An artificial intelligence-based functional safety auditing method, device, medium, and apparatus

Abstract

The application provides an artificial intelligence function safety auditing method and device, a medium and equipment, specifically comprising: receiving a file to be audited; identifying the file format of the file to be audited; extracting text information of the file to be audited; performing feature extraction on the text information of the file to be audited to obtain element information of the file to be audited; generating an element list of the file to be audited; generating an audit result of the file to be audited based on the element list of the file to be audited and corresponding evaluation criteria; after a customer initiates function safety auditing, identifying and extracting text information of the file to be audited, further extracting element information in the text information and generating an element list, evaluating and scoring the confidence of each item of element information in the element list to obtain an audit result with confidence, which can greatly reduce the artificial workload, and the element information of the audit can be divided into confidence to ensure the accuracy of the audit.

Description

Technical Field

[0001] This application relates to the field of functional safety auditing technology, specifically to a functional safety auditing method, apparatus, medium, and equipment based on artificial intelligence. Background Technology

[0002] With the increasing complexity of electronic / electrical systems in modern road vehicles, ensuring their compliance with functional safety standards has become a crucial aspect of guaranteeing vehicle safety. Traditional functional safety audits typically rely on manual review and verification of numerous documents item by item. This process is cumbersome, time-consuming, and susceptible to human error, leading to omissions of important audit items or inconsistencies in audit results. Therefore, there is an urgent need to develop an automated functional safety audit solution to reduce human intervention, improve audit efficiency, and ensure audit accuracy. Summary of the Invention

[0003] To address the aforementioned technical problems, this application is proposed. Embodiments of this application provide a method, apparatus, medium, and device for functional safety auditing based on artificial intelligence.

[0004] According to one aspect of this application, an artificial intelligence-based functional safety audit method is provided, comprising: responding to a functional safety audit instruction initiated by a customer and receiving a document to be audited; identifying the file format of the document to be audited; wherein the file format of the document to be audited includes text format, image format, and portable file format; extracting text information of the document to be audited based on the file format of the document to be audited; performing feature extraction on the text information of the document to be audited to obtain element information of the document to be audited; generating an element list of the document to be audited based on the element information of the document to be audited; wherein the element list of the document to be audited stores structured data; generating an audit result of the document to be audited based on the element list of the document to be audited and corresponding evaluation criteria; wherein the audit result includes an evaluation result and a corresponding confidence level for each element information.

[0005] In one embodiment, the step of responding to a functional safety audit instruction initiated by a customer and receiving a file to be audited includes: receiving a functional safety audit instruction initiated by a customer and obtaining the file to be audited based on a path to the file to be audited included in the functional safety audit execution.

[0006] In one embodiment, extracting the text information of the document to be reviewed based on its file format includes: extracting the text information of the document to be reviewed using a corresponding text extraction method based on its file format; wherein the text information includes the document information and document content of the document to be reviewed.

[0007] In one embodiment, the step of extracting the text information of the file to be reviewed based on its file format and using a corresponding text extraction method includes: if the file format of the file to be reviewed is an image format or a portable file format, then converting the file to be reviewed into readable text; extracting the text content from the readable text to obtain the text information of the file to be reviewed.

[0008] In one embodiment, the step of extracting features from the text information of the document to be reviewed to obtain the element information of the document to be reviewed includes: extracting the initial text of the document to be reviewed based on semantic understanding and contextual association; and performing verification and error correction processing on the initial text to obtain the element information of the document to be reviewed.

[0009] In one embodiment, generating the element list of the document to be reviewed based on the element information of the document to be reviewed includes: filling the element information of the document to be reviewed into the table corresponding to the corresponding header based on the context information of the element information of the document to be reviewed, thereby obtaining the element list of the document to be reviewed.

[0010] In one embodiment, generating the review result of the document to be reviewed based on the element list and corresponding evaluation criteria includes: an evaluation result for each element based on the element list and corresponding evaluation criteria; and calculating the confidence level of the evaluation result based on the average log probability, similarity and consistency of the cited evidence for each element.

[0011] According to another aspect of this application, an artificial intelligence-based functional safety audit device is provided, comprising: a file receiving module for receiving a file to be audited in response to a functional safety audit instruction initiated by a customer; a file format recognition module for recognizing the file format of the file to be audited; wherein the file format of the file to be audited includes text format, image format, and portable file format; a text information extraction module for extracting text information of the file to be audited based on the file format of the file to be audited; an element information extraction module for performing feature extraction on the text information of the file to be audited to obtain element information of the file to be audited; an element list generation module for generating an element list of the file to be audited based on the element information of the file to be audited; wherein the element list of the file to be audited stores structured data; and an audit result generation module for generating an audit result of the file to be audited based on the element list of the file to be audited and the corresponding evaluation criteria; wherein the audit result includes the evaluation result and corresponding confidence level of each element information.

[0012] According to another aspect of this application, a computer-readable storage medium is provided, the storage medium storing a computer program for performing any of the methods described above.

[0013] According to another aspect of this application, an electronic device is provided, comprising: a processor; a memory for storing processor-executable instructions; the processor being configured to perform any of the methods described above.

[0014] This application provides a method, apparatus, medium, and device for functional safety auditing based on artificial intelligence, specifically including: responding to a functional safety audit instruction initiated by a customer and receiving a document to be audited; identifying the file format of the document to be audited; wherein the file format of the document to be audited includes text format, image format, and portable file format; extracting text information of the document to be audited based on the file format; performing feature extraction on the text information of the document to be audited to obtain element information of the document to be audited; generating an element list of the document to be audited based on the element information of the document to be audited; wherein structured data is stored in the element list of the document to be audited; generating an audit result of the document to be audited based on the element list of the document to be audited and the corresponding evaluation criteria; wherein the audit result includes the evaluation result and corresponding confidence level of each element information; after the customer initiates a functional safety audit, the text information of the document to be audited is identified and extracted, the element information in the text information is further extracted and an element list is generated, and each element information in the element list is evaluated and scored with confidence level to obtain an audit result with confidence level. This not only significantly reduces the amount of manual work, but also allows for confidence level classification of the audited element information to ensure the accuracy of the audit. Attached Figure Description

[0015] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0016] Figure 1 This is a flowchart illustrating an exemplary embodiment of the functional safety audit method based on artificial intelligence provided in this application.

[0017] Figure 2 This is a schematic diagram of the structure of an artificial intelligence-based functional safety audit device provided in an exemplary embodiment of this application.

[0018] Figure 3 This is a structural diagram of an electronic device provided in an exemplary embodiment of this application. Detailed Implementation

[0019] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0020] Figure 1 This is a flowchart illustrating an exemplary embodiment of the functional safety audit method based on artificial intelligence provided in this application. Figure 1 As shown, this AI-based functional safety audit method includes the following steps:

[0021] Step 110: Respond to the functional safety audit instruction initiated by the customer and receive the documents to be audited.

[0022] When a customer initiates a functional safety audit instruction, the system responds to the instruction, receives the document to be audited, and performs a functional safety audit on the document.

[0023] Step 120: Identify the file format of the document to be reviewed.

[0024] The file formats of the documents to be reviewed include text, image, and portable file formats. This application identifies the file format of the documents to be reviewed to determine its format and adopts corresponding processing methods for different file formats.

[0025] Step 130: Extract the text information of the document to be reviewed based on its file format.

[0026] This application extracts the text information of the document to be reviewed based on its file format and the corresponding processing method, so as to obtain the text content for review.

[0027] Step 140: Extract features from the text information of the document to be reviewed to obtain the element information of the document to be reviewed.

[0028] This application uses feature extraction to obtain element information from the text information of the document to be reviewed, thereby simplifying the document and reducing the difficulty of subsequent review, and also obtaining the review object and review content.

[0029] Step 150: Generate a list of elements for the documents to be reviewed based on the element information of the documents to be reviewed.

[0030] The element list of the document to be reviewed stores structured data. This application converts the element information of the document to be reviewed into structured data and stores it in the element list to obtain the element list of the document to be reviewed.

[0031] Step 160: Based on the list of elements in the document to be reviewed and the corresponding evaluation criteria, generate the review results for the document to be reviewed.

[0032] The review results include the evaluation result and corresponding confidence level for each element. This application evaluates each element in the element list of the document to be reviewed according to the corresponding evaluation criteria to obtain the corresponding evaluation result and confidence level, thereby obtaining the review result of the document to be reviewed.

[0033] This application provides an artificial intelligence-based functional safety audit method, specifically including: responding to a functional safety audit instruction initiated by a customer and receiving a document to be audited; identifying the file format of the document to be audited; wherein the file format of the document to be audited includes text format, image format, and portable file format; extracting text information of the document to be audited based on the file format; performing feature extraction on the text information of the document to be audited to obtain element information of the document to be audited; generating an element list of the document to be audited based on the element information of the document to be audited; wherein structured data is stored in the element list of the document to be audited; generating an audit result of the document to be audited based on the element list of the document to be audited and the corresponding evaluation criteria; wherein the audit result includes the evaluation result and corresponding confidence level of each element information; after the customer initiates a functional safety audit, the method identifies and extracts the text information of the document to be audited, further extracts the element information in the text information and generates an element list, evaluates and scores the confidence level of each element information in the element list to obtain an audit result with confidence level, which can significantly reduce the workload of manual labor and also classify the confidence level of the audited element information to ensure the accuracy of the audit.

[0034] In one embodiment, step 110 can be implemented by receiving a functional safety audit instruction initiated by a customer and obtaining the file to be audited based on the path to the file to be audited included in the functional safety audit execution.

[0035] This application receives a "Start Functional Safety Audit" instruction from a client (e.g., manually initiated by a functional safety engineer or auditor), which includes a path to the set of files to be audited or an uploaded file package. Optionally, this application verifies the validity of the functional safety audit instruction based on a predefined audit knowledge base and checklist that conforms to audit standards and relevant enterprise processes. For example, it confirms whether the files are accessible and valid. Upon successful verification, it initializes a unique identifier for this audit session and initiates the audit process.

[0036] In one embodiment, step 130 can be implemented as follows: based on the file format of the document to be reviewed, the text information of the document to be reviewed is extracted using the corresponding text extraction method; wherein, the text information includes the document information and document content of the document to be reviewed.

[0037] Specifically, the file format of the document to be reviewed may include text format (Word, Excel, PPT, etc.), image format, and portable file format (PDF). This application identifies the file format of the document to be reviewed and extracts the text information of the document to be reviewed by adopting the corresponding text extraction method according to the identification result. Specifically, it includes document information and document content. The document information includes document title, version, author, date, etc., and the document content includes safety life cycle stage identifiers (such as Hazard Analysis and Risk Assessment (HARA), Technical Safety Concept (TSC), Safety Requirements Specification (FSR / TSR), verification report, etc.) and core safety artifacts (such as Safety Goals, Automotive Safety Integrity Level (ASIL), Fault Tolerance Time Interval (FTTI), Safety Mechanisms, Metrics, etc.).

[0038] In one embodiment, step 130 can be implemented as follows: if the file format of the document to be reviewed is an image format or a portable file format, the document to be reviewed is converted into readable text; the text content in the readable text is extracted to obtain the text information of the document to be reviewed.

[0039] For non-text formats (such as images and scanned PDFs) of documents to be reviewed, an optical character recognition (OCR) engine can be called to convert the documents into machine-readable text. After obtaining the readable text, basic natural language processing technologies (such as named entity recognition, keyword matching, regular expressions, etc.) are applied to extract structured or semi-structured text information related to functional safety management from the readable text.

[0040] In one embodiment, step 140 can be implemented as follows: based on semantic understanding and contextual association, the initial text of the document to be reviewed is extracted; the initial text is verified and corrected to obtain the element information of the document to be reviewed.

[0041] Specifically, this application can employ a fine-tuned language model for text classification and sequence labeling. For example, a model can be trained to identify Hazard Analysis and Risk Assessment (HARA), Technical Safety Concepts (TSC), Safety Requirements Specifications (FSR / TSR), verification reports, etc. Even if abbreviations such as HARA, TSC, and FSR / TSR do not appear in the text, the model can still correctly classify them based on descriptive contexts such as "identifying hazards" and "risk assessment." For the extraction of core safety artifacts, a relation extraction model is used to identify and associate entities. For example, "ASIL D" is automatically paired with a specific "safety target" to extract core safety artifacts from the document. The extracted text is then validated and corrected to obtain a structured data set (e.g., a JSON object) containing key elements extracted from the initial text and establishing a mapping relationship with the corresponding positions in the document to be reviewed.

[0042] Preferably, this application can employ multi-path parallel extraction of the initial text of the document to be reviewed, specifically including precise extraction, semantic extraction, and location extraction. Precise extraction uses preset rules and regular expressions to perform high-precision extraction in structured areas (such as well-defined tables). Semantic extraction uses fine-tuned named entity recognition and relation extraction models to identify security elements and their associated information in unstructured text. Location extraction combines coordinate information from the document to be reviewed to extract metadata located in specific layout areas (such as the homepage title or header). Preferably, this application adjudicates and merges the results of the same element obtained from different paths based on source confidence (e.g., precise extraction > semantic extraction > location extraction) and logical consistency.

[0043] Preferably, after extracting the initial text, this application performs verification and error correction processing on the initial text. Specifically, for inconsistent spaces and separators in the initial text, standardization is performed (e.g., replacing all consecutive spaces / line breaks with single spaces), and regular expressions are used to allow flexible spacing. For typos and synonyms in the initial text, a business thesaurus is established, and the semantic similarity between text fragments and target keywords is calculated using a word vector model to perform fuzzy matching and achieve targeted replacement. For cases where poor image / scanning quality in the initial text leads to text distortion, an error correction mode is enabled and an engine that supports multiple languages ​​is selected. The extraction results are post-processed and corrected based on a dictionary, especially for professional terms. For key areas (such as titles and tables), a multi-optical character recognition engine result voting can be tried to select the result with the highest confidence. For cases where the information in the initial text is diverse or the position is not fixed, a "broad recall" strategy is adopted, using multiple modes to extract all possible candidate information (e.g., extracting all strings with similar dates), and then filtering is performed through business rule verification (e.g., selecting "document dates" that conform to logical order from multiple dates).

[0044] In one embodiment, step 150 can be implemented as follows: based on the context information of the element information of the document to be reviewed, the element information of the document to be reviewed is filled into the table corresponding to the corresponding header to obtain the element list of the document to be reviewed.

[0045] After extracting the element information of the document to be reviewed, this application fills the element information of the document to be reviewed into the corresponding table header according to the context information of the element information, thus obtaining the element list of the document to be reviewed. Optionally, this application can also accurately locate and annotate each original text segment, table cell, or chart title associated with each element information in the element list. Furthermore, this application can also perform problem diagnosis and annotation, using different colors of highlighting to distinguish the information status in the generated element list. For example, if the inspection item requires the definition of FTTI, but no FTTI-related description is found in the element list, the corresponding position in the element list is marked in red; for ambiguous, incomplete, or unclarified elements, without a clear explanation of their diagnostic coverage, the corresponding position in the element list is marked in yellow; for elements that have been identified and preliminarily meet the requirements, the corresponding position in the element list is marked in green. This application can also automatically insert structured annotations next to each highlight or at the end of the context, such as concise expenditure issues, citing specific clause numbers and content, citing authoritative sources, and providing modification or supplementary suggestions based on the knowledge base.

[0046] In one embodiment, step 160 can be implemented as follows: based on the list of elements of the document to be reviewed and the corresponding evaluation criteria, the evaluation result of each element information; based on the average log probability of each element information, the similarity and consistency of the cited evidence, the confidence level of the evaluation result is calculated.

[0047] This application evaluates each element in the element list based on corresponding evaluation criteria. For example, if the evidence is sufficient to meet the requirements, it is evaluated as compliant; if the evidence is missing or conflicts with the requirements, it is evaluated as non-compliant; if the evidence is incomplete or ambiguous, it is evaluated as pending / requires clarification; and if the evidence is inapplicable, it is evaluated as inapplicable. Furthermore, this application calculates a confidence level (between 0 and 1) for the evaluation result of each element. The formula for calculating the confidence level is: ,in, The average log probability of the corresponding element information. The similarity of the cited evidence for the corresponding element information (the semantic similarity between the cited evidence and the reference evidence). The degree of consistency of the corresponding element information (the degree of consistency of data extracted through multiple paths or multiple extractions). These are the weighting coefficients for the average log probability, the similarity of cited evidence, and the degree of consistency, respectively.

[0048] For each element of information, if its confidence level is greater than the first set value (e.g., 0.8), the evaluation result is included in the preliminary review qualified audit result for the auditor to confirm quickly; if its confidence level is between the first set value and the second set value (e.g., 0.5), the evaluation result is marked to remind the auditor to review; if its confidence level is less than the second set value, the manual review process is triggered for the auditor to review.

[0049] After the auditor completes the review, this application can generate an audit report (with a pre-defined template) and provide the report to the client. Optionally, this application can use the auditor's review results as training samples to retrain the model and optimize its audit accuracy.

[0050] Figure 2 This is a schematic diagram of the structure of an artificial intelligence-based functional safety audit device provided in an exemplary embodiment of this application. Figure 2As shown, the AI-based functional safety audit device 20 includes: a file receiving module 21, used to receive files to be audited in response to functional safety audit instructions initiated by a customer; a file format recognition module 22, used to recognize the file format of the files to be audited; wherein the file format of the files to be audited includes text format, image format, and portable file format; a text information extraction module 23, used to extract text information of the files to be audited based on the file format of the files to be audited; an element information extraction module 24, used to extract features from the text information of the files to be audited to obtain element information of the files to be audited; an element list generation module 25, used to generate an element list of the files to be audited based on the element information of the files to be audited; wherein the element list of the files to be audited stores structured data; and an audit result generation module 26, used to generate an audit result of the files to be audited based on the element list of the files to be audited and the corresponding evaluation criteria; wherein the audit result includes the evaluation result and corresponding confidence level of each element information.

[0051] This application provides an artificial intelligence-based functional safety audit device, which receives documents to be audited in response to functional safety audit instructions initiated by customers through a document receiving module 21; a document format recognition module 22 identifies the document format to be audited; wherein, the document format to be audited includes text format, image format, and portable file format; a text information extraction module 23 extracts the text information of the document to be audited based on the document format; an element information extraction module 24 extracts features from the text information of the document to be audited to obtain the element information of the document to be audited; and an element list generation module 25 generates an element list of the document to be audited based on the element information of the document to be audited. The document to be reviewed stores structured data in its element list. The review result generation module 26 generates the review result of the document to be reviewed based on the element list and the corresponding evaluation criteria. The review result includes the evaluation result and the corresponding confidence level of each element. After the customer initiates a functional safety review, the module identifies and extracts the text information of the document to be reviewed, further extracts the element information from the text information and generates an element list. Each element in the element list is evaluated and scored with confidence level to obtain a review result with confidence level. This not only greatly reduces the amount of manual work, but also allows for confidence level classification of the reviewed element information to ensure the accuracy of the review.

[0052] In one embodiment, the file receiving module 21 can be further configured to: receive a functional safety audit instruction initiated by a customer, and obtain the file to be audited based on the path to the file to be audited included in the functional safety audit execution.

[0053] In one embodiment, the text information extraction module 23 can be further configured to: extract the text information of the document to be reviewed based on the file format of the document to be reviewed, using the corresponding text extraction method; wherein, the text information includes the document information and document content of the document to be reviewed.

[0054] In one embodiment, the text information extraction module 23 can be further configured to: if the file format of the document to be reviewed is an image format or a portable file format, convert the document to be reviewed into readable text; extract the text content from the readable text to obtain the text information of the document to be reviewed.

[0055] In one embodiment, the element information extraction module 24 can be further configured to: extract the initial text of the document to be reviewed based on semantic understanding and contextual association; perform verification and error correction on the initial text to obtain the element information of the document to be reviewed.

[0056] In one embodiment, the element list generation module 25 can be further configured to: fill the element information of the document to be reviewed into the table corresponding to the corresponding header based on the context information of the element information of the document to be reviewed, so as to obtain the element list of the document to be reviewed.

[0057] In one embodiment, the above-mentioned audit result generation module 26 can be further configured to: evaluate the evaluation result of each element information based on the element list of the document to be audited and the corresponding evaluation criteria; and calculate the confidence level of the evaluation result based on the average log probability of each element information, the similarity and consistency of the cited evidence.

[0058] Below, for reference Figure 3 This application describes an electronic device according to embodiments thereof. The electronic device may be either or both of a first device and a second device, or a standalone device independent of them, which may communicate with the first device and the second device to receive acquired input signals from them.

[0059] Figure 3 A block diagram of an electronic device according to an embodiment of this application is illustrated.

[0060] like Figure 3 As shown, the electronic device 10 includes one or more processors 11 and memory 12.

[0061] The processor 11 may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0062] The memory 12 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 11 may execute the program instructions to implement the methods of the various embodiments of this application described above and / or other desired functions. Various contents such as input signals, signal components, and noise components may also be stored in the computer-readable storage medium.

[0063] In one example, the electronic device 10 may also include an input device 13 and an output device 14, which are interconnected via a bus system and / or other forms of connection mechanism (not shown).

[0064] When the electronic device is a standalone device, the input device 13 can be a communication network connector for receiving the collected input signals from the first device and the second device.

[0065] In addition, the input device 13 may also include, for example, a keyboard, a mouse, etc.

[0066] The output device 14 can output various information to the outside, including determined distance information, direction information, etc. The output device 14 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0067] Of course, for the sake of simplicity, Figure 3 Only some of the components of the electronic device 10 relevant to this application are shown in this illustration; components such as buses, input / output interfaces, etc., are omitted. In addition, the electronic device 10 may include any other suitable components depending on the specific application.

[0068] In addition to the methods and apparatus described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the methods according to various embodiments of this application described in the "Exemplary Methods" section above.

[0069] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0070] Furthermore, embodiments of this application may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform the steps in the methods according to various embodiments of this application described in the "Exemplary Methods" section above.

[0071] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable 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.

[0072] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0073] The block diagrams of devices, apparatuses, devices, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0074] It should also be noted that in the apparatus, equipment, and methods of this application, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions of this application.

[0075] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0076] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. An artificial intelligence-based functional safety auditing method, characterized by, The method comprises the following steps: In response to a functional safety audit instruction initiated by a client, a file to be audited is received; The file format of the file to be audited is identified; wherein the file format of the file to be audited includes a text format, a picture format, and a portable file format; Based on the file format of the file to be audited, text information of the file to be audited is extracted; Feature extraction is performed on the text information of the file to be audited to obtain element information of the file to be audited; Based on the element information of the file to be audited, an element list of the file to be audited is generated; wherein structured data is stored in the element list of the file to be audited; Based on the element list of the file to be audited and the corresponding evaluation standard, an audit result of the file to be audited is generated; wherein the audit result includes an evaluation result of each item of element information and a corresponding confidence level; The feature extraction on the text information of the file to be audited to obtain the element information of the file to be audited comprises: Based on semantic understanding and context association, initial text of the file to be audited is extracted; Verification and error correction processing is performed on the initial text to obtain the element information of the file to be audited; The extraction of the initial text of the file to be audited based on semantic understanding and context association comprises: Multi-path parallel extraction is adopted to extract the initial text of the file to be audited; wherein the multi-path includes accurate extraction, semantic extraction, and positioning extraction, the accurate extraction is high-precision extraction using pre-set rules and regular expressions in structured areas, the semantic extraction is to identify safety elements and their associated information using a fine-tuned named entity recognition model and a relation extraction model in unstructured text, and the positioning extraction is to extract metadata located in a specific layout area in combination with coordinate information in the file to be audited; The generation of the audit result of the file to be audited based on the element list of the file to be audited and the corresponding evaluation standard comprises: Based on the element list of the file to be audited and the corresponding evaluation standard, an evaluation result of each item of element information is obtained; The confidence of the evaluation result is calculated based on the average logarithmic probability of each item of element information, the similarity of the cited evidence, and the consistency degree, wherein the calculation formula of the confidence of the evaluation result is: wherein, is the average logarithmic probability of the corresponding item of element information, is the semantic similarity between the cited evidence and the reference evidence of the corresponding item of element information, is the data consistency degree of the multi-path extraction or multi-extraction of the corresponding item of element information, are weight coefficients of the average logarithmic probability, the similarity of the cited evidence, and the consistency degree, respectively. 2.The AI-based functional safety audit method of claim 1, wherein The receiving of the file to be audited in response to the functional safety audit instruction initiated by the client comprises: The functional safety audit instruction initiated by the client is received, and the file to be audited is obtained based on the path pointing to the file to be audited included in the functional safety audit execution. 3.The AI-based functional safety audit method of claim 1, wherein The extraction of the text information of the file to be audited based on the file format of the file to be audited comprises: Based on the file format of the file to be audited, a corresponding text extraction method is adopted to extract the text information of the file to be audited; wherein the text information includes document information and document content of the file to be audited. 4.The AI-based functional safety audit method of claim 3, wherein, The extraction of the text information of the file to be audited based on the file format of the file to be audited and the corresponding text extraction method comprises: If the file format of the file to be audited is a picture format or a portable file format, the file to be audited is converted into readable text; Text content in the readable text is extracted to obtain the text information of the file to be audited. 5.The AI-based functional safety audit method of claim 1, wherein The generation of the element list of the file to be audited based on the element information of the file to be audited comprises: Fill the element information of the file to be audited into the table corresponding to the corresponding table header based on the context information of the element information of the file to be audited, to obtain the element list of the file to be audited.

6. An apparatus for safety audit based on artificial intelligence function, characterized by, Comprise: A file receiving module is configured to receive a file to be audited in response to a functional safety audit instruction initiated by a client. A file format identification module is configured to identify the file format of the file to be audited; wherein the file format of the file to be audited comprises a text format, a picture format, and a portable file format. A text information extraction module is configured to extract text information of the file to be audited based on the file format of the file to be audited. An element information extraction module is configured to perform feature extraction on the text information of the file to be audited to obtain element information of the file to be audited. An element list generation module is configured to generate an element list of the file to be audited based on the element information of the file to be audited; wherein the element list of the file to be audited stores structured data. An audit result generation module is configured to generate an audit result of the file to be audited based on the element list of the file to be audited and corresponding evaluation criteria; wherein the audit result comprises evaluation results of each item of element information and corresponding confidence levels. The element information extraction module is further configured to: Extract initial text of the file to be audited based on semantic understanding and context association; Perform verification and error correction processing on the initial text to obtain element information of the file to be audited. The element information extraction module is further configured to: Extract initial text of the file to be audited in multiple paths in parallel; wherein the multiple paths comprise accurate extraction, semantic extraction, and positioning extraction, the accurate extraction is high-precision extraction using pre-set rules and regular expressions in structured areas, the semantic extraction is identifying safety elements and their associated information using a fine-tuned named entity recognition model and a relationship extraction model in unstructured text, and the positioning extraction is extracting metadata located in a specific layout area in combination with coordinate information in the file to be audited. The audit result generation module is further configured to: Obtain evaluation results of each item of element information based on the element list of the file to be audited and corresponding evaluation criteria. The confidence of the evaluation result is calculated based on the average logarithmic probability of each item of element information, the similarity of the cited evidence, and the consistency degree, wherein the calculation formula of the confidence of the evaluation result is: wherein, is the average logarithmic probability of the corresponding item of element information, is the semantic similarity between the cited evidence and the reference evidence of the corresponding item of element information, is the data consistency degree of the multi-path extraction or multi-extraction of the corresponding item of element information, are weight coefficients of the average logarithmic probability, the similarity of the cited evidence, and the consistency degree, respectively.

7. A computer readable storage medium characterized in that, The storage medium stores a computer program, and the computer program is used to execute the method of any one of claims 1-5.

8. An electronic device, comprising: Comprise: A processor; A memory for storing instructions executable by the processor; The processor is configured to execute the method of any one of claims 1-5.

Images