Abnormal information processing method and device, electronic equipment, storage medium and program

Abstract

The invention provides an abnormal information processing method and device, electronic equipment, a storage medium and a program. The method comprises the following steps: in response to detecting multiple pieces of attribute information of a current abnormal condition, packaging the multiple pieces of attribute information into multiple pieces of attribute data in a data structure body; generating a natural statement by using at least one item of target attribute data in each item of attribute data of the data structure body, wherein the natural statement is used for describing each item of target attribute data; judging whether the natural statement representation meets a preset display standard or not; and in response to determining that the display criteria are met, displaying the natural statement. According to the embodiment, the attribute information of the abnormal condition is packaged into the data structure body, so that all the attribute information is unified into a standardized expression form, all the attribute data with the standardized expression form in the data structure body can be used as effective context information in the process of reasoning the natural statement, and the efficiency of reasoning the natural statement is improved. Identification errors of attribute information in different forms are avoided, and a more accurate result is obtained.

Description

Technical Field

[0001] The embodiments of this application relate to the technical field of anomaly error detection, and in particular to a method, apparatus, electronic device, storage medium, and program for processing anomaly information. Background Technology

[0002] When users run business systems or perform tasks through electronic devices or processors, errors or exceptions often occur due to improper user operation or system problems, and error or exception messages are displayed to the user. These error or exception messages may include, for example, Server Internal Error, Unexpected Error, or garbled text.

[0003] Because such error messages or exception messages are usually presented in English, technical terms, code languages, or garbled characters, they are very difficult for non-technical users to understand. This makes it difficult for users to comprehend the specific error or exception that the error or exception indicates, nor can they know the cause of the error or exception, making it difficult to successfully resolve the error or exception. Summary of the Invention

[0004] In view of this, embodiments of the present disclosure provide a method, apparatus, electronic device, storage medium, and program for processing abnormal information.

[0005] In a first aspect, embodiments of this disclosure provide a method for processing abnormal information, the method comprising: In response to the detection of multiple attribute information of the current abnormal situation, the multiple attribute information is encapsulated into multiple attribute data in the data structure; Generate a natural statement using at least one target attribute data from each attribute data of the data structure. The natural statement is used to describe each target attribute data. Determine whether the natural language representation conforms to the preset display standards; In response to determining that the display criteria are met, natural language is displayed.

[0006] In some alternative implementations, natural language statements are generated using at least one target attribute from the attribute data of the data structure, including: Identify prompt words that include multiple indicators, among which at least one is a content indicator pointing to the target attribute data and at least one is a descriptive indicator constraining the format and semantic content of the statement; Determine the target attribute data corresponding to each content instruction from the data structure according to each content instruction; Generate prompt statements that point to the target attribute data according to the description instructions.

[0007] In some optional implementations, the target attribute data includes the error code corresponding to the current abnormal situation, scenario data, interface data, and operation data, and the description indication includes format indication for constraining word count and / or sentence format, and language indication for constraining semantic content. Accordingly, prompt statements are generated according to the description instructions, pointing to the target attribute data for each item, including: Map error codes to corresponding preset semantic tags; The system uses a pre-defined large language model to generate prompts that describe semantic tags, scene data, interface data, and operation data, and that conform to description and language instructions.

[0008] In some alternative implementations, the natural statement also includes an operation statement, and the multiple instructions also include at least one guiding instruction for defining the operation action; Accordingly, after generating prompt statements pointing to the target attribute data according to the description instructions, the method also includes: Based on the target attribute data of the current abnormal situation, determine the target operation action corresponding to the current abnormal situation from a number of preset operation actions; Generate operation statements that point to the target operation actions according to the various guidance instructions.

[0009] In some optional implementations, determining whether a natural language representation conforms to a preset display standard includes: Determine the semantic readability of natural language sentences and use it as the primary semantic indicator; Determine the emotional positivity of natural language statements and use it as a secondary semantic indicator; Determine the degree of introductory effect of natural language statements and use it as a third semantic indicator; Determine the semantic consistency between natural language statements and the target attribute data, and use it as the fourth semantic indicator; Determine the weighted result of the first semantic index, the second semantic index, the third semantic index, and the fourth semantic index; In response to determining that the weighted result is greater than or equal to a preset evaluation threshold, it is determined that the natural language representation meets the display criteria.

[0010] In some alternative implementations, after determining the weighted result of the first semantic index, the second semantic index, the third semantic index, and the fourth semantic index, the method further includes: In response to the determination that the weighted result is less than the evaluation threshold, according to the preset mapping relationship between multiple candidate statements and various abnormal situations, the candidate statement corresponding to the current abnormal situation is selected from multiple candidate statements and determined as a natural statement.

[0011] Secondly, embodiments of this disclosure provide an apparatus for processing abnormal information, the apparatus comprising: The encapsulation module is configured to encapsulate multiple attribute information into multiple attribute data in a data structure in response to the detection of current abnormal conditions. The generation module is configured to generate natural statements using at least one target attribute data from each attribute data of the data structure. The natural statements are used to describe each target attribute data. The judgment module is configured to determine whether the natural language representation conforms to the preset display standards; The presentation module is configured to display natural language in response to determining that the presentation criteria are met.

[0012] In some optional implementations, the generation module is further configured to: Identify prompt words that include multiple indicators, among which at least one is a content indicator pointing to the target attribute data and at least one is a descriptive indicator constraining the format and semantic content of the statement; Determine the target attribute data corresponding to each content instruction from the data structure according to each content instruction; Generate prompt statements that point to the target attribute data according to the description instructions.

[0013] The target attribute data includes the error code, scenario data, interface data, and operation data corresponding to the current abnormal situation. The description instructions include format instructions for constraining word count and / or sentence format, and language instructions for constraining semantic content. Accordingly, prompt statements are generated according to the description instructions, pointing to the target attribute data for each item, including: Map error codes to corresponding preset semantic tags; The system uses a pre-defined large language model to generate prompts that describe semantic tags, scene data, interface data, and operation data, and that conform to description and language instructions.

[0014] In some alternative implementations, the natural statement also includes an operation statement; the multiple instructions also include at least one guiding instruction for defining the operation action; Accordingly, the generation module is further configured to, after generating prompt statements pointing to target attribute data according to the description instructions, determine the target operation action corresponding to the current abnormal situation from a set of preset operation actions based on the target attribute data of the current abnormal situation. In some optional implementations, the determination module is further configured to: Determine the semantic readability of natural language sentences and use it as the primary semantic indicator; Determine the emotional positivity of natural language statements and use it as a secondary semantic indicator; Determine the degree of introductory effect of natural language statements and use it as a third semantic indicator; Determine the semantic consistency between natural language statements and the target attribute data, and use it as the fourth semantic indicator; Determine the weighted result of the first semantic index, the second semantic index, the third semantic index, and the fourth semantic index; In response to determining that the weighted result is greater than or equal to a preset evaluation threshold, it is determined that the natural language representation meets the display criteria.

[0015] Accordingly, the judgment module is further configured to, after determining the weighted result of the first semantic indicator, the second semantic indicator, the third semantic indicator and the fourth semantic indicator, in response to determining that the weighted result is less than the evaluation threshold, select the candidate statement corresponding to the current abnormal situation from the multiple candidate statements according to the preset mapping relationship between multiple preset candidate statements and various abnormal situations, and determine it as a natural statement.

[0016] Thirdly, embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, which, when executed by the one or more processors, cause the one or more processors to implement the method described in any implementation of the first aspect.

[0017] Fourthly, embodiments of this disclosure provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by one or more processors, implements the method as described in any implementation of the first aspect.

[0018] Fifthly, embodiments of this disclosure provide a computer program product, including a computer program / instructions that, when executed by a processor, implement the method described in any of the implementations of the first aspect.

[0019] To address the problem of users struggling to understand obscure and technical error messages or abnormal line prompts, the embodiments of this disclosure provide an error information processing method, apparatus, electronic device, storage medium, and program. Through a process of encapsulating and displaying standard verification, this achieves more accurate and understandable error prompts. The method first encapsulates multiple attribute information of the abnormal situation into a unified data structure, avoiding identification errors caused by different forms of attribute information. This ensures that the generated natural language can fully utilize the complete context, improving the accuracy of the statement description. Simultaneously, by judging whether the natural language conforms to preset display standards, prompt content that aligns with the user's comprehension is selected. This effectively solves the problems of traditional error prompts being obscure and difficult to understand, leaving users unclear about the cause of the error and the direction of processing. This allows users to quickly understand the abnormal situation and proceed with relevant processes. Furthermore, the standardized data encapsulation method provides a reliable foundation for the traceability and subsequent optimization of error information, improving the controllability and efficiency of the entire error handling process. Attached Figure Description

[0020] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings. The drawings are for illustrative purposes only and are not intended to limit the invention. In the drawings: Figure 1 This is an exemplary system architecture diagram to which one embodiment of this disclosure may be applied; Figure 2A A flowchart illustrating one embodiment of the method for handling abnormal information according to this disclosure; Figure 2B This is a breakdown flowchart of an embodiment of the breakdown process 2020 according to one embodiment of step 202 of this disclosure; Figure 2C This is a breakdown flowchart of an embodiment of the breakdown process 2030 according to step 203 of this disclosure; Figure 3 This is a schematic diagram of the structure of an embodiment of an anomaly information processing apparatus according to the present disclosure; Figure 4 A schematic diagram of the structure of a computer system suitable for implementing embodiments of the present disclosure. Detailed Implementation

[0021] The present disclosure will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other. This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] Figure 1 An exemplary system architecture 100 is shown, illustrating embodiments of methods, apparatuses, electronic devices, storage media, and programs for processing abnormal information according to this disclosure.

[0024] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0025] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as neural network applications, natural language processing applications, large model applications, speech recognition applications, applications running specified systems, and / or applications running various services, etc.

[0026] Terminal devices 101, 102, and 103 can be hardware or software. When terminal devices 101, 102, and 103 are hardware, they can be various electronic devices with sound acquisition devices (e.g., microphones), video acquisition devices (e.g., cameras), and displays, including but not limited to smartphones, tablets, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 players (Moving Picture Experts Group Audio Layer IV), laptops, and desktop computers, etc. Accordingly, terminal devices 101, 102, and 103 can be used to run business systems or execute task processes. In the event of operational errors or abnormalities, terminal devices 101, 102, and 103 can send the error or abnormality to server 105 and receive relevant prompts corresponding to the error or abnormality from server 105, thereby enabling terminal devices 101, 102, and 103 to display prompts related to the error or abnormality through their display devices, such as displays.

[0027] When terminal devices 101, 102, and 103 are software, they can be installed on the terminal devices listed above. They can be implemented as multiple software programs or software modules (e.g., services related to handling exception information) or as a single software program or software module. Terminal devices 101, 102, and 103 can be the business system or task process itself. In the event of an error or exception during their operation, they can send the error or exception to server 105 and receive relevant prompts corresponding to the error or exception from server 105. This allows terminal devices 101, 102, and 103 to display the prompts related to the error or exception through a display device such as a screen.

[0028] In some cases, the exception information handling method provided in this disclosure can be executed by terminal devices 101, 102, and 103. Accordingly, the exception information handling device can be set in terminal devices 101, 102, and 103. In this case, system architecture 100 may not include server 105, and when an error or exception occurs in the business system or task process, terminal devices 101, 102, and 103 will generate and display relevant prompts corresponding to the error or exception.

[0029] In some cases, the abnormal information processing method provided in this disclosure can be jointly executed by terminal devices 101, 102, and 103 and server 105. For example, the step of "displaying natural language in response to determining that it conforms to the display standard" can be executed by terminal devices 101, 102, and 103, and the step of "encapsulating multiple attribute information into multiple attribute data in a data structure in response to detecting multiple attribute information of the current abnormal situation" can be executed by server 105. This disclosure does not limit this. Accordingly, the abnormal information processing device can also be respectively set in terminal devices 101, 102, and 103 and server 105.

[0030] In some cases, the abnormal information processing method provided in this disclosure can be executed by server 105. Correspondingly, the abnormal information processing device can also be set in server 105. In this case, system architecture 100 may not include terminal devices 101, 102, and 103.

[0031] It should be noted that server 105 can be either hardware or software. When server 105 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. When server 105 is software, it can be implemented as multiple software programs or software modules (e.g., services related to providing distributed exception information processing), or as a single software program or software module. No specific limitations are made here.

[0032] It should be understood that Figure 1The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0033] In some alternative implementations, refer to Figure 2A The diagram illustrates a flow 200 of an embodiment of a method for processing abnormal information according to the present disclosure, which includes the following steps 201 to 204: Step 201: In response to the detection of multiple attribute information of the current abnormal situation, the multiple attribute information is encapsulated into multiple attribute data in a data structure.

[0034] Among them, abnormal situations can be unexpected errors or exceptions that occur during the operation of the business system or task process. For example, in the task process of running form submission, a form reading exception or output error occurs, or in the process of running the business system according to the detected operation button, an operation exception occurs, etc.

[0035] The multiple attribute information of the abnormal situation can be related information reflecting multiple dimensions of the corresponding abnormal situation; for example, error code related to the name of the abnormal situation, descriptive information related to the specific content of the abnormal situation, scene information related to the abnormal situation, interface information related to the abnormal situation, operation action information related to the abnormal situation, and timestamp information related to the abnormal situation, etc.

[0036] Based on this, when a current abnormal situation occurs in a currently running business system or task process, multiple attribute information related to that abnormal situation can be detected.

[0037] On the one hand, because the detected multiple attribute information is represented in different structural forms, for example, error codes can be represented in the structural form of programming languages, description information can be represented in the structural form of programming languages ​​or professional natural languages, scene information, interface information and operation action information can be represented in the structural form of predefined data structures, professional natural languages ​​or predefined programming languages, and timestamp information can be represented in the structural form of predefined data structures, users usually find it difficult to know the true meaning of each attribute information.

[0038] On the other hand, due to the different representations of various attribute information, if these attribute information are used as input data for artificial intelligence models such as machine learning models, and the machine learning models generate natural language corresponding to these attribute information to describe the current abnormal situation, it is difficult to generate accurate natural language.

[0039] Furthermore, since the attribute information is scattered, when describing the current abnormal situation using natural language through machine learning models, it is not only easy to miss important attribute information, but also difficult to effectively trace the attribute information that caused the inaccurate natural language when the generated natural language is inaccurate.

[0040] Based on this, in this step, the detected attribute information can be encapsulated into a unified and standardized structural form. Specifically, each detected attribute information can be converted into the same data structure to obtain the corresponding attribute data, and each attribute data can be encapsulated into a single data structure to avoid missing important attribute information. Furthermore, a data structure that can be applied to machine learning models can be obtained for each attribute data.

[0041] The data structure can be, for example, a JSON (JavaScript Object Notation) structure.

[0042] In some specific examples, the data structure corresponding to the current error or exception can be represented as a JSON structure as shown below: JSON{ "errorCode": "SERVER_INTERNAL_ERROR", "errorMessage": "InternalServer Error", "scene": "FORM_SUBMIT", "pageId": "order_create", "userAction": "click_submit", "timestamp": 1700000000} Each row represents one attribute of the anomaly.

[0043] Here, errorCode represents the error code for this exception in the JSON structure format. Specifically, the error code is SERVER_INTERNAL_ERROR (internal system error).

[0044] errorMessage represents the attribute data, or information data, in JSON structure form that describes the specific content of the exception, namely Internal Server Error.

[0045] "scene" represents the attribute data related to the scenario of the exception in the form of a JSON structure, i.e., the scenario data. Specifically, the scenario is FORM_SUBMIT (submit scenario).

[0046] pageId represents the attribute data, or interface data, related to the page in the abnormal situation under the JSON structure. Specifically, this interface information is order_create (order creation). The interface data can be related data when the interface is a page or other interfaces.

[0047] userAction represents the attribute data, or operation data, related to the user's action in this exception situation, in the form of a JSON structure. The action is click_submit (click to submit).

[0048] The timestamp represents the timestamp data, which is the attribute data related to the timestamp of the exception in the JSON structure. The timestamp is 1700000000.

[0049] Step 202: Generate natural sentences using at least one target attribute data from each attribute data of the data structure.

[0050] Among them, a natural statement is a statement expressed in natural language. This natural statement can describe the current error or exception through one or more attribute data. Accordingly, the attribute data involved in the natural statement are the target attribute data among the attribute data related to the current error or exception. That is, the target attribute data are the attribute data required to generate the natural statement.

[0051] In this step, based on the data structure obtained in step 201 above, a machine learning model can be used to obtain the target attribute data required to generate natural sentences from the data structure, and the corresponding natural sentences can be generated using the target attribute data.

[0052] The machine learning model can be, for example, a large language model. This large language model can be a finely tuned large language model for the required scenario. Specifically, during the fine-tuning process, the large language model can be trained using JSON structures under various error or abnormal conditions as input data, thereby adjusting the parameters of the large language model and enabling the trained large language model to output more accurate results.

[0053] In the process of utilizing a large language model, prompt words can be identified first for generating natural sentences using that large language model.

[0054] The prompt includes several instructions that constrain the generation actions of the large language model.

[0055] Based on this, the large language model can generate target attribute data that conforms to each instruction in the prompt word from the data structure, and use the target attribute data to generate natural sentences according to each instruction in the prompt word.

[0056] Step 203: Determine whether the natural language representation conforms to the preset display standards.

[0057] The display criteria include multiple dimensions of metrics used to verify whether the natural language can accurately and understandably convey information related to the current error or anomaly to the user.

[0058] In this step, based on the natural language statement generated in step 202 above, there are still situations where users have difficulty understanding the natural language statement and / or do not know how to handle it. Therefore, before displaying the natural language statement, the various contents described in the natural language statement can be verified from multiple dimensions through the display standard to see if they can accurately and understandably convey information related to the current error or anomaly.

[0059] Specifically, since the display standard can include indicators of multiple dimensions, the indicator values ​​of each dimension can be determined separately, and the indicator values ​​of each dimension can be weighted to obtain a weighted result for the natural language statement.

[0060] Furthermore, based on a pre-set evaluation threshold, if the weighted result is greater than or equal to the evaluation threshold, it can be determined that the natural statement meets the display criteria.

[0061] Step 204: In response to determining that the display criteria are met, display natural language statements.

[0062] In this step, based on the judgment result in step 203 above, if the natural statement meets the display standard, it can be considered that the natural statement can accurately and easily describe to the user the information related to the current error or anomaly.

[0063] Based on this, a complete process of encapsulating attribute information, generating natural language statements, and validating display standards has been implemented to ensure that exception information is presented in an easy-to-understand and accurate manner. Specifically, by encapsulating multiple attribute information of an exception into a unified data structure, the identification bias of different forms of attribute information is avoided, ensuring that the complete context is fully utilized when generating natural language statements, thus improving the accuracy of the statement description; by generating natural language statements based on standardized attribute data, the use of professionally encoded or garbled characters for exception prompts is avoided, making it easier for users to understand; by validating natural language statements through preset display standards, prompts that meet the requirements are selected, further ensuring the effectiveness of information delivery; and by displaying compliant natural language statements, the problems of traditional exception prompts being obscure and difficult to understand, and users not being able to clearly understand the cause of the error and the direction of handling are solved.

[0064] In some alternative implementations, refer to Figure 2BThis illustrates a decomposition process 2020 of one embodiment of step 202 of this disclosure. The decomposition process 2020 includes steps 2021 to 2023: Step 2021: Identify prompt words that include multiple instructions.

[0065] Among the multiple indications of the prompt word, there is at least one content indication pointing to the target attribute data and at least one description indication constraining the format and semantic content of the statement.

[0066] Specifically, the content indicator can be used to constrain the appearance or absence of the target attribute data pointed to by the content indicator in the natural language statement. For example, constraining the absence of error codes in the generated natural language statement.

[0067] Furthermore, the description instructions may include formatting instructions for constraining word count and / or sentence format, for example, constraining the number of words in the generated natural sentences to not exceed a predetermined number.

[0068] Description instructions may also include language instructions for constraining semantic content, such as constraining the generated natural language to not contain technical terms, and / or using a specified language (e.g., using Chinese), etc.

[0069] In addition, since natural language statements can not only describe specific information about the current error or exception to the user, but also convey relevant operations for handling the error or exception, natural language statements can include prompt statements describing the current error or exception, as well as operation statements that guide the user to take action.

[0070] Accordingly, multiple instructions may also include at least one guiding instruction to define the action. For example, the generated natural language statement may be required to include one or more actions that the user can perform to handle the current error or exception.

[0071] Based on this, in the process of generating natural sentences using a large language model, we can first determine the various instructions used to constrain the generation operation and generate prompt words containing these instructions.

[0072] In a specific example, where the identified instructions include the aforementioned content instructions, formatting instructions, language instructions, and guidance instructions, the prompt words composed of these instructions could be, for example: "In response to the current exception, the generated natural language statement should: not contain technical terms, not contain error codes, not exceed 20 Chinese characters, describe the scenario in which the exception occurred and the corresponding page, and include a user-executable action."

[0073] Step 2022: Determine the target attribute data corresponding to each content instruction from the data structure according to the content instructions.

[0074] In this step, based on the prompt words determined in step 2021 above, it can be determined that the content of the prompt words indicates an error code, scene data related to the scene, interface data related to the page, and operation data related to the operation action.

[0075] Based on this, the aforementioned error codes, scene data, interface data, and operation data can be used as target attribute data. Scene data, interface data, and operation data can be selected from the various attribute data of the data structure, while error codes are not selected.

[0076] Step 2023: Generate prompt statements pointing to the target attribute data of each item according to the description instructions.

[0077] Based on the prompt words determined in step 2021 above, and the target attribute data determined in step 2022 above, prompt statements describing the target data can be generated according to the format and language instructions contained in the prompt words.

[0078] Specifically, error codes are typically used to identify specific types of errors, such as "SERVER_INTERNAL_ERROR", "PARAM_INVALID", or "AUTH_FAILED". These identifiers are usually not represented in natural language, making it difficult for large language models to effectively utilize the content represented by the error code. Therefore, error codes in data structures can be converted into semantic forms, enabling large language models to effectively obtain the semantic information contained in the error code.

[0079] Specifically, all possible error codes for various types of errors and anomalies can be predetermined, and a corresponding semantic label can be set for each error code, thereby forming a corresponding mapping relationship between the error code and the semantic label.

[0080] Semantic labels can be labels that identify corresponding errors or anomalies in natural language form. Due to their natural language form, the semantic information contained in the semantic label can be recognized by large language models.

[0081] In some cases, each semantic tag may correspond to only a single error code; in other cases, each semantic tag may correspond to one or more error codes.

[0082] Based on this, during the generation of prompt statements, the semantic tags corresponding to the error codes can be mapped out using the mapping relationship of the error codes in the data structure.

[0083] In specific examples, the semantic label corresponding to the error code "SERVER_INTERNAL_ERROR" could be, for example, "SYSTEM_EXCEPTION", the semantic label corresponding to the error code "PARAM_INVALID" could be, for example, "USER_INPUT_ERROR", and the semantic label corresponding to the error code "AUTH_FAILED" could be, for example, "PERMISSION_ERROR".

[0084] Furthermore, based on the target attribute data determined in step 2022, the semantic tag and the target attribute data can be input into the large language model, and the large language model can be used to generate corresponding prompt statements according to the prompt words determined in step 2021.

[0085] In a specific example of the JSON structure in step 201, the target attribute data determined in step 2022 includes: error code "SERVER_INTERNAL_ERROR", scenario data "FORM_SUBMIT", interface data "order_create", and operation data "click_submit"; the semantic tag is "SYSTEM_EXCEPTION"; the prompt obtained in step 2021, "For the current exception, the generated natural language statement should: not contain technical terms, not contain error codes, not exceed 20 Chinese characters, describe the scenario in which the exception occurred, and the corresponding page, and include a user-executable action", the description instruction includes "not exceed 20 Chinese characters", the language instruction includes "not contain technical terms", and the content instruction includes "not contain error codes" and "the scenario in which the exception occurred, and the corresponding page".

[0086] Based on this, the error code "SERVER_INTERNAL_ERROR", the scene data "FORM_SUBMIT", the interface data "order_create", the operation data "click_submit", the semantic label "SYSTEM_EXCEPTION", and the prompt words can be input into the large language model.

[0087] Accordingly, the large language model can describe the scene data "FORM_SUBMIT", interface data "order_create", and operation data "click_submit" that point to the content instruction "does not contain technical terms" and "the scenario in which the exception occurs and the corresponding page" with the description instruction "no more than 20 Chinese characters" according to the requirement of the language instruction "does not contain technical terms", thus forming a prompt statement that does not contain the error code "SERVER_INTERNAL_ERROR".

[0088] Furthermore, based on the guidance instruction in the prompt word obtained in step 2021, which "includes a user-executable operation action", after the prompt word is input into the large language model, the large language model can generate an operation statement to handle the current error or exception according to the guidance instruction.

[0089] Specifically, multiple user actions can be pre-set, such as viewing a form sample, re-filling out the form, contacting technical staff, and submitting feedback.

[0090] Accordingly, the large language model can determine the target operation action from multiple preset operation actions based on the target attribute data of the current error or anomaly.

[0091] During this process, severity levels can be pre-set for various errors and exceptions, and corresponding action plans can be set for each severity level.

[0092] In a specific example, if the error code "SERVER_INTERNAL_ERROR" appears for the first time in the target attribute data, it can be classified as a low severity level. If the same error code appears three or more times, it can be classified as a high severity level.

[0093] The low severity level can correspond to actions such as viewing a sample form or filling out a new form, while the high severity level can correspond to actions such as contacting technical personnel or submitting feedback.

[0094] Based on this, the large language model can determine the number of times an error or anomaly occurs and its corresponding severity level based on the target attribute data, and determine the corresponding operation action based on the severity level, thereby generating the operation statement corresponding to the operation action.

[0095] In specific examples, you can set the corresponding operation statement "View Filling Example" for the operation of viewing the form, "Refill Form" for the operation of refilling the form, "Contact Technical Personnel" for the operation of contacting technical personnel, and "Submit Feedback" for the operation of submitting feedback.

[0096] Therefore, when the error code "SERVER_INTERNAL_ERROR" appears for the first time, the large language model can generate the corresponding operation statement "View the filling example" or operation statement "Fill in the form again".

[0097] In some cases, one or more effective time periods for an operation can be preset. For example, the effective time period for contacting technical personnel can be set to 8:00 to 17:00, and the rest of the time period can be considered ineffective.

[0098] Based on this, if the corresponding action is determined to be contacting technical personnel, and the current time period is not a valid time period, the large language model can select other actions as the target action.

[0099] Based on this, a refined process of identifying prompt words, extracting target attribute data, and generating prompt statements and / or guiding statements achieves precision and adaptability in anomaly prompts. By defining prompt words containing multiple indicators, constraints are clearly defined from the dimensions of content, format, and semantics, providing clear guidance for subsequent statement generation. Target attribute data is precisely filtered according to content indicators, avoiding redundant or irrelevant data interference. Prompt statements are generated based on descriptive instructions, ensuring that anomaly information is presented in a form that meets the constraints. It is evident that through multi-level indicator constraints and data filtering, the generated prompt statements are ensured to not only align with the core attributes of the anomaly but also meet preset format and semantic specifications. This effectively solves the problems of traditional anomaly prompts lacking specificity and having non-standard expressions, allowing users to quickly and accurately understand the anomaly situation while improving the standardization and controllability of anomaly prompt generation.

[0100] In some alternative implementations, refer to Figure 2C The diagram illustrates a flow 2030 of one embodiment of step 203 of this disclosure. Flow 2030 includes steps 2031 to 2033: Step 2031: Determine the semantic readability of the natural language statement and use it as the first semantic indicator; determine the emotional positivity of the natural language statement and use it as the second semantic indicator; determine the guiding effect of the natural language statement and use it as the third semantic indicator; determine the semantic consistency between the natural language statement and the target attribute data and use it as the fourth semantic indicator.

[0101] Based on the natural language statement determined in step 202 above, before displaying the natural language statement to the user, it can be determined whether the natural language statement can accurately and easily convey information related to the current error or exception to the user.

[0102] Specifically, multiple semantic indicators can be pre-set, such as a first semantic indicator to characterize the semantic readability of natural sentences, a second semantic indicator to characterize the emotional positivity of natural sentences, a third semantic indicator to characterize the guiding nature of natural sentences, and a fourth semantic indicator to characterize the semantic consistency between natural sentences and various target attribute data.

[0103] Among them, the large language model can determine the first semantic indicator by such as the text length, word length, and number of sentences of the natural sentence; the large language model can determine the second semantic indicator by identifying the negative and positive emotions of the natural sentence; the large language model can determine the third semantic indicator by whether the natural sentence contains operation statements; and the large language model can determine the fourth semantic indicator by the semantic similarity between the natural sentence and the semantic label.

[0104] Step 2032: Determine the weighted result of the first semantic indicator, the second semantic indicator, the third semantic indicator, and the fourth semantic indicator.

[0105] In this step, based on the first semantic index, the second semantic index, the third semantic index and the fourth semantic index determined in step 2031 above, each semantic index can be weighted.

[0106] Specifically, the first semantic indicator, the second semantic indicator, the third semantic indicator, and the fourth semantic indicator can be assigned the same weight, or they can be assigned different weights.

[0107] Step 2033: In response to determining that the weighted result is greater than or equal to a preset evaluation threshold, determine that the natural language representation meets the display standard.

[0108] In this step, an evaluation threshold can be preset. Based on the weighted result determined in step 2033 above, it can be determined whether the weighted result is greater than or equal to the evaluation threshold.

[0109] If the weighted result is greater than or equal to the evaluation threshold, it can be considered that the current natural language statement can accurately and easily convey to the user the specific information about the current error or anomaly, as well as the corresponding processing actions that the user can take, and it can be determined that the natural language statement meets the display standards.

[0110] If the weighted result is less than the evaluation threshold, it can be considered that the current natural language statement is unable to accurately and understandably convey to the user the specific information about the current error or anomaly, or is unable to clearly convey to the user the corresponding processing actions, and it can be determined that the natural language statement does not meet the display standards.

[0111] Based on this, semantic indicators are determined from four dimensions: semantic readability, emotional positivity, guidance, and semantic consistency. This comprehensively covers the core elements of natural language sentences conveying information to users, avoiding the one-sidedness of single-dimensional verification. By weighting the four semantic indicators, the importance of each dimension is comprehensively considered to form an objective and comprehensive evaluation result. By comparing with the preset evaluation threshold, it is determined whether the natural language sentences meet the display standards, ensuring that only sentences that meet the requirements are displayed to users.

[0112] In some alternative implementations, after step 204 of this disclosure, if the operation is to contact a technical personnel, relevant information may be sent to the target address in the corresponding location.

[0113] In this embodiment, the data structure may also include the business topic corresponding to the current error or exception. Specifically, the business topic may be the business system or task process that has the current error or exception, and the contact address corresponding to each business system and task process may be preset.

[0114] Based on this, after the user performs the corresponding operation according to the displayed operation statement, if the operation is to contact technical personnel, the large language model can determine the corresponding business system or task process and the corresponding contact address of the business system or task from the data structure, and use it as the target address.

[0115] Furthermore, large language models can utilize data structures to generate structured summaries; specifically, they can combine semantic tags and target attribute data to form structured summaries.

[0116] Accordingly, the large language model can send the structured summary to the corresponding target address, so that when a user performs the action of contacting technical personnel, the relevant information data can be sent to the corresponding target address with a single click of the corresponding button.

[0117] Based on this, through the linkage design of business data and contact operations, efficient closed-loop processing of abnormal issues is achieved. After displaying natural statements, if the user triggers a contact operation, the corresponding preset target address can be accurately located based on the business data in the data structure, and the complete abnormal data structure or its structured summary can be sent directly without the user needing to supplement any additional information. This provides the recipient with a complete context to quickly locate the root cause of the problem.

[0118] Further reference Figure 3 As an implementation of the methods shown in the above figures, this disclosure provides an embodiment of an abnormal information processing device, which corresponds to the method embodiment shown in Figure 2, and the device can be specifically applied to various electronic devices.

[0119] like Figure 3 As shown, the abnormal information processing device 300 in this embodiment includes: an encapsulation module 301, a generation module 302, a judgment module 303, and a display module 304; The encapsulation module 301 is configured to encapsulate multiple attribute information into multiple attribute data in a data structure in response to the detection of multiple attribute information of the current abnormal situation. The generation module 302 is configured to generate natural statements using at least one target attribute data from each attribute data of the data structure, and the natural statements are used to describe each target attribute data. The judgment module 303 is configured to determine whether the natural language representation conforms to the preset display standard; Display module 304 is configured to display natural language in response to determining that it conforms to the display criteria.

[0120] In this embodiment, the specific processing of the encapsulation module 301, generation module 302, judgment module 303, and display module 304 of the abnormal information processing device 300, and the resulting technical effects, can be referred to respectively. Figure 2A The relevant descriptions of steps 201, 202, 203 and 204 in the corresponding embodiments will not be repeated here.

[0121] In some alternative implementations, the generation module 302 may further be configured as follows: Identify prompt words that include multiple indicators, among which at least one is a content indicator pointing to the target attribute data and at least one is a descriptive indicator constraining the format and semantic content of the statement; Determine the target attribute data corresponding to each content instruction from the data structure according to each content instruction; Generate prompt statements that point to the target attribute data according to the description instructions.

[0122] The target attribute data includes the error code, scenario data, interface data, and operation data corresponding to the current abnormal situation; the description instructions include format instructions for constraining word count and / or sentence format, and language instructions for constraining semantic content. Accordingly, prompt statements are generated according to the description instructions, pointing to the target attribute data for each item, including: Map error codes to corresponding preset semantic tags; The system uses a pre-defined large language model to generate prompts that describe semantic tags, scene data, interface data, and operation data, and that conform to description and language instructions.

[0123] In some alternative implementations, the natural statement may also include an operation statement; the multiple instructions may also include at least one guiding instruction for defining the operation action; Accordingly, the generation module 302 can be further configured to, after generating prompt statements pointing to target attribute data according to the description instructions, determine the target operation action corresponding to the current abnormal situation from a set of preset operation actions based on the target attribute data of the current abnormal situation. In some alternative implementations, the determination module 303 may be further configured as follows: Determine the semantic readability of natural language sentences and use it as the primary semantic indicator; Determine the emotional positivity of natural language statements and use it as a secondary semantic indicator; Determine the degree of introductory effect of natural language statements and use it as a third semantic indicator; Determine the semantic consistency between natural language statements and the target attribute data, and use it as the fourth semantic indicator; Determine the weighted result of the first semantic index, the second semantic index, the third semantic index, and the fourth semantic index; In response to determining that the weighted result is greater than or equal to a preset evaluation threshold, it is determined that the natural language representation meets the display criteria.

[0124] Accordingly, the judgment module 303 can be further configured to, after determining the weighted result of the first semantic indicator, the second semantic indicator, the third semantic indicator and the fourth semantic indicator, in response to determining that the weighted result is less than the evaluation threshold, select the candidate statement corresponding to the current abnormal situation from the multiple candidate statements according to the preset mapping relationship between multiple preset candidate statements and various abnormal situations, and determine it as a natural statement.

[0125] In some alternative implementations, the exception information processing device may further include an execution module 305, configured to, after displaying the natural statement, in response to determining that a trigger contact operation has been initiated, send the data structure corresponding to the current exception to a preset target address corresponding to the business data.

[0126] It should be noted that the implementation details and technical effects of each module in the abnormal information processing device provided in the embodiments of this disclosure can be referred to the descriptions of other embodiments in this disclosure, and will not be repeated here.

[0127] The following is for reference. Figure 4 It shows a schematic diagram of the structure of a computer system 400 suitable for implementing the electronic device of the present disclosure. Figure 4 The computer system 400 shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0128] like Figure 4 As shown, the computer system 400 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage device 408 into a random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of the computer system 400. The processing device 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0129] Typically, the following devices can be connected to I / O interface 405: input devices 406 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, etc.; output devices 407 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 408 including, for example, magnetic tapes, hard disks, etc.; and communication devices 409. Communication device 409 allows computer system 400 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 4 A computer system 400 with various electronic 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.

[0130] 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 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 409, or installed from storage device 408, or installed from ROM 402. When the computer program is executed by processing device 401, it performs the functions defined in the methods of embodiments of this disclosure.

[0131] 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 conjunction 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.

[0132] 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.

[0133] The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the following functions: Figure 2A The embodiments shown and their alternative implementations illustrate methods for handling abnormal information.

[0134] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and 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).

[0135] 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.

[0136] The units or modules described in the embodiments of this disclosure can be implemented in software or hardware. The names of the units or modules do not necessarily limit the unit itself; for example, an encapsulation module can also be described as "a module that encapsulates multiple attribute information into multiple attribute data in a data structure in response to the detection of multiple attribute information of a current abnormal situation."

[0137] 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.

Claims

1. A method for processing abnormal information, comprising: In response to the detection of multiple attribute information of the current abnormal situation, the multiple attribute information is encapsulated into multiple attribute data in the data structure; A natural language statement is generated using at least one target attribute data from each attribute data of the data structure, and the natural language statement is used to describe each target attribute data. Determine whether the natural language representation conforms to the preset display standards; In response to determining that the display criteria are met, the natural language statement is displayed.

2. The method according to claim 1, wherein, The natural language statements include prompt statements; and The step of generating natural language statements using at least one target attribute from each attribute data item of the data structure includes: Identify prompt words that include multiple indicators, among which at least one is a content indicator pointing to the target attribute data and at least one is a descriptive indicator constraining the format and semantic content of the statement; According to the content instructions, determine the target attribute data corresponding to each content instruction from the data structure; Generate prompt statements that point to the target attribute data of each item according to the description instructions.

3. The method according to claim 2, wherein, The target attribute data includes the error code, scene data, interface data, and operation data corresponding to the current abnormal situation. The description indication includes format indication for constraining word count and / or sentence format, and language indication for constraining semantic content. as well as The step of generating the prompt statement pointing to the target attribute data according to each description instruction includes: The error codes are mapped to corresponding preset semantic tags; The system uses a pre-defined large language model to generate prompt statements that describe the semantic tags, the scene data, the interface data, and the operation data, and that conform to the description instructions and the language instructions.

4. The method according to claim 2, wherein, The natural statements also include operation statements, and the multiple instructions also include at least one guiding instruction for defining the operation action; as well as After generating the prompt statement pointing to the target attribute data according to the description instructions, the method further includes: Based on the target attribute data of the current abnormal situation, determine the target operation action corresponding to the current abnormal situation from a number of preset operation actions; The operation statements are generated according to the various guiding instructions, which are used to perform the operation actions for each target.

5. The method according to claim 1, wherein, The determination of whether the natural language representation conforms to the preset display standard includes: The semantic readability of the natural language is determined and used as a first semantic indicator; The emotional positivity of the natural language statement is determined and used as a second semantic indicator; The degree of guidance of the natural statement is determined and used as a third semantic indicator; Determine the semantic consistency between the natural language statement and the target attribute data, and use it as the fourth semantic indicator; Determine the weighted result of the first semantic indicator, the second semantic indicator, the third semantic indicator, and the fourth semantic indicator; In response to determining that the weighted result is greater than or equal to a preset evaluation threshold, the natural language representation is determined to conform to the display standard.

6. The method according to claim 5, wherein, After determining the weighted result of the first semantic indicator, the second semantic indicator, the third semantic indicator, and the fourth semantic indicator, the method further includes: In response to determining that the weighted result is less than the evaluation threshold, according to a preset mapping relationship between multiple preset candidate statements and various abnormal situations, a candidate statement corresponding to the current abnormal situation is selected from the multiple candidate statements and determined as the natural statement.

7. An apparatus for processing abnormal information, comprising: The encapsulation module is configured to encapsulate multiple attribute information into multiple attribute data in a data structure in response to the detection of current abnormal conditions. The generation module is configured to generate a natural statement using at least one target attribute data from each attribute data of the data structure, the natural statement being used to describe each target attribute data. The judgment module is configured to determine whether the natural language representation conforms to a preset display standard; The display module is configured to display the natural language statement in response to determining that it conforms to the display criteria.

8. An electronic device, comprising: One or more processors; Storage device, on which one or more programs are stored, When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, wherein, When the computer program is executed by one or more processors, it implements the method as described in any one of claims 1-6.

10. A computer program product comprising computer program instructions, characterized in that, When the computer program instructions are executed on the computer, the computer causes the computer to perform the method as described in any one of claims 1-6.

Images