Task information generation method, object processing method, electronic device, storage medium and program product

Abstract

The embodiment of the invention provides a task information generation method, an object processing method, electronic equipment, a storage medium and a program product. The method comprises the following steps: acquiring classification demand information; classification dimension analysis is carried out according to the classification demand information, type labels under different classification dimensions and label description information of the type labels are generated, and the different classification dimensions correspond to different type labels and the label description information of the type labels; and for at least one classification dimension in different classification dimensions, type identification task information corresponding to the type label is generated based on the type label and the label description information under the classification dimension, and different types of identification task information are used for indicating object type identification under different classification dimensions. By means of the technical scheme, the accuracy of the generated type recognition task information can be improved, and then the accuracy of object type recognition based on the type recognition task information is improved.

Description

Technical Field

[0001] This disclosure relates to the field of type recognition technology, and in particular to a task information generation method, an object processing method, an electronic device, a storage medium, and a program product. Background Technology

[0002] In some scenarios, there is a need for object classification and processing, such as identifying the type of an object and processing it accordingly. However, the accuracy of object classification in related technologies is relatively low. Summary of the Invention

[0003] This disclosure provides a task information generation method, an object processing method, an electronic device, a storage medium, and a program product to improve the accuracy of object classification.

[0004] In a first aspect, embodiments of this disclosure provide a method for generating task information, including:

[0005] Obtain classification requirement information;

[0006] Based on the classification requirement information, a classification dimension analysis is performed to generate type labels under different classification dimensions and label description information for the type labels. Different classification dimensions correspond to different type labels and label description information for the type labels.

[0007] For at least one of the different classification dimensions, based on the type label and the label description information under the classification dimension, type recognition task information corresponding to the type label is generated, wherein the different type recognition task information is used to indicate object type recognition under different classification dimensions.

[0008] Secondly, embodiments of this disclosure also provide an object processing method, including:

[0009] In response to a type recognition instruction for a preset object, the preset object is subjected to object type recognition based on different type recognition task information to obtain the type recognition result of the preset object. The different type recognition task information is generated based on type recognition labels and label description information under different classification dimensions. The different type recognition task information is used to indicate object type recognition under different classification dimensions.

[0010] The preset object is processed based on the type recognition result.

[0011] Thirdly, embodiments of this disclosure also provide a task information generation apparatus, including:

[0012] The information acquisition module is used to acquire classification requirement information;

[0013] The tag generation module is used to perform classification dimension analysis based on the classification requirement information, and generate type tags under different classification dimensions and tag description information of the type tags, wherein different classification dimensions correspond to different type tags and tag description information of the type tags;

[0014] The information generation module is used to generate type recognition task information corresponding to the type label based on the type label and the label description information under the different classification dimensions for at least one of the different classification dimensions. The different type recognition task information is used to indicate object type recognition under different classification dimensions.

[0015] Fourthly, embodiments of this disclosure also provide an object processing apparatus, including:

[0016] The type recognition module is used to respond to the type recognition instruction for the preset object, and perform object type recognition on the preset object based on different type recognition task information to obtain the type recognition result of the preset object. The different type recognition task information is generated based on type recognition labels and label description information under different classification dimensions. The different type recognition task information is used to indicate the object type recognition under different classification dimensions.

[0017] The object processing module is used to process the preset object based on the type identification result.

[0018] Fifthly, embodiments of this disclosure also provide an electronic device, including:

[0019] One or more processors;

[0020] Memory, used to store one or more programs.

[0021] When the one or more programs are executed by the one or more processors, the one or more processors implement the task information generation method or object processing method as described in the embodiments of this disclosure.

[0022] Sixthly, embodiments of this disclosure also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the task information generation method or object processing method as described in embodiments of this disclosure.

[0023] In a seventh aspect, embodiments of this disclosure also provide a computer program product that, when executed by a computer, enables the computer to implement the task information generation method or object processing method as described in embodiments of this disclosure.

[0024] The task information generation method, object processing method, electronic device, storage medium, and program product provided in this disclosure first generate type labels and label description information of type labels based on classification requirement information. Then, type recognition task information corresponding to the type label is generated based on the type label and label description information. Object type recognition is performed based on the type recognition task information under different classification dimensions, and object processing is performed based on the object type recognition results. This enables the generation of type recognition tasks under different classification dimensions, improves the recognition accuracy of the generated type recognition task information and the accuracy of object type recognition based on the type recognition task information, and further improves the accuracy of object processing based on the type recognition results. Attached Figure Description

[0025] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0026] Figure 1 A flowchart illustrating a task information generation method provided in this embodiment of the disclosure;

[0027] Figure 2 A flowchart illustrating an object processing method provided in an embodiment of this disclosure;

[0028] Figure 3 This is a schematic diagram illustrating the process of generating marking task information according to an embodiment of the present disclosure;

[0029] Figure 4 A schematic diagram illustrating a content marking process provided in an embodiment of this disclosure;

[0030] Figure 5 A structural block diagram of a task information generation device provided in an embodiment of this disclosure;

[0031] Figure 6 This is a structural block diagram of an object processing apparatus provided in an embodiment of the present disclosure;

[0032] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0033] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0034] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0035] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below.

[0036] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0037] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0038] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0039] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0040] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the software or hardware, such as the electronic device, application, server, or storage medium performing the operations of this disclosed technical solution, based on the prompt message.

[0041] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0042] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0043] It is understood 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 related provisions.

[0044] Figure 1 This is a flowchart illustrating a task information generation method provided in an embodiment of this disclosure. The method can be executed by a task information generation device, which can be implemented in software and / or hardware and can be configured in an electronic device, typically a computer, mobile phone, or tablet. The task information generation method provided in this disclosure is applicable to scenarios involving the generation of type identification task information under different classification dimensions. Figure 1 As shown, the task information generation method provided in this embodiment may include:

[0045] S101. Obtain classification requirement information.

[0046] The classification requirement information can be used to describe the needs of the relevant user (such as object processing personnel) for the generated type tags, such as the requirements or demands for the generated type tags. Type tags can be understood as tags used to label the type of a corresponding object, such as tags used to label the type of anomalies that have occurred in the corresponding object. Taking a tagging scenario as an example, the classification requirement information can be the requirements of the relevant personnel for the generated tagging labels, such as tagging requirement information. Tagging requirement information can be descriptive information about the tagging requirements. In some examples, tagging requirement information may include, but is not limited to, at least one of tagging requirement information and tagging example information. This tagging requirement information can be used to describe the requirements for the generated packaged tags, such as the quantity or format requirements for the generated tagging labels, etc. This tagging example can be understood as a case that can be used for reference when performing classification dimension analysis. For example, tagging examples may include tagging example examples and tag description information examples corresponding to the packaged tag examples. The number of tagging examples can be one or more.

[0047] In this embodiment, classification requirement information can be obtained, such as the classification requirement information input by the user when a task information generation instruction is received, so as to generate type labels and description information of type labels based on this classification requirement information.

[0048] S102. Perform classification dimension analysis based on the classification requirement information to generate type labels under different classification dimensions and label description information of the type labels, wherein different classification dimensions correspond to different type labels and label description information of the type labels.

[0049] Type labels can be understood as tags used to identify object types, such as type identifiers. Different type labels may correspond to the same or different classification dimensions. A single classification dimension can have one or more type labels, and type labels for different classification dimensions may be different. This classification dimension can be set as needed; for example, it may include dimensions of object anomalies, with different classification dimensions corresponding to different object anomalies. In some examples, type labels can be used to indicate whether an object has an anomaly under a specific dimension. For instance, when an anomaly is detected in an object under a certain dimension, a type label corresponding to this anomaly can be added to the object to indicate the existence of the anomaly corresponding to this type label. The label description information of a type label can be used to describe the meaning of the type label and / or the type characteristics of the corresponding type. The label description information for type labels under different classification dimensions may be different.

[0050] For example, after obtaining the classification requirement information, classification dimension analysis can be performed based on this classification requirement information. For instance, classification dimensions can be decomposed based on the classification requirement information to obtain type labels under different dimensions that match this classification requirement information, as well as label description information for each type label.

[0051] Taking the tagging scenario as an example, after obtaining the target requirement information (i.e., classification requirement information), tagging dimension analysis can be performed based on this tagging requirement information. Multiple tagging dimensions can be extracted from this tagging requirement information, and tagging labels (i.e. type labels) under each tagging dimension and label description information of each tag can be generated.

[0052] In this embodiment, the generation method of type labels and their descriptions is not limited. For example, type labels and their descriptions can be generated based on a pre-set label generation strategy; alternatively, a pre-trained second model can be used to generate type labels and their descriptions.

[0053] Optionally, the step of performing classification dimension analysis based on the classification requirement information to generate type labels under different classification dimensions and label description information of the type labels includes: generating a second prompt word based on the classification requirement information and a second instruction, wherein the second instruction is used to instruct the performance of classification dimension analysis; inputting the second prompt word into a second model, and generating type labels under different classification dimensions and label description information of the type labels through the second model based on the second prompt word.

[0054] The second instruction is used to instruct the performance of classification dimension analysis. This second instruction can be pre-set or generated based on a pre-set instruction generation strategy. The second prompt word can be a prompt word that triggers the second model to perform classification dimension analysis. The second model can be a model used to generate type labels and label descriptions based on the user's classification needs. Taking a tagging scenario as an example, the second model can be used to generate tagging labels and label descriptions. The type of the second model is not limited; it can be a large language model, or other models besides a large language model, such as a deep learning model. The following explanation uses a large language model as an example for the second model.

[0055] Specifically, the acquired classification requirement information can be input into a pre-trained second model, which then performs classification dimension analysis based on the word classification requirement information. For example, a second instruction can be acquired or generated; a second prompt word can be generated based on the acquired second instruction and classification requirement information, for example, by concatenating the second instruction and classification requirement information into a second prompt word according to a pre-set splicing method; then, this second prompt word can be input into the second model, which performs classification dimension analysis based on it. For example, the second model responds to the second instruction in the second prompt word, performs classification dimension analysis based on the classification requirement information in the second prompt word, and generates type labels under one or more classification dimensions, as well as label description information for each type label, to further improve the rationality and practicality of the generated type labels.

[0056] S103. For at least one of the different classification dimensions, based on the type label and the label description information under the classification dimension, generate type recognition task information corresponding to the type label, wherein the different type recognition task information is used to indicate object type recognition under different classification dimensions.

[0057] The type recognition task information can be task information used to instruct on object type recognition. For example, the task objective indicated by the type recognition task information can be to identify the type of an object under the corresponding classification dimension. Taking the scenario of tagging anomalies that occur in an object as an example, the type recognition task information corresponding to a certain type label can be used to indicate whether the object has an anomaly corresponding to this type label, and can add this type label to the object when it is determined that the object has an anomaly corresponding to this type label. Different type recognition task information can correspond to different type labels, used to instruct on the type recognition of the object under the classification dimension corresponding to this type label, in order to determine whether it is necessary to add its corresponding type label to the object. Taking the tagging scenario as an example, this type recognition task information can be tagging task information, and different tagging task information can correspond to different tagging labels, used to instruct on whether it is necessary to add this tag to the corresponding object, and can further add this tag to the object when it is determined that it is necessary to add this tag.

[0058] Specifically, after generating type labels and label descriptions under different classification dimensions, for each classification dimension or at least some classification dimensions, type recognition task information corresponding to the type labels under this classification dimension can be generated based on the type labels and label descriptions under this classification dimension, so as to facilitate subsequent identification and determination of whether the corresponding object needs to be labeled with this type through this type recognition task information.

[0059] In this embodiment, the method for generating type recognition task information can be set as needed. For example, type recognition task information corresponding to type labels under each classification dimension can be generated based on a pre-set task information generation strategy; alternatively, a pre-trained third model can be used to generate type recognition task information corresponding to type labels under each classification dimension.

[0060] Optionally, generating type recognition task information corresponding to the type label based on the type label and the label description information under the classification dimension includes: generating a third prompt word based on the type label, the label description information and a third instruction under the classification dimension, wherein the third instruction is used to instruct the generation of type recognition task information; inputting the third prompt word into a third model, and generating type recognition task information corresponding to the type label through the third model based on the third prompt word.

[0061] The third instruction is used to generate type recognition task information. The third instruction can be pre-set or generated based on a pre-set instruction generation strategy. The third prompt word can be a prompt word that triggers the third model to perform classification dimension analysis. The third model can be a model used to generate type recognition task information corresponding to the type label based on the type label and its description information. The type of the third model is not limited; for example, the third model can be a large language model or other models besides a large language model. The following explanation uses a large language model as an example. The third model and the second model can be the same or different models; this embodiment does not limit this.

[0062] Specifically, for each classification dimension or at least some classification dimensions with classification labels, the type label and label description information under this classification dimension are input into a pre-trained third model, and the third model generates type recognition task information under this classification dimension based on this classification label and label description information.

[0063] For example, a third instruction can be acquired or generated. Furthermore, a third prompt word is generated based on the acquired third instruction, the category label and label description information for this category dimension; and this third prompt word is input into a third model, which generates type recognition task information for this category dimension based on the third prompt word. For instance, in response to the third instruction in the third prompt word, the third model generates type recognition task information corresponding to the type label based on the type label in the third prompt word and the label description information of the type label, thereby further improving the recognition accuracy of the generated type recognition task information.

[0064] In this embodiment, the content of the type recognition task information is not limited. For example, the type recognition task information may include type recognition code. For instance, type recognition code corresponding to a type label can be generated. By running this type recognition code, the corresponding object can be type-recognized under the classification dimension corresponding to the type label, such as determining whether a type recognition label needs to be added to the corresponding object. Alternatively, the type recognition task information may include a first instruction. For instance, a first instruction corresponding to a type label can be generated. This first instruction can be used to instruct the corresponding object to be type-recognized under the classification dimension corresponding to the type recognition label, such as indicating whether a corresponding type recognition label needs to be added. Thus, the type recognition task information can be executed based on this first instruction. In this case, optionally, the type recognition task information includes a first instruction. Generating the type recognition task information corresponding to the type label includes: generating a first instruction corresponding to the type label and outputting the first instruction. The first instruction is used to instruct the first model to perform object type recognition under the classification dimension corresponding to the type label. The first model can be used to perform object type recognition based on the type recognition task information to determine the object type of the corresponding object under the classification dimension corresponding to the type recognition task information. The first model and the second and third models can be the same or different models. The type of the first model is not limited. For example, the first model can be a large language model; or other models besides large language models, such as deep learning models. The following explanation uses a large language model as the first model.

[0065] In some implementations, considering that the type label generation method and type recognition task information generation method (such as strategy or model) adopted in this embodiment are feasible generation methods with high accuracy, after generating the type recognition task information corresponding to each type recognition label, the generated type recognition task information can be directly applied. For example, the generated type recognition task information can be deployed. After deployment, object type recognition can be performed based on the generated type recognition task information without testing the generated type recognition task information, thereby further improving the generation and application speed of type recognition task information and reducing the resources occupied by generating type recognition task information.

[0066] In some implementations, after generating type recognition task information, the generated type recognition task information can be tested using test samples in a test set to determine the recognition accuracy of the type recognition task information. The generated type recognition task information is only applied when its recognition accuracy is high (e.g., it meets a preset accuracy condition). When its recognition accuracy is low (e.g., it does not meet the preset accuracy condition), the type labels and type recognition task information are regenerated until the recognition accuracy of the generated type recognition task information is high, so as to further improve the recognition accuracy of the generated type recognition task information, and thus further improve the accuracy of object type recognition based on the type recognition task information.

[0067] Optionally, after generating the type recognition task information corresponding to the classification dimension, the method further includes: testing at least a portion of the type recognition task information using test samples from a test set, and determining the recognition accuracy of the at least a portion of the type recognition task information based on the test results of the type recognition task information, wherein the test set includes multiple test samples; determining that the generation of the type recognition task information is complete in response to the recognition accuracy meeting a preset accuracy condition; and regenerating type labels and label description information for different classification dimensions, and regenerating the type recognition task information corresponding to the type labels, in response to the recognition accuracy of the at least a portion of the type recognition task information not meeting the preset accuracy condition.

[0068] The test set can be understood as a collection of test samples used to test the generated type recognition task information. The test set includes multiple test samples, each of which can be used to test various types of generated type recognition task information. The test set can be pre-set or input by the user; this embodiment does not limit this. Different test samples can correspond to different test object samples. For example, each test sample may include a test object sample and the actual first object type information of that test object sample. The test result of the type recognition task information can be understood as the type recognition result of the type recognition task information on the test object samples in one or more test samples, such as the second object type information of the corresponding test object sample identified and determined by the type recognition task information.

[0069] The preset accuracy conditions can be flexibly set as needed. For example, preset accuracy conditions may include: the recognition accuracy corresponding to each type of recognition task information is higher than a preset accuracy threshold; the recognition accuracy of each test sample is higher than a preset accuracy threshold; the proportion of recognition task information with a recognition accuracy higher than the preset accuracy threshold in each type of recognition task information reaches a preset proportion threshold; or, the proportion of test samples with a recognition accuracy higher than the preset accuracy threshold in each test sample reaches a preset proportion threshold, etc. Both the preset accuracy threshold and the preset proportion threshold can be set as needed, such as setting the preset accuracy threshold to 0.99 or 0.98, etc.; and / or setting the preset proportion threshold to 0.95 or 0.96, etc.

[0070] For example, for each or at least a portion of the test samples in the test set, type recognition tests can be performed on the test object samples in the test sample based on the generated type recognition task information, generating test results (such as the type of the identified object) for the test object sample. Based on the actual object type contained in the test sample, it can be determined whether the test results of the type recognition test on the test object sample based on the type recognition task information are accurate. Then, the recognition accuracy corresponding to each type recognition task information can be determined based on whether the test results of each test sample are accurate. If the recognition accuracy meets the preset accuracy conditions (such as being greater than the preset accuracy threshold), it can be determined that the generated type recognition task information can be applied, that is, the type recognition task information generation is complete. If the recognition accuracy does not meet the preset accuracy conditions (such as being less than or equal to the preset accuracy threshold), the generated type labels and the type recognition task information corresponding to each type label can be discarded, and the type labels and the type recognition task information corresponding to each type label can be regenerated.

[0071] In the above embodiments, when regenerating type labels, erroneous test samples in the test set may or may not be referenced. In some examples, when regenerating type labels, erroneous test samples in the test set may be referenced. For example, erroneous test samples in the test set can be used as reference cases. Based on the test requirement information and these erroneous test samples, classification dimension analysis is performed to regenerate type labels and label description information for each type label that match the test requirement information under different classification dimensions, thereby further improving the accuracy of the regenerated type labels and label description information. In this case, optionally, the regeneration of multiple type labels and label description information for the type labels includes: the regeneration of type labels and label description information for different classification dimensions includes: based on the classification requirement information and the erroneous test samples in the test set, classification dimension analysis is performed to regenerate type labels and label description information for different classification dimensions, wherein the erroneous test samples are test samples that have not been correctly identified based on the type identification task information. Among them, the error test sample can be understood as the test sample in the test set that did not correctly perform type identification based on the type identification task information. It can include test samples that did not correctly perform type identification in one or more tests before this regeneration of type labels, such as test samples that did not correctly perform type identification based on the type identification task information generated in each test before this regeneration of type labels.

[0072] In the above embodiments, when regenerating type recognition task information, erroneous test samples in the test set may or may not be referenced. In some examples, when regenerating type recognition task information, erroneous test samples in the test set may be referenced. For example, erroneous test samples in the test set can be used as reference cases. Based on the regenerated type recognition labels, the regenerated label description information, and the erroneous test samples, the type recognition task information corresponding to each type label is regenerated to further improve the recognition accuracy of the regenerated type recognition task information. In this case, optionally, regenerating the type recognition task information corresponding to the type label includes: regenerating the type recognition task information corresponding to the type label based on the erroneous test samples in the test set, the type label, and the label description information.

[0073] The task information generation method provided in this embodiment obtains classification requirement information; performs classification dimension analysis based on this classification requirement information to generate type labels and label descriptions for at least one classification dimension, wherein different classification dimensions correspond to different type labels and label descriptions; for at least one of the above different classification dimensions, based on the type labels and label descriptions for this classification dimension, type recognition task information corresponding to this type label is generated, wherein different type recognition task information is used to indicate object type recognition under different classification dimensions. This embodiment utilizes the above technical solution to first generate type labels and label descriptions based on classification requirement information, and then generate type recognition task information corresponding to this type label based on this type label and label description information. This enables the generation of type recognition task information under different classification dimensions, improves the recognition accuracy of the generated type recognition task information, and thus improves the accuracy of object type recognition based on type recognition task information under different classification dimensions.

[0074] Figure 2 This is a flowchart illustrating an object processing method provided in an embodiment of this disclosure. The method can be executed by an object processing device, which can be implemented in software and / or hardware and can be configured in an electronic device, typically a computer, mobile phone, or tablet computer. The object processing method provided in this disclosure is applicable to scenarios where object type identification is performed based on pre-generated type identification task information. Figure 1 As shown, the object processing method provided in this embodiment may include:

[0075] S201. In response to a type recognition instruction for a preset object, the preset object is subjected to object type recognition based on different type recognition task information to obtain the type recognition result of the preset object. The different type recognition task information is generated based on type recognition labels and label description information under different classification dimensions. The different type recognition task information is used to indicate object type recognition under different classification dimensions.

[0076] In this context, a type identification instruction can be understood as an instruction to perform object type identification. The default object can be the object corresponding to this type identification instruction, which in turn instructs the object to be type-identified.

[0077] Specifically, upon receiving a type recognition instruction for a certain object (such as a preset object), the system can respond to this instruction by using at least a portion of the generated type recognition task information to perform object type recognition on the object, thereby obtaining the type recognition result for the preset object. For example, the system can use the generated type recognition task information to perform object type recognition on the object under different classification dimensions, thereby obtaining the object type of the preset object under different classification dimensions, and generating the type recognition result of the preset object based on the object type of the preset object under different classification dimensions.

[0078] In this embodiment, the execution method of the type recognition task information is not limited. For example, the generated type recognition task information can be executed directly, such as executing the type recognition task information based on a preset task information execution strategy; or the first model can be called to execute the generated type recognition task information, and so on.

[0079] In some implementations, taking the type recognition task information including a first instruction and calling a first model to execute the generated type recognition task information as an example, optionally, the type recognition task information includes a first instruction, which instructs the first model to perform object type recognition under the classification dimension corresponding to the type label. The step of performing object type recognition on the preset object based on different type recognition task information includes: performing object type recognition on the preset object based on different first instructions. For example, the first instruction and the object information of the preset object can be input into the first model. Thus, the first model can respond to the input first instruction, and based on the input object information, perform type recognition on the preset object under the classification dimension corresponding to the first instruction, determining the object type of the preset object under this classification dimension. The object information of the preset object can be understood as information related to type recognition of the preset object, such as one or more attribute information of the preset object. Taking the preset object as media content as an example, the object information of the preset object can, for instance, include the content data of this media content.

[0080] In some implementations, the step of identifying the object type of the preset object based on different first instructions includes: for at least one of the different first instructions, generating a first prompt word based on the first instruction and the object information of the preset object, inputting the first prompt word into a first model, and identifying the object type of the preset object based on the first prompt word through the first model.

[0081] For example, a first prompt word can be generated based on a first instruction under at least one pre-generated classification dimension and the object information of a preset object. For instance, the first instruction and the object information of the preset object can be concatenated into a first prompt word according to a pre-set concatenation method. This first prompt word is then input into a first model. The first model performs object type recognition on the preset object based on the input first prompt word. For example, the first model responds to the first instruction in the first prompt word and performs object type recognition on the preset object based on the object information of the preset object in the first prompt word, determining whether the preset object needs to have a type label corresponding to this first instruction added, thus obtaining the recognition result for this type recognition task information. After obtaining the recognition results for each type recognition task information, the object recognition result of the first object can be generated based on each recognition result.

[0082] In some implementations, the type label may include an anomaly label, and the type identification task information can be used to indicate anomaly identification of a preset object under different dimensions to determine the anomaly that occurs in the preset object. Optionally, the type label includes an anomaly label, and the different type identification task information includes first type identification task information and second type identification task information. The step of performing object type identification on the preset object based on the different type identification task information to obtain the type identification result of the preset object includes: performing object anomaly identification on the preset object using the first type identification task information and the second type identification task information respectively to obtain a first anomaly identification result of the preset object in a first dimension and a second anomaly identification result of the preset object in a second dimension; generating a third anomaly identification result of the preset object based on the first anomaly result and the second anomaly result as the type identification result of the preset object.

[0083] Here, an anomaly label can be understood as a label used to identify an anomaly that has occurred in an object. For example, an anomaly label can be used to indicate that a corresponding object (such as a preset object) has an anomaly corresponding to this anomaly label. The first type identification task information and the second type identification task information can be type identification task information under different classification dimensions generated. The number of generated type identification task information can be multiple, and these multiple type identification task information can include at least the first type identification task information and the second type identification task information. The first dimension can be the classification dimension corresponding to the first type identification task information, such as the anomaly identification dimension corresponding to the first type identification task information. The second dimension can be the classification dimension corresponding to the second type identification task information, such as the anomaly identification dimension corresponding to the second type identification task information. The first dimension and the second dimension can be different classification dimensions. The first anomaly identification result can be the anomaly identification result generated by performing anomaly identification on the preset object based on the first anomaly identification task information. The second anomaly identification result can be the anomaly identification result generated by performing anomaly identification on the preset object based on the second type identification task information. The third anomaly identification result can be the anomaly identification result of the preset object generated based on the anomaly identification results of each type identification task information (including the first type identification task information and the second type identification task information).

[0084] For example, anomaly identification can be performed on a preset object based on first type identification task information to generate a first anomaly identification result for the preset object in a first dimension; and anomaly identification can be performed on the preset object based on second type identification task information to generate a second anomaly identification result for the preset object in a second dimension. Then, a third anomaly identification result for the preset object can be generated based on the first and second anomaly identification results, such as obtaining the union of the first and second anomaly identification results as the third anomaly identification result for the preset object, and this third anomaly identification result can be used as the type identification result for the preset object.

[0085] S202. Process the preset object based on the type recognition result.

[0086] In this step, after obtaining the type identification result of the preset object, the preset object can be processed using the processing method corresponding to this type identification result, such as marking the preset object or other processing.

[0087] Taking media content as a preset object as an example, it can be demonstrated that the presence of anomalies in the media content can be determined based on the media content type identification result, and when the media content type identification result indicates that the media content is abnormal, the user who published the media content can be prompted to modify the media content, and so on.

[0088] The object processing method provided in this embodiment, in response to a type identification instruction for a preset object, performs object type identification on the preset object based on different type identification task information to obtain the type identification result of the preset object. The different type identification task information is generated based on type identification labels and label description information under different classification dimensions, and is used to indicate object type identification under different classification dimensions. The preset object is then processed based on the type identification result. This embodiment utilizes the above technical solution to perform object type identification on the preset object based on type identification task information under different classification dimensions, obtain the type identification result of the preset object, and process the preset object based on this type identification result. This improves the accuracy of the type identification result, thereby further improving the accuracy of processing the preset object.

[0089] Figure 3 This is a schematic diagram illustrating the process of generating marking task information according to an embodiment of this disclosure. In some optional embodiments, such as... Figure 3 As shown, taking the anomaly tagging scenario of content and the first, second, and third models being the same large language model as an example, the generation process of tagging task information (i.e., type recognition task information) can be described as follows:

[0090] After receiving the user's classification requirements and test set for anomaly labels via the user interface, this classification requirement information can be input into a Large Language Model (MLM). The MLM performs classification dimension analysis and label decomposition based on this requirement, generating multiple anomaly labels and their descriptions. These generated anomaly labels and their descriptions can then be input back into the MLM. The MLM then generates labeling task information based on these labels and descriptions, and the corresponding labeling task information for each anomaly label is obtained from the MLM's output. After obtaining the tagging task information corresponding to each anomaly label, the user-input test set can be used to verify (i.e., test) each tagging task information. If the verification passes, the generated tagging task information can be deployed online to facilitate subsequent anomaly tagging of content based on the generated tagging task information. If the verification fails, the test sample with anomaly identification error during verification can be used as a case, and the process of performing intent understanding and tag decomposition based on the classification requirements using this large language model can be returned until the generated tagging task information passes verification.

[0091] Figure 4 This is a schematic diagram of a content marking process provided in an embodiment of the present disclosure, such as... Figure 4As shown, after the labeling task information is deployed online, when there is content that needs to be labeled with anomalies, anomaly labeling can be performed on this content based on the information from each labeling task. For example, a large language model can be called to identify whether the content has anomalies corresponding to the anomaly tags under different dimensions based on the information from each labeling task, generating anomaly identification results for this content. Afterwards, the content can be processed based on these anomaly identification results. For example, it can be determined whether the content has anomalies; if so, the content can be marked as anomalous content, such as adding anomaly tags corresponding to the anomalies it contains; otherwise, the content can be marked as normal content.

[0092] The technical solution provided in this embodiment automatically splits multiple abnormal tags based on the user's classification requirements, generates at least one tagging task information corresponding to an abnormal tag, and tags the content for abnormality using each tagging task information. Compared with related technologies that use models to perform binary classification of content based on all abnormal features to determine whether there is an anomaly, and compared with related technologies that use keywords for anomaly identification, this solution eliminates the need for manual splitting and tagging of abnormal tags, reduces the human resources required for tag splitting and tagging, improves the accuracy of anomaly identification based on tagging task information, and thus improves the accuracy of content tagging.

[0093] Figure 5 This is a structural block diagram of a task information generation device provided in an embodiment of this disclosure. The device can be implemented by software and / or hardware, and can be configured in an electronic device, typically a computer, mobile phone, or tablet computer. It can generate type identification task information under different classification dimensions by executing a task information generation method. Figure 5 As shown, the task information generation device provided in this embodiment may include: an information acquisition module 501, a tag generation module 502, and an information generation module 503, wherein,

[0094] Information acquisition module 501 is used to acquire classification requirement information;

[0095] The tag generation module 502 is used to perform classification intent analysis based on the classification requirement information, and generate type tags under at least one classification dimension and tag description information of the type tags;

[0096] The information generation module 503 is used to generate type recognition task information corresponding to the type label based on the type label and the label description information, wherein different type recognition task information is used to perform object type recognition under different classification dimensions.

[0097] The task information generation device provided in this embodiment acquires classification requirement information through an information acquisition module; performs classification dimension analysis based on this classification requirement information through a tag generation module, generating type tags and tag descriptions for at least one classification dimension, wherein different classification dimensions correspond to different type tags and tag descriptions; and generates type recognition task information corresponding to at least one of the different classification dimensions based on the type tags and tag descriptions for that classification dimension, wherein different type recognition task information is used to indicate object type recognition under different classification dimensions. This embodiment utilizes the above technical solution to first generate type tags and tag descriptions based on the classification requirement information, and then generate type recognition task information corresponding to that type tag based on the type tags and tag descriptions. This enables the generation of type recognition task information under different classification dimensions, improves the recognition accuracy of the generated type recognition task information, and thus improves the accuracy of object type recognition based on type recognition task information under different classification dimensions.

[0098] Furthermore, the task information generation device may further include: a testing module, configured to, after generating type recognition task information corresponding to the classification dimension, test at least a portion of the type recognition task information using test samples from a test set, and determine the recognition accuracy of the at least a portion of the type recognition task information based on the test results, wherein the test set includes multiple test samples; in response to the recognition accuracy meeting a preset accuracy condition, determine that the generation of the type recognition task information is complete; in response to the recognition accuracy of the at least a portion of the type recognition task information not meeting the preset accuracy condition, regenerate type labels under different classification dimensions and label description information of the type labels, and regenerate type recognition task information corresponding to the type labels.

[0099] Optionally, the testing module may be specifically used to: perform classification dimension analysis based on the classification requirement information and the erroneous test samples in the test set, regenerate type labels under different classification dimensions and label description information of the type labels, wherein the erroneous test samples are test samples that have not been correctly identified based on the type identification task information.

[0100] Optionally, the testing module may be specifically used to: regenerate type recognition task information corresponding to the type label based on the erroneous test samples in the test set, the type label, and the label description information.

[0101] Optionally, the type recognition task information includes a first instruction, and the information generation module 503 can be specifically used to: generate a first instruction corresponding to the type label and output the first instruction, wherein the first instruction is used to instruct the first model to perform object type recognition under the classification dimension corresponding to the type label.

[0102] Optionally, the tag generation module 502 may be specifically used to: generate a second prompt word based on the classification requirement information and the second instruction, the second instruction being used to instruct the classification dimension analysis; input the second prompt word into the second model, and generate type tags under different classification dimensions and tag description information of the type tags based on the second prompt word through the second model.

[0103] Optionally, the information generation module 503 may be specifically used to: generate a third prompt word based on the type label under the classification dimension, the label description information, and a third instruction, wherein the third instruction is used to instruct the generation of type recognition task information; input the third prompt word into a third model, and generate type recognition task information corresponding to the type label based on the third prompt word through the third model.

[0104] The task information generation apparatus provided in this disclosure can execute the task information generation method provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects for executing the task information generation method. Technical details not described in detail in this embodiment can be found in the task information generation method provided in any embodiment of this disclosure.

[0105] Figure 6 This is a structural block diagram of an object processing apparatus provided in an embodiment of the present disclosure. The apparatus can be implemented by software and / or hardware, and can be configured in an electronic device, typically a computer, mobile phone, or tablet computer. It can perform object type identification based on pre-generated type identification task information by executing an object processing method. Figure 6 As shown, the object processing device provided in this embodiment may include: a type identification module 601 and an object processing module 602, wherein,

[0106] Type recognition module 601 is used to respond to a type recognition instruction for a preset object, perform object type recognition on the preset object based on different type recognition task information, and obtain the type recognition result of the preset object. The different type recognition task information is generated based on type recognition labels and label description information under different classification dimensions. The different type recognition task information is used to indicate object type recognition under different classification dimensions.

[0107] The object processing module 602 is used to process the preset object based on the type recognition result.

[0108] The object processing device provided in this embodiment responds to a type recognition instruction for a preset object through a type recognition module. It performs object type recognition on the preset object based on different type recognition task information, obtaining the type recognition result of the preset object. The different type recognition task information is generated based on type recognition labels and label description information under different classification dimensions, and is used to indicate object type recognition under different classification dimensions. The object processing module then processes the preset object based on the type recognition result. This embodiment utilizes the above technical solution to perform object type recognition on the preset object based on type recognition task information under different classification dimensions, obtaining the type recognition result of the preset object, and processing the preset object based on this type recognition result. This improves the accuracy of the type recognition result, thereby further improving the accuracy of processing the preset object.

[0109] Optionally, the type label includes an anomaly label, and the different type identification task information includes first type identification task information and second type identification task information. The type identification module 601 can be specifically used to: perform object anomaly identification on the preset object using the first type identification task information and the second type identification task information respectively, to obtain a first anomaly identification result of the preset object in a first dimension and a second anomaly identification result of the preset object in a second dimension; and generate a third anomaly identification result of the preset object based on the first anomaly result and the second anomaly result, as the type identification result of the preset object.

[0110] Optionally, the type recognition task information includes a first instruction, which is used to instruct the first model to perform object type recognition under the classification dimension corresponding to the type label. The type recognition module 601 can be specifically used to perform object type recognition on the preset object based on different first instructions.

[0111] Optionally, the type recognition module 601 may be specifically used to: generate task information description information for the type recognition task information corresponding to the type label; the type recognition module may be specifically used to: generate a first prompt word based on the first instruction and the object information of the preset object for at least one of the different first instructions, input the first prompt word into a first model, and perform object type recognition of the preset object based on the first prompt word through the first model.

[0112] The task information generation apparatus provided in this disclosure can execute the task information generation method provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects for executing the task information generation method. Technical details not described in detail in this embodiment can be found in the task information generation method provided in any embodiment of this disclosure.

[0113] The following is for reference. Figure 7 The diagram illustrates a structural schematic of an electronic device (e.g., a server) 700 suitable for implementing embodiments of the present disclosure. Terminal devices in embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 7 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0114] like Figure 7 As shown, the electronic device 700 may include a processing unit (e.g., a central processing unit, a graphics processor, etc.) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage device 708 into a random access memory (RAM) 703. The RAM 703 also stores various programs and data required for the operation of the electronic device 700. The processing unit 701, ROM 702, and RAM 703 are interconnected via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0115] Typically, the following devices can be connected to I / O interface 705: input devices 706 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 707 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 708 including, for example, magnetic tapes, hard disks, etc.; and communication devices 709. Communication device 709 allows electronic device 700 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 7 An electronic device 700 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0116] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 709, or installed from storage device 708, or installed from ROM 702. When the computer program is executed by processing device 701, it performs the functions defined in the methods of embodiments of this disclosure.

[0117] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, 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 device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0118] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0119] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0120] The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire classification requirement information; acquire classification requirement information; perform classification dimension analysis based on the classification requirement information, generating type labels under different classification dimensions and label description information for the type labels, wherein different classification dimensions correspond to different type labels and label description information for the type labels; for at least one of the different classification dimensions, based on the type labels and label description information under the classification dimension, generate type recognition task information corresponding to the type labels, wherein different type recognition task information is used to instruct object type recognition to be performed under different classification dimensions. Alternatively,

[0121] In response to a type recognition instruction for a preset object, object type recognition is performed on the preset object based on different type recognition task information to obtain the type recognition result of the preset object. The different type recognition task information is generated based on type recognition labels and label description information under different classification dimensions, and is used to indicate object type recognition under different classification dimensions. The preset object is then processed based on the type recognition result.

[0122] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including but not limited to object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0123] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0124] The units described in the embodiments of this disclosure can be implemented in software or hardware. The names of modules do not, in some cases, constitute a limitation on the unit itself.

[0125] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0126] 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.

[0127] According to one or more embodiments of this disclosure, Example 1 provides a task information generation method, including:

[0128] Obtain classification requirement information;

[0129] Based on the classification requirement information, a classification dimension analysis is performed to generate type labels under different classification dimensions and label description information for the type labels. Different classification dimensions correspond to different type labels and label description information for the type labels.

[0130] For at least one of the different classification dimensions, based on the type label and the label description information under the classification dimension, type recognition task information corresponding to the type label is generated, wherein the different type recognition task information is used to indicate object type recognition under different classification dimensions.

[0131] According to one or more embodiments of this disclosure, Example 2, based on the method of Example 1, further includes, after generating the type recognition task information corresponding to the classification dimension:

[0132] The test set includes multiple test samples. At least some of the type recognition task information is tested using test samples from the test set. The recognition accuracy of the at least some type recognition task information is determined based on the test results of the type recognition task information.

[0133] In response to the recognition accuracy meeting a preset accuracy condition, it is determined that the type recognition task information generation is complete;

[0134] In response to the fact that the recognition accuracy of at least some of the type recognition task information does not meet the preset accuracy condition, type labels under different classification dimensions and label description information of the type labels are regenerated, and type recognition task information corresponding to the type labels is regenerated.

[0135] According to one or more embodiments of this disclosure, Example 3, based on the method described in Example 2, includes regenerating type labels under different classification dimensions and label description information for the type labels, comprising:

[0136] Based on the classification requirement information and the erroneous test samples in the test set, a classification dimension analysis is performed to regenerate type labels under different classification dimensions and label description information for the type labels. The erroneous test samples are test samples that have not been correctly identified based on the type identification task information.

[0137] According to one or more embodiments of this disclosure, Example 4, based on the method described in Example 2, includes regenerating the type recognition task information corresponding to the type label, comprising:

[0138] Based on the erroneous test samples in the test set, the type labels, and the label description information, the type recognition task information corresponding to the type labels is regenerated.

[0139] According to one or more embodiments of this disclosure, Example 5 describes the method described in Example 1, wherein the type identification task information includes a first instruction, and the generation of type identification task information corresponding to the type label includes:

[0140] Generate a first instruction corresponding to the type label and output the first instruction, which is used to instruct the first model to perform object type recognition under the classification dimension corresponding to the type label.

[0141] According to one or more embodiments of this disclosure, Example 6 describes the method described in any of Examples 1-5, wherein the step of performing classification dimension analysis based on the classification requirement information to generate type labels under different classification dimensions and label description information for the type labels includes:

[0142] A second prompt word is generated based on the classification requirement information and the second instruction, and the second instruction is used to instruct the classification dimension analysis to be performed.

[0143] The second prompt word is input into the second model, and the second model generates type labels under different classification dimensions and label description information of the type labels based on the second prompt word.

[0144] According to one or more embodiments of this disclosure, Example 7, based on any one of Examples 1-5, the step of generating type recognition task information corresponding to the type label based on the type label and the label description information under the classification dimension includes:

[0145] A third prompt word is generated based on the type label under the classification dimension, the label description information, and the third instruction, wherein the third instruction is used to instruct the generation of type recognition task information;

[0146] The third prompt word is input into the third model, and the third model generates type recognition task information corresponding to the type label based on the third prompt word.

[0147] According to one or more embodiments of this disclosure, Example 8 provides an object processing method, including:

[0148] In response to a type recognition instruction for a preset object, the preset object is subjected to object type recognition based on different type recognition task information to obtain the type recognition result of the preset object. The different type recognition task information is generated based on type recognition labels and label description information under different classification dimensions. The different type recognition task information is used to indicate object type recognition under different classification dimensions.

[0149] The preset object is processed based on the type recognition result.

[0150] According to one or more embodiments of this disclosure, Example 9 describes the method described in Example 8, wherein the type label includes an anomaly label, the different type identification task information includes first type identification task information and second type identification task information, and the step of performing object type identification on the preset object based on the different type identification task information to obtain the type identification result of the preset object includes:

[0151] The preset object is subjected to object anomaly identification using the first type of identification task information and the second type of identification task information, respectively, to obtain the first anomaly identification result of the preset object in the first dimension and the second anomaly identification result of the preset object in the second dimension.

[0152] A third anomaly identification result for the preset object is generated based on the first anomaly result and the second anomaly result, and is used as the type identification result for the preset object.

[0153] According to one or more embodiments of this disclosure, Example 10 describes the method described in Example 8, wherein the type recognition task information includes a first instruction, the first instruction being used to instruct a first model to perform object type recognition under a classification dimension corresponding to the type label, and the step of performing object type recognition on the preset object based on different type recognition task information includes:

[0154] The preset object is identified based on different first instructions.

[0155] According to one or more embodiments of this disclosure, Example 11, based on the method described in Example 10, includes identifying the object type of the preset object based on different first instructions, comprising:

[0156] For at least one of the different first instructions, a first prompt word is generated based on the first instruction and the object information of the preset object. The first prompt word is input into a first model, and the object type of the preset object is identified by the first model based on the first prompt word.

[0157] According to one or more embodiments of this disclosure, Example 12 provides an electronic device, including:

[0158] One or more processors;

[0159] Memory, used to store one or more programs.

[0160] When the one or more programs are executed by the one or more processors, the one or more processors implement the task information generation method as described in any of Examples 1-7 or the object processing method as described in any of Examples 8-11.

[0161] According to one or more embodiments of the present disclosure, Example 13 provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements a task information generation method as described in any of Examples 1-7 or an object processing method as described in any of Examples 8-11.

[0162] According to one or more embodiments of this disclosure, Example 14 provides a computer program product that, when executed by a computer, causes the computer to implement the task information generation method as described in any of Examples 1-7 or the object processing method as described in any of Examples 8-11.

[0163] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

[0164] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0165] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A method for generating task information, characterized in that, include: Obtain classification requirement information; Based on the classification requirement information, a classification dimension analysis is performed to generate type labels under different classification dimensions and label description information for the type labels. Different classification dimensions correspond to different type labels and label description information for the type labels. For at least one of the different classification dimensions, based on the type label and the label description information under the classification dimension, type recognition task information corresponding to the type label is generated, wherein the different type recognition task information is used to indicate object type recognition under different classification dimensions.

2. The method according to claim 1, characterized in that, After generating the type recognition task information corresponding to the classification dimension, the method further includes: The test set includes multiple test samples. At least some of the type recognition task information is tested using test samples from the test set. The recognition accuracy of the at least some type recognition task information is determined based on the test results of the type recognition task information. In response to the recognition accuracy meeting a preset accuracy condition, it is determined that the type recognition task information generation is complete; In response to the fact that the recognition accuracy of at least some of the type recognition task information does not meet the preset accuracy condition, type labels under different classification dimensions and label description information of the type labels are regenerated, and type recognition task information corresponding to the type labels is regenerated.

3. The method according to claim 2, characterized in that, The process of regenerating type labels under different classification dimensions and the label description information of the type labels includes: Based on the classification requirement information and the erroneous test samples in the test set, a classification dimension analysis is performed to regenerate type labels under different classification dimensions and label description information for the type labels. The erroneous test samples are test samples that have not been correctly identified based on the type identification task information.

4. The method according to claim 2, characterized in that, The step of regenerating the type recognition task information corresponding to the type label includes: Based on the erroneous test samples in the test set, the type labels, and the label description information, the type recognition task information corresponding to the type labels is regenerated.

5. The method according to claim 1, characterized in that, The type recognition task information includes a first instruction, and the generation of type recognition task information corresponding to the type label includes: Generate a first instruction corresponding to the type label and output the first instruction, which is used to instruct the first model to perform object type recognition under the classification dimension corresponding to the type label.

6. The method according to any one of claims 1-5, characterized in that, The step of performing classification dimension analysis based on the classification requirement information to generate type labels under different classification dimensions and label description information for the type labels includes: A second prompt word is generated based on the classification requirement information and the second instruction, and the second instruction is used to instruct the classification dimension analysis to be performed. The second prompt word is input into the second model, and the second model generates type labels under different classification dimensions and label description information of the type labels based on the second prompt word.

7. The method according to any one of claims 1-5, characterized in that, The step of generating type recognition task information corresponding to the type label based on the type label and the label description information under the classification dimension includes: A third prompt word is generated based on the type label under the classification dimension, the label description information, and the third instruction, wherein the third instruction is used to instruct the generation of type recognition task information; The third prompt word is input into the third model, and the third model generates type recognition task information corresponding to the type label based on the third prompt word.

8. An object processing method, characterized in that, include: In response to a type recognition instruction for a preset object, the preset object is subjected to object type recognition based on different type recognition task information to obtain the type recognition result of the preset object. The different type recognition task information is generated based on type recognition labels and label description information under different classification dimensions. The different type recognition task information is used to indicate object type recognition under different classification dimensions. The preset object is processed based on the type recognition result.

9. The method according to claim 8, characterized in that, The type label includes an anomaly label, and the different type identification task information includes first type identification task information and second type identification task information. The step of performing object type identification on the preset object based on the different type identification task information to obtain the type identification result of the preset object includes: The preset object is subjected to object anomaly identification using the first type of identification task information and the second type of identification task information, respectively, to obtain the first anomaly identification result of the preset object in the first dimension and the second anomaly identification result of the preset object in the second dimension. A third anomaly identification result for the preset object is generated based on the first anomaly result and the second anomaly result, and is used as the type identification result for the preset object.

10. The method according to claim 8, characterized in that, The type recognition task information includes a first instruction, which instructs the first model to perform object type recognition under the classification dimension corresponding to the type label. The step of performing object type recognition on the preset object based on different type recognition task information includes: The preset object is identified based on different first instructions.

11. The method according to claim 10, characterized in that, The step of identifying the object type of the preset object based on different first instructions includes: For at least one of the different first instructions, a first prompt word is generated based on the first instruction and the object information of the preset object. The first prompt word is input into a first model, and the object type of the preset object is identified by the first model based on the first prompt word.

12. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the task information generation method of any one of claims 1-7 or the object processing method of any one of claims 8-11.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the task information generation method of any one of claims 1-7 or the object processing method of any one of claims 8-11.

14. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the task information generation method of any one of claims 1-7 or the object processing method of any one of claims 8-11.

Images