification strategy corresponding to the target modification strategy and the execution script of the modification strategy; In response to determining that the target modification strategy is available based on the modification strategy description information, the modification strategy execution script is run to modify the candidate delivery object and obtain the target return object; In response to determining that the target modification strategy is unavailable based on the modification strategy description information, the target modification strategy is adjusted based on the target filtering strategy to obtain a new target modification strategy until the new target modification strategy becomes available. Based on the new target modification strategy, the candidate delivery object is modified to obtain the target return object.

11. An object modification apparatus, wherein, The device includes: The acquisition module is used to acquire the number of returned objects received by each of the candidate object requesting end sets, and to determine the target abnormal requesting end with an abnormal number of returned objects from the candidate object requesting end set based on the number of returned objects. The determination module is used to obtain the target filtering policy corresponding to the target abnormal request terminal, and determine the target modification policy corresponding to the filtering policy; The modification module is used to obtain the target return object of the target abnormal request end that is not filtered out by the target filtering strategy, based on the target modification strategy.

12. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method of any one of claims 1-10.

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

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

Images