Identification letter auditing method and device, equipment, storage medium and program product
By recognizing images of disability assessment reports and generating prompts, and utilizing pre-trained language models and knowledge graphs to automate the review of disability assessment reports, the problem of low efficiency in traditional manual review is solved, and efficient and accurate review results are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-06
- Publication Date
- 2026-06-12
AI Technical Summary
Traditional manual review of disability assessment reports is inefficient, prone to omissions or misjudgments, and affects the accuracy of the review results.
By recognizing the image of the disability assessment report, generating prompts and inputting them into a pre-trained language model, the disability level and entry number are determined. Combined with the review standards of a pre-set knowledge graph, the review result is automatically determined.
It improved the efficiency of searching for disability levels and entry numbers, reduced the tedious process of manual review, avoided the decrease in review accuracy caused by large language models, and ensured the accuracy of review results.
Smart Images
Figure CN122198057A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of data processing technology, and in particular to a method, apparatus, equipment, storage medium, and program product for reviewing certificates. Background Technology
[0002] In areas such as insurance claims and medical dispute mediation, the review of disability assessment reports is a crucial step in ensuring the legality, impartiality, and scientific validity of the assessment conclusions.
[0003] Currently, under the traditional manual review model, appraisers need to compare the injury descriptions in the appraisal documents with the standard clauses one by one, and verify whether each item meets the judgment criteria.
[0004] However, the inventors have found that the relevant technology has at least the following technical problems: manual verification is inefficient and prone to omissions or misjudgments. Summary of the Invention
[0005] This application provides a method, apparatus, equipment, storage medium, and program product for reviewing certificates, in order to solve the problems of low efficiency and easy omissions or misjudgments in manual verification.
[0006] In a first aspect, embodiments of this application provide a method for reviewing disability assessment certificates, comprising: acquiring an image of a disability assessment certificate; determining the disability assessment text based on the image of the disability assessment certificate; generating prompt words based on the disability assessment text; inputting the prompt words into a first pre-trained language model to obtain the disability level and corresponding entry number output by the first pre-trained language model; determining the review standard based on the disability level and entry number; and determining the review result based on the review standard and the disability assessment text.
[0007] In one possible implementation, the review criteria include judgment conditions and judgment logic; the review result is determined based on the review criteria and the disability assessment text, including: determining the logical input based on the judgment conditions and the disability assessment text; and determining the review result based on the logical input and the judgment logic.
[0008] In one possible implementation, the logical input is determined based on the judgment condition and the disability assessment text, including: if the judgment condition is a semantic condition, generating a judgment prompt word based on the judgment condition and the disability assessment text; inputting the judgment prompt word into a second pre-trained language model to obtain the logical input output by the second pre-trained language model; if the judgment condition is a computational condition, reading the input variable corresponding to the judgment condition in the disability assessment text; and inputting the input variable into the judgment condition to obtain the logical input.
[0009] In one possible implementation, the audit result is determined based on the logical input and the judgment logic, including: substituting the logical input into the judgment logic to obtain the logic to be executed; and executing the logic to be executed to obtain the audit result.
[0010] In one possible implementation, the review criteria are determined based on the disability level and the entry number, including: reading the sub-entry corresponding to the disability level in the preset knowledge graph; identifying the sub-entry corresponding to the entry number in the sub-entry as the target sub-entry; and reading the judgment conditions and judgment logic under the target sub-entry to determine it as the review criteria.
[0011] In one possible implementation, before reading the sub-entries corresponding to the disability level in the preset knowledge graph, the method further includes: generating graph construction prompts based on the preset disability level classification standard text; inputting the graph construction prompts into a third pre-trained language model to obtain the preset knowledge graph output by the third pre-trained language model.
[0012] Secondly, embodiments of this application provide an appraisal certificate review device, comprising: an image acquisition module for acquiring an image of a disability appraisal certificate; a text determination module for determining disability appraisal text based on the image of the disability appraisal certificate; a prompt word generation module for generating prompt words based on the disability appraisal text; an information acquisition module for inputting the prompt words into a first pre-trained language model to obtain the disability level and corresponding entry number output by the first pre-trained language model; a standard determination module for determining review standards based on the disability level and entry number; and a result determination module for determining the review result based on the review standards and the disability appraisal text.
[0013] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;
[0014] The memory stores the instructions that the computer executes;
[0015] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.
[0016] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.
[0017] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.
[0018] The assessment certificate review method, apparatus, equipment, storage medium, and program product provided in this application acquire an image of a disability assessment certificate and determine the corresponding text. Based on the disability assessment text, prompt words are generated and input into a first pre-trained language model to obtain the disability level and corresponding entry number output by the first pre-trained language model. Based on the disability level and entry number, review standards are determined. Then, combining the review standards and the disability assessment text, the review result is determined. This achieves automated conversion of image data into text and uses a pre-trained language model to find the corresponding disability level and entry number, increasing the efficiency of finding the disability level and entry number. In subsequent review processes, the corresponding review result is determined based on the review standards and the disability assessment text, avoiding the problem of decreased review accuracy due to the illusion of a large model when directly using a large language model for review. Attached Figure Description
[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.
[0020] Figure 1 A schematic diagram illustrating the scenario of the authentication certificate review method provided for this application;
[0021] Figure 2 A flowchart illustrating the certificate review method provided in this application embodiment;
[0022] Figure 3 A schematic diagram of the structure of the certificate review device provided in the embodiments of this application;
[0023] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.
[0024] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation
[0025] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.
[0026] In core areas such as insurance claims, medical dispute mediation, and judicial appraisal, the review of disability assessment reports is a crucial support for ensuring the legality, impartiality, and scientific validity of disability assessment conclusions, and a core link in safeguarding the legitimate rights and interests of relevant parties.
[0027] Currently, disability assessment review is still mainly conducted using the traditional manual method. Assessors need to manually compare the description of the injury and the examination results in the assessment documents with the standard clauses to verify whether the relevant content meets the conditions for determining the disability level.
[0028] However, through practical research, the inventors discovered that the existing traditional manual review model has obvious technical shortcomings. The core problems are mainly reflected in two aspects: First, the manual verification process is cumbersome and inefficient, making it difficult to adapt to the review needs of a large number of appraisal documents; second, manual comparison is easily affected by subjective experience, workload and other factors, which can easily lead to omissions of clauses and judgment biases, thereby affecting the accuracy of the review results.
[0029] To address the aforementioned technical problems, the inventors provide the following technical concept: by recognizing the image of the disability assessment certificate, the disability assessment text is obtained; prompt words are generated based on the disability assessment text; the prompt words are input into a large model to obtain the corresponding disability level and item number; the corresponding review standard is determined based on the disability level and item number; and the review result is determined using the review standard and the review criteria.
[0030] This application is used in the scenario of reviewing appraisal reports. It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with relevant laws, regulations and standards, and corresponding operation entry points are provided for users to choose to authorize or refuse.
[0031] Figure 1 This is a schematic diagram illustrating the scenario for reviewing the authentication certificate provided in this application. For example... Figure 1 In this scenario, the following components are included: terminal device 101 and server 102.
[0032] In the specific implementation process, the terminal device 101 may include a computer, server, tablet, mobile phone, PDA (Personal Digital Assistant), and laptop, etc., which can input data.
[0033] Server 102 can be implemented using a single server or a cluster of multiple servers with more powerful processing capabilities and higher security. Where possible, it can also be replaced by a computer or laptop with strong computing power.
[0034] The connection between server 102 and terminal device 101 can be either wired or wireless.
[0035] Server 102 is used to receive the disability assessment certificate image sent by the terminal device, determine the disability assessment text based on the disability assessment certificate image, generate corresponding prompt words, input the prompt words into a pre-trained language model to obtain the disability level and corresponding item number, and then determine the review standard based on the disability level and item number, and determine the review result based on the review standard and the disability assessment text.
[0036] It is understood that the scenarios illustrated in the embodiments of this application do not constitute a specific limitation on the method for reviewing appraisal certificates. In other feasible embodiments of this application, the above scenarios may include more or fewer components than illustrated, or combine some components, or split some components, or arrange different components, which can be determined according to the actual application scenario and are not limited here. Figure 1 The scenario shown can be implemented by hardware, software, or a combination of both.
[0037] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.
[0038] Figure 2 This is a flowchart illustrating the certificate review method provided in this application embodiment. The execution entity of this application embodiment can be... Figure 1 The server 102 in this embodiment can also be a computer and / or a mobile phone, etc., and this embodiment does not impose any special restrictions on it. Figure 2 As shown, the method includes:
[0039] S201: Obtain the image of the disability assessment report.
[0040] In this step, the disability assessment certificate image can be obtained through the system's preset input interface, or it can be manually uploaded by the user on the user terminal (such as uploading from a web page or client), or imported through external system interfaces (such as push from the assessment agency system or insurance system). The disability assessment certificate image can be in any image format (JPG, PNG, PDF scan, etc.). After acquisition, the image can be pre-processed (such as noise reduction and bias correction) to ensure the accuracy of subsequent analysis. No manual intervention is required in the image acquisition process; only the integrity of the image needs to be verified (to avoid missing pages).
[0041] For example, insurance claims personnel can use the "image upload" function of the claims review system to upload scanned copies of the disability assessment report submitted by the policyholder (which may include physical examination records, injury descriptions, preliminary assessment opinions, etc.) to the server. The server then recognizes the complete image of the disability assessment report for subsequent text parsing.
[0042] S202: Determine the text of the disability assessment based on the image of the disability assessment report.
[0043] In this step, an optical character recognition algorithm can be used to recognize each page of the acquired disability assessment report image, restoring the text, numbers, symbols, and other information in the image into an editable and parsable plain text format. For blurry, tilted, or occluded areas in the image, OCR enhancement algorithms (such as edge detection and character completion) can be used to optimize the recognition effect. After recognition, the text of each page is stitched together to form a complete disability assessment text. Optionally, blurry paragraphs can also be marked for subsequent manual review.
[0044] For example, the system identifies each page of the disability assessment report obtained in S201. The "Physical Examination Record" on page 3 is slightly blurry. The OCR enhancement algorithm automatically completes the blurry characters and finally generates the complete text, which includes key information such as "left upper limb pronation degree 20°, supination degree 10°, fracture of one vertebra and corresponding accessory structures, surgical treatment has been performed".
[0045] S203: Generate prompt words based on the disability assessment text.
[0046] In this step, based on the preset prompt word template and combined with the content of the disability assessment text, key segments (such as the assessment opinion section and the injury diagnosis section) are extracted from the text and embedded into the prompt word template. The prompt word template can be preset with clear task instructions, such as requiring the model to extract the "disability level" and "corresponding item number", so that the system can automatically fill the template to generate prompt words.
[0047] S204: Input the prompt word into the first pre-trained language model to obtain the disability level and corresponding entry number output by the first pre-trained language model.
[0048] In this step, the first pre-trained language model can be a large language model fine-tuned with business data in the field of disability assessment (such as historical assessment reports and national standard clauses), which is adapted to the task of extracting disability level and entry number. The prompt words generated by S203 are input into the model according to the preset model input format (such as text strings). The model recognizes the specific numerical value of the disability level (such as level 9 and level 10) and the corresponding national standard entry number (such as 5.9.6.2) in the text through semantic understanding, filters irrelevant information (such as redundant injury descriptions), and outputs structured results (including disability level and entry number).
[0049] The prompt words generated by S203 are input into the fine-tuned first pre-trained language model. The model identifies the clear disability level in the text as level 9 through semantic parsing. Combined with the injury description (fracture of a vertebra and corresponding appendages, surgical treatment has been performed), it matches the corresponding national standard entry number 5.9.6.2. The final output result is: "Level: 9, Entry Number: 5.9.6.2".
[0050] S205: Determine the review criteria based on the disability level and item number.
[0051] In this step, the source of the review criteria is a pre-set knowledge graph. The system first uses the disability level as the primary search condition to locate all sub-items corresponding to that level in the knowledge graph; then, it uses the item number as the secondary search condition to filter out the uniquely matching target sub-item from the sub-item set; finally, it extracts the pre-set judgment conditions (semantic conditions or calculation conditions) and judgment logic under the target sub-item, and integrates them as the review criteria for this review.
[0052] The knowledge graph stores the sub-entries after the terms are broken down, the judgment conditions, and the judgment logic.
[0053] For example, if the disability level output in step S204 is level 9 and the entry number is 5.9.6.2, the system first searches the preset knowledge graph for all sub-entries corresponding to level 9 disability (including 5.9.6.1, 5.9.6.2, 5.9.6.3, etc.), then filters out the target sub-entry with entry number 5.9.6.2, extracts the judgment conditions ("fracture of one vertebral body and corresponding appendages, after surgical treatment; compression fracture of two vertebral bodies;") and judgment logic ("the three conditions are OR relations, where the first condition is decomposed into two sub-conditions, which are AND relations"), and determines them as the review criteria for this review.
[0054] S206: Determine the audit results based on the audit standards and disability assessment documents.
[0055] In this step, the review criteria can include judgment conditions and judgment logic. The implementation process can be divided into two steps: First, based on the disability assessment text, the judgment conditions in the review criteria are judged one by one to obtain the result of each judgment condition (i.e., logical input); Second, all logical inputs are substituted into the preset judgment logic, and the final review result (meets or does not meet the requirements) is obtained through logical calculation.
[0056] As described in the above embodiments, this disclosure acquires an image of a disability assessment certificate and determines the corresponding text. Based on the disability assessment text, prompts are generated, and these prompts are input into a first pre-trained language model to obtain the disability level and corresponding entry number output by the first pre-trained language model. Based on the disability level and entry number, review standards are determined. Then, combining the review standards and the disability assessment text, the review result is determined. This achieves automated conversion of image data into text and uses a pre-trained language model to find the corresponding disability level and entry number, increasing the efficiency of finding the disability level and entry number. In subsequent review processes, the corresponding review result is determined based on the review standards and the disability assessment text, avoiding the problem of decreased review accuracy due to the illusion of a large model when directly using a large language model for review.
[0057] In one possible implementation, the review criteria include judgment conditions and judgment logic.
[0058] In step S206 above, the review result is determined based on the review standards and the disability assessment document, including:
[0059] S2061: Determine the logical input based on the judgment conditions and the disability assessment text.
[0060] In this step, the logical input is the specific judgment result (such as true / false, specific numerical value) for each judgment condition. Different implementation methods are adopted according to the type of judgment condition (semantic condition, computational condition): semantic conditions focus on text semantic matching and are judged by a second pre-trained language model; computational conditions focus on numerical calculation and are calculated by extracting variables from the text and substituting them into the formula, so as to ensure the accuracy of judgment for different types of conditions and avoid errors caused by a single method.
[0061] S2062: Determine the audit result based on the logical input and judgment logic.
[0062] In this step, all logical inputs (such as true / false, specific numerical values) are first substituted into the judgment logic according to the format requirements of the judgment logic (such as AND, OR, numerical comparison) to be transformed into a directly executable logical expression (i.e., the logic to be executed); then the expression is executed by the server's built-in logic calculation engine to obtain the final audit result (compliant / non-compliant).
[0063] As can be seen from the description of the above embodiments, the embodiments of this disclosure determine the logical input based on the judgment conditions and the disability assessment text, and then determine the review result based on the logical input and the judgment logic, thereby achieving accurate acquisition of the logical input. Furthermore, the entire logical calculation process is executed by code, avoiding errors in manual calculation or large model logical judgment.
[0064] In one possible implementation, step S2061 above, determining the logical input based on the judgment conditions and the disability assessment text, includes:
[0065] S611: If the judgment condition is a semantic condition, then generate judgment prompt words based on the judgment condition and the disability assessment text. Input the judgment prompt words into the second pre-trained language model to obtain the logical input of the output of the second pre-trained language model.
[0066] In this step, semantic conditions refer to the conditions that need to be determined through text semantic matching (without specific numerical calculations, but focusing on descriptive judgment); the judgment prompt words are generated based on a preset template, which includes the judgment task (judging whether the text meets the conditions), the judgment conditions, and key fragments of the disability assessment text, ensuring that the prompt words are clear; the second pre-trained language model recognizes the degree of matching between the text and the conditions and outputs the logical input (true / false).
[0067] For example, the judgment condition (semantic condition) is "fracture of a vertebral body and corresponding appendages, after surgical treatment," and the disability assessment text contains "the patient suffered a fracture of the L1 vertebral body and corresponding appendages due to trauma, underwent open reduction and internal fixation of the fracture, the surgery was successful, and the postoperative recovery was acceptable." Based on the judgment condition and the disability assessment text, a judgment prompt is generated: "Please determine whether the following disability assessment text meets the given semantic condition, outputting only true or false. Semantic condition: fracture of a vertebral body and corresponding appendages, after surgical treatment; Disability assessment text: the patient suffered a fracture of the L1 vertebral body and corresponding appendages due to trauma, underwent open reduction and internal fixation of the fracture, the surgery was successful, and the postoperative recovery was acceptable." This prompt is input into the second pre-trained language model. The model, through semantic matching, determines that the text meets the condition and outputs the logical input "true."
[0068] S612: If the judgment condition is a calculation condition, then read the input variable corresponding to the judgment condition from the disability assessment text. Input the input variable into the judgment condition to obtain the logical input.
[0069] In this step, the calculation conditions can be conditions that need to be determined through numerical calculation (including specific calculation formulas and variables). First, by using keyword matching (such as "pronation degree" and "supination degree"), the input variables (specific values) required for the determination conditions are extracted from the disability assessment text. The completeness of the variables can also be verified to avoid missing necessary values. Then, the extracted variables are substituted into the calculation formula preset for the determination conditions, and the calculation is performed by the system code to obtain the calculation result, which is the logical input.
[0070] As can be seen from the description of the above embodiments, the embodiments of this disclosure generate corresponding prompt words based on the judgment conditions and disability assessment text when the judgment condition is a semantic condition. The prompt words are then input into a large language model, which performs semantic judgment. The language processing advantages of the large language model are used to quickly obtain the judgment result. For judgment conditions of type calculation condition, the input variables corresponding to the judgment conditions in the disability assessment text are read and input into the judgment conditions to obtain logical input. This achieves accurate calculation of judgment conditions of type calculation condition and obtains accurate logical input, avoiding the decrease in accuracy caused by the illusion of the large model.
[0071] In one possible implementation, step S2062 above, determining the audit result based on the logical input and judgment logic, includes:
[0072] S621: Substitute the logical input into the decision logic to obtain the logic to be executed.
[0073] In this step, the decision logic is a preset logical relationship (such as AND, OR, numerical comparison, etc.), which includes logical operators (>, <, ≥, ≤, etc.) and condition placeholders. The logical inputs (true / false, specific numerical values) obtained in S2061 are used to replace the condition placeholders in the decision logic one by one, and transformed into logical expressions that can be directly executed by the logic calculation engine (i.e., logic to be executed). The replacement process ensures that the variables and placeholders correspond one-to-one, avoiding replacement errors.
[0074] For example, the decision logic is "(left upper limb loss ≥ 75%) and (right upper limb loss ≥ 75%)", and the logic input is 83.3% (left) and 27.8% (right). Substituting the logic input into the decision logic and replacing the placeholders, we get the logic to be executed: "(83.3% ≥ 75%) and (27.8% ≥ 75%)".
[0075] S622: Execute the pending logic to obtain the audit result.
[0076] In this step, for Boolean logic (true / false judgment), logical operations are performed to obtain Boolean results; for numerical comparison logic, numerical comparisons are performed to obtain comparison results; and finally, the operation results are transformed into business-level audit results.
[0077] As can be seen from the description of the above embodiments, the embodiments of this disclosure obtain the audit result by substituting the logical input into the judgment logic, obtaining the logic to be executed, executing the logic to be executed, and obtaining the audit result.
[0078] In one possible implementation, step S205 above, determining the review criteria based on the disability level and item number, includes:
[0079] S2051: Read the sub-entry corresponding to the disability level in the preset knowledge graph.
[0080] In this step, the pre-defined knowledge graph uses disability level as the first-level node, with multiple sub-item nodes associated under each level node. The nodes are linked through the relationship of "level-sub-item". The system uses disability level (e.g., level 9) as the search keyword to call the knowledge graph, retrieve all associated sub-item nodes under that level node, and obtain the complete information of the sub-items (item number, clause content). The search results are sorted in ascending order by item number for easy subsequent filtering.
[0081] S2052: Identify the sub-entry corresponding to the entry number in the sub-entry as the target sub-entry.
[0082] In this step, the system matches all sub-entries read in S2051 by entry number and filters out sub-entries with the same entry number as those output in S204. The matching adopts an exact matching method (entry numbers are exactly the same, such as 5.9.6.2 matching the sub-entry with entry number 5.9.6.2) to avoid errors in target sub-entries caused by fuzzy matching. If no corresponding entry number is matched, the system will prompt "No corresponding sub-entry, please check the entry number".
[0083] S2053: Read the judgment conditions and judgment logic under the target sub-item and determine them as the audit criteria.
[0084] In this step, the knowledge graph contains associated decision condition nodes (which store all decision conditions under this entry, divided into semantic conditions and computational conditions) and decision logic nodes (which store the logical relationships between conditions) under the target sub-entry. The system reads the content of these two sub-nodes under the target sub-entry and integrates the decision conditions and decision logic into structured review standards to ensure that the review standards correspond one-to-one with the target sub-entry without omissions.
[0085] As can be seen from the description of the above embodiments, the embodiments of this disclosure read the sub-entries corresponding to the disability level in the atlas, and then determine the target sub-entry from the read sub-entries according to the entry number. The judgment conditions and judgment logic under the target sub-entry are determined as the review standards, thereby realizing the automatic and accurate positioning of the review standards.
[0086] In one possible implementation, before step S2051 (reading the sub-entry corresponding to the disability level in the preset knowledge graph) mentioned above, the following is also included:
[0087] S220: Generate map construction prompts based on the preset disability level classification standard text.
[0088] In this step, the disability grading standard text can be a pre-entered standard document by staff. The disability grading standard text can be combined with preset map generation prompts to obtain map construction prompts.
[0089] The prompts for generating the graph can include graph structure requirements (e.g., first-level nodes are disability levels, second-level nodes are sub-entries, and third-level nodes are judgment conditions and judgment logic).
[0090] S221: Input the knowledge graph construction prompts into the third pre-trained language model to obtain the preset knowledge graph output by the third pre-trained language model.
[0091] In this step, the third pre-trained language model can decompose the clauses and extract the relationships in the disability grading standard text, and generate knowledge graph data containing nodes (disability grade, sub-items, judgment conditions, judgment logic), node attributes (item number, condition content), and relationships between nodes. The output format conforms to the system's preset graph storage specifications.
[0092] After step S221 above, the preset knowledge graph can also be output for staff to verify.
[0093] As can be seen from the description of the above embodiments, the embodiments of this disclosure generate knowledge graph construction prompts based on preset disability level classification standard text, input them into a pre-trained language model, and obtain a knowledge graph, thereby achieving fast and accurate knowledge graph construction.
[0094] Figure 3 This is a schematic diagram of the structure of the certificate review device provided in an embodiment of this application. Figure 3 As shown, the certificate review device 300 includes: an image acquisition module 301, a text determination module 302, a prompt word generation module 303, an information acquisition module 304, a standard determination module 305, and a result determination module 306.
[0095] Image acquisition module 301 is used to acquire images of disability assessment reports;
[0096] The text determination module 302 is used to determine the disability assessment text based on the image of the disability assessment report;
[0097] The prompt word generation module 303 is used to generate prompt words based on the disability assessment text;
[0098] The information acquisition module 304 is used to input the prompt words into the first pre-trained language model and obtain the disability level and corresponding entry number output by the first pre-trained language model.
[0099] Standard determination module 305 is used to determine the review standards based on the disability level and item number;
[0100] The result determination module 306 is used to determine the audit result based on the audit standards and the disability assessment text.
[0101] The apparatus provided in this embodiment can be used to execute the technical solutions of the above method embodiments. Its implementation principle and technical effects are similar, and will not be described again here.
[0102] In one possible implementation, the review criteria include judgment conditions and judgment logic; the result determination module 306 is used to determine the logical input based on the judgment conditions and the disability assessment text; and to determine the review result based on the logical input and the judgment logic.
[0103] The apparatus provided in this embodiment can be used to execute the technical solutions of the above method embodiments. Its implementation principle and technical effects are similar, and will not be described again here.
[0104] In one possible implementation, the result determination module 306 is used to generate a judgment prompt word based on the judgment condition and the disability assessment text if the judgment condition is a semantic condition; input the judgment prompt word into the second pre-trained language model to obtain the logical input output by the second pre-trained language model; if the judgment condition is a computational condition, read the input variable corresponding to the judgment condition in the disability assessment text; input the input variable into the judgment condition to obtain the logical input.
[0105] The apparatus provided in this embodiment can be used to execute the technical solutions of the above method embodiments. Its implementation principle and technical effects are similar, and will not be described again here.
[0106] In one possible implementation, the result determination module 306 is used to substitute the logical input into the judgment logic to obtain the logic to be executed; and to execute the logic to be executed to obtain the audit result.
[0107] The apparatus provided in this embodiment can be used to execute the technical solutions of the above method embodiments. Its implementation principle and technical effects are similar, and will not be described again here.
[0108] In one possible implementation, the standard determination module 305 is used to read the sub-entries corresponding to the disability level in the preset knowledge graph; determine the sub-entries corresponding to the entry number in the sub-entries as target sub-entries; and read the judgment conditions and judgment logic under the target sub-entries to determine them as the review standards.
[0109] The apparatus provided in this embodiment can be used to execute the technical solutions of the above method embodiments. Its implementation principle and technical effects are similar, and will not be described again here.
[0110] In one possible implementation, the assessment certificate review device 300 further includes a knowledge graph construction module 307. The knowledge graph construction module 307 is used to generate knowledge graph construction prompts based on preset disability level classification standard text; input the knowledge graph construction prompts into a third pre-trained language model to obtain a preset knowledge graph output by the third pre-trained language model.
[0111] The apparatus provided in this embodiment can be used to execute the technical solutions of the above method embodiments. Its implementation principle and technical effects are similar, and will not be described again here.
[0112] To implement the above embodiments, this application also provides an electronic device.
[0113] refer to Figure 4 The diagram illustrates a structural schematic of an electronic device 400 suitable for implementing embodiments of this application. The electronic device 400 can be a terminal device or a server. The terminal device can include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, personal digital assistants (PDAs), portable Android devices (PADs), portable media players (PMPs), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 4 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.
[0114] like Figure 4 As shown, the electronic device 400 may include a processor (e.g., a central processing unit, a graphics processing unit, etc.) 401 and a memory 402 communicatively connected to the processor. The processor can perform various appropriate actions and processes based on programs stored in the memory 402, computer-executed instructions, or programs loaded from storage device 408 into random access memory (RAM) 403, to implement the certificate verification method in any of the above embodiments. The memory may be a read-only memory (ROM). The RAM 403 also stores various programs and data required for the operation of the electronic device 400. The processing device 401, the memory 402, and the RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.
[0115] Typically, the following devices can be connected to I / O interface 405: input devices 406 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, 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 electronic device 400 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 4 An electronic device 400 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] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable storage 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 memory 402. When the computer program is executed by processing device 401, it performs the functions defined in the methods of embodiments of this application.
[0117] It should be noted that the computer-readable storage medium described above in this application can be a computer-readable signal medium, a computer storage medium, or any combination of the two. 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 application, 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 application, 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 may also be any computer-readable storage 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 storage 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] The aforementioned computer-readable storage medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.
[0119] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the method shown in the above embodiments.
[0120] Computer program code for performing the operations of this application 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).
[0121] 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 application. 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.
[0122] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the units are not necessarily limiting of the module itself; for example, an image acquisition module can also be described as an "authentication certificate image acquisition module".
[0123] 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-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.
[0124] This application also provides a computer-readable storage medium storing computer-executable instructions. When a processor executes the computer-executable instructions, it implements the technical solution of the certificate review method in any of the above embodiments. Its implementation principle and beneficial effects are similar to those of the certificate review method, and can be found in the implementation principle and beneficial effects of the certificate review method, which will not be repeated here.
[0125] In the context of this application, 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. Machine-readable media 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 fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
[0126] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the technical solution of the certificate review method in any of the above embodiments. Its implementation principle and beneficial effects are similar to those of the certificate review method, and can be found in the implementation principle and beneficial effects of the certificate review method, which will not be repeated here.
[0127] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application 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 with similar functions disclosed in this application.
[0128] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.
[0129] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.
Claims
1. A method for verifying authentication certificates, characterized in that, include: Obtain an image of the disability assessment report; Based on the image of the disability assessment report, determine the text of the disability assessment report; Based on the disability assessment text, generate prompt words; The prompt word is input into the first pre-trained language model to obtain the disability level and corresponding entry number output by the first pre-trained language model; The review criteria are determined based on the disability level and the item number. The review results are determined based on the review criteria and the disability assessment documents.
2. The method according to claim 1, characterized in that, The review criteria include judgment conditions and judgment logic; The process of determining the review result based on the review standards and the disability assessment document includes: Based on the aforementioned judgment criteria and the aforementioned disability assessment text, determine the logical input; The audit result is determined based on the logical input and the judgment logic.
3. The method according to claim 2, characterized in that, The step of determining the logical input based on the judgment conditions and the disability assessment text includes: If the determination condition is a semantic condition, then a determination prompt word is generated based on the determination condition and the disability assessment text; the determination prompt word is input into the second pre-trained language model to obtain the logical input output by the second pre-trained language model; If the determination condition is a calculation condition, then read the input variable corresponding to the determination condition in the disability assessment text; input the input variable into the determination condition to obtain the logical input.
4. The method according to claim 2, characterized in that, The step of determining the audit result based on the logical input and the judgment logic includes: Substitute the logical input into the decision logic to obtain the logic to be executed; The pending logic is executed to obtain the audit result.
5. The method according to any one of claims 1 to 4, characterized in that, The process of determining the review criteria based on the disability level and the item number includes: Read the sub-entries corresponding to the disability level from the preset knowledge graph; The sub-entry corresponding to the entry number in the sub-entry is determined as the target sub-entry; Read the judgment conditions and judgment logic under the target sub-item and determine them as the audit criteria.
6. The method according to claim 5, characterized in that, Before the sub-entry corresponding to the disability level in the preset knowledge graph, the following is also included: Based on the preset disability level classification standard text, generate a map to construct prompt words; The knowledge graph is constructed by inputting the prompt words into a third pre-trained language model to obtain the preset knowledge graph output by the third pre-trained language model.
7. A certificate verification device, characterized in that, include: The image acquisition module is used to acquire images of the disability assessment report; The text determination module is used to determine the disability assessment text based on the image of the disability assessment report; The prompt word generation module is used to generate prompt words based on the disability assessment text; The information acquisition module is used to input the prompt words into the first pre-trained language model to obtain the disability level and corresponding entry number output by the first pre-trained language model. The standard determination module is used to determine the review standards based on the disability level and the item number; The result determination module is used to determine the review result based on the review standards and the disability assessment text.
8. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1 to 6.
10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1 to 6.