Intelligent question and answer method and system for enterprise safety and environmental protection management

Abstract

The invention discloses an intelligent question and answer method and system for enterprise safety and environmental protection management, and belongs to the technical field of intelligent question and answer. The method comprises the following steps: acquiring a security and environmental protection problem asked by a user, analyzing the security and environmental protection problem asked by the user by using a security and environmental protection semantic analysis model to obtain an analysis result, and judging whether the security and environmental protection problem asked by the current user belongs to the field of security and environmental protection systems or not according to the analysis result; if the method belongs to the field of security and environmental protection systems, rewriting a security and environmental protection problem proposed by a user by using a statement rewriting model to obtain an instruction statement; searching a rule unit in a security and environmental protection knowledge base based on the instruction statement to obtain a search result; and the recall summary model extracts content directly related to the user intention from the search result, and generates a safety and environmental protection answer. According to the method, the corresponding safety and environmental protection result can be generated in time according to the questions of the employees, the operation risk increase is reduced, and the safety and environmental protection system awareness rate of the employees is improved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent question-and-answer technology, specifically relating to an intelligent question-and-answer method, system, device, medium, and program for enterprise safety and environmental protection management. Background Technology

[0002] In the course of business operations, safe production and environmental protection are crucial cornerstones for ensuring the steady development of enterprises, concerning the safety of employees' lives, the safety of corporate property, and the sustainable development of the social ecological environment. From a legal perspective, with the continuous improvement of laws and regulations such as the "Production Safety Law" and the "Environmental Protection Law," enterprises must strictly comply with relevant safety and environmental protection regulations; otherwise, they will face the risk of hefty fines, production shutdowns, or even license revocation. From an operational perspective, safety and environmental accidents not only cause huge economic losses but also seriously damage the company's reputation, affecting its market competitiveness and sustainable development capabilities. From a social perspective, as important players in socio-economic activities, the safety and environmental performance of enterprises directly relates to the quality of life of surrounding communities and the balance of the ecological environment.

[0003] However, traditional safety and environmental management models have gradually revealed some problems. Previously, when employees encountered questions about safety and environmental regulations, they often had to resort to consulting numerous paper documents and safety administrators. This not only consumed a significant amount of time but could also lead to improper operation and increased safety and environmental risks due to untimely or inaccurate information transmission. Simultaneously, traditional safety and environmental training was mostly conducted in a centralized, offline manner, which was not only costly but also limited by time and space, resulting in poor employee participation and knowledge absorption, and hindering the improvement of policy awareness. Furthermore, safety administrators had to handle a large number of repetitive inquiries daily, consuming considerable energy and affecting their focus on core safety and environmental management work. Summary of the Invention

[0004] To address the problems in existing safety production system Q&A methods, such as untimely and inaccurate information delivery leading to non-standard operations, and time and space limitations in training resulting in poor employee participation and knowledge absorption, this invention provides an intelligent Q&A method for enterprise safety and environmental management. This method can generate corresponding safety and environmental protection results in a timely manner based on employee questions, reducing operational risks and improving employee awareness of safety and environmental protection regulations.

[0005] To achieve the above objectives, the present invention provides the following technical solution.

[0006] In a first aspect, the present invention provides an intelligent question-and-answer method for enterprise safety and environmental protection management, comprising: The system acquires safety and environmental protection questions raised by users, uses a safety and environmental protection semantic parsing model to parse these questions, obtains the parsing results, and determines whether the current safety and environmental protection questions raised by users fall within the scope of safety and environmental protection regulations based on the parsing results. If the safety and environmental protection question raised by the user falls under the scope of safety and environmental protection regulations, the statement rewriting model is used to rewrite the safety and environmental protection question raised by the user to obtain the instruction statement; The search results are obtained by searching for rule units in the safety and environmental protection knowledge base based on command statements, summarizing the matched rule units; The recall summary model extracts content directly related to user intent from search results to generate safe and environmentally friendly answers.

[0007] As a further improvement of the present invention, the step of obtaining the safety and environmental protection questions raised by the user, parsing the safety and environmental protection questions raised by the user using a safety and environmental protection semantic parsing model, obtaining the parsing results, and determining whether the safety and environmental protection questions raised by the user fall under the category of safety and environmental protection regulations based on the parsing results includes: The system receives safety and environmental protection questions from users, authenticates the user's identity and permissions, and then uses a safety and environmental protection semantic parsing model to parse the questions and obtain the parsing results. The analysis results are classified according to the preset prompt words and classification rules to determine whether the safety and environmental protection issues raised by the current user fall under the scope of safety and environmental protection regulations. If the safety and environmental protection question raised by the user does not fall under the scope of safety and environmental protection regulations, a general answer model will be invoked to answer the question.

[0008] As a further improvement of the present invention, if the safety and environmental protection issue raised by the current user belongs to the field of safety and environmental protection regulations, the user's safety and environmental protection issue is rewritten using a statement rewriting model to obtain instruction statements, including: If the safety and environmental protection issue raised by the user falls under the scope of safety and environmental protection regulations, then the user's safety and environmental protection issue will be input into the statement rewriting model. The statement rewriting model rewrites the security and environmental protection questions raised by the current user based on natural language, and obtains the instruction statement.

[0009] As a further improvement of the present invention, the step of searching for rule units in the safety and environmental protection knowledge base based on instruction statements, summarizing the matched rule units, and obtaining the search results includes: The system searches for rule units in the safety and environmental protection knowledge base based on command statements. When the matching degree between a rule unit and the safety and environmental protection question raised by the user is greater than or equal to a set value, the rule unit is considered a hit rule unit. The search results are obtained by summarizing the rule units that have at least one match. If no matching rule unit is found, the reply will be that no safety and environmental protection regulations related to your question were found.

[0010] As a further improvement of the present invention, the step of replying with relevant information if no matching rule unit is found includes: If no rule unit is matched, the user is prompted to supplement the information, and the instruction statement is regenerated based on the supplemented information. Based on the regenerated instruction statement, search for rule units in the safety and environmental protection knowledge base to obtain new query results; If no rule unit is found this time, we will reply that no safety and environmental protection regulations related to your question were found, and record the question in the missing question database.

[0011] As a further improvement of the present invention, the recall summary model extracts content directly related to the user's intent from the search results to generate safe and environmentally friendly answers, including: The recall summary model extracts key information from the search results to obtain the extracted content; The extracted content is logically organized, deduplicated, and integrated to generate a safe and environmentally friendly answer.

[0012] Secondly, the present invention provides an intelligent question-and-answer system for enterprise safety and environmental protection management, comprising: The parsing and judgment module is used to obtain the safety and environmental protection questions raised by users, use the safety and environmental protection semantic parsing model to parse the safety and environmental protection questions raised by users, obtain the parsing results, and determine whether the safety and environmental protection questions raised by users fall within the scope of safety and environmental protection regulations based on the parsing results. Instruction Statement Module: If the safety and environmental protection question raised by the current user belongs to the field of safety and environmental protection regulations, the module uses the statement rewriting model to rewrite the user's safety and environmental protection question to obtain instruction statements; Search Results Module: Used to search for rule units in the safety and environmental protection knowledge base based on command statements, summarize the matched rule units, and obtain the search results; Answer generation module: Used by the recall summary model to extract content directly related to the user's intent from the search results and generate safe and environmentally friendly answers.

[0013] Thirdly, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the intelligent question-and-answer method for enterprise safety and environmental management.

[0014] Fourthly, this invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the intelligent question-and-answer method for enterprise safety and environmental management.

[0015] Fifthly, the present invention provides a computer program product, including computer instructions, which, when executed by a processor, implement the aforementioned intelligent question-and-answer method for enterprise safety and environmental management.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This method acquires user-submitted safety and environmental protection questions and uses a safety and environmental protection semantic parsing model for precise analysis to determine whether the question falls within the scope of safety and environmental protection regulations. This avoids confusion and delays in information transmission. If the question belongs to this scope, a statement rewriting model is used to transform the question into an instruction statement, making the question expression more standardized and accurate, laying the foundation for accurate answer retrieval. Based on the instruction statement, rule units are searched in the safety and environmental protection knowledge base, and the hit results are summarized. Finally, a recall summary model extracts content directly related to the user's intent to generate the answer. The entire process is tightly integrated and operates efficiently, generating corresponding safety and environmental protection results in a timely manner based on employee questions, allowing employees to obtain accurate information immediately and effectively reducing operational risks caused by untimely and inaccurate information. Secondly, this method does not restrict users to specific times and locations, allowing them to obtain the necessary safety and environmental protection knowledge anytime, anywhere by asking questions, thus increasing employee participation. Attached Figure Description

[0017] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of the invention in any way. In the drawings: Figure 1 This is a flowchart illustrating an intelligent question-and-answer method for enterprise safety and environmental management according to the present invention. Figure 2 This is a schematic diagram of the business process in an intelligent question-and-answer method for enterprise safety and environmental management according to the present invention. Figure 3 This is a workflow architecture diagram of an intelligent question-and-answer method for enterprise safety and environmental management according to the present invention. Figure 4 This is a schematic diagram of the structure of an intelligent question-and-answer system for enterprise safety and environmental management according to the present invention; Figure 5 This is a schematic diagram of an electronic device in an embodiment of the present invention. Detailed Implementation

[0018] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. The described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0020] To address the problems in existing technologies regarding safety production system Q&A, such as untimely and inaccurate information delivery leading to increased operational risks, high training costs due to time and space constraints resulting in poor employee participation and knowledge absorption, and the distraction of safety administrators' attention by repeated inquiries affecting core management efforts, this invention provides an intelligent question-and-answer method for enterprise safety and environmental management, such as... Figure 1 As shown, it includes: S100: Obtain the safety and environmental protection questions raised by users, use the safety and environmental protection semantic parsing model to parse the safety and environmental protection questions raised by users, obtain the parsing results, and determine whether the safety and environmental protection questions raised by users belong to the field of safety and environmental protection regulations based on the parsing results; S200: If the safety and environmental protection question raised by the user belongs to the field of safety and environmental protection regulations, the user's safety and environmental protection question is rewritten using the statement rewriting model to obtain the instruction statement; S300: Based on the command statement, search for rule units in the safety and environmental protection knowledge base, summarize the matched rule units, and obtain the search results; S400: The recall summary model extracts content directly related to the user's intent from the search results to generate safe and environmentally friendly answers.

[0021] This method can generate corresponding safety and environmental protection results in a timely manner based on employee inquiries, reducing operational risks and increasing employees' awareness of safety and environmental protection regulations. It achieves multiple advantages: timely and accurate answers to safety production regulations to reduce operational risks; breaking down time and space limitations in training to improve employee knowledge absorption efficiency; and freeing up safety administrators' energy to focus on core management work.

[0022] The present invention will be further explained and described below with reference to the accompanying drawings.

[0023] like Figure 2As shown, this invention proposes an intelligent question-and-answer method for enterprise safety and environmental management, including: S1: Obtain information about users' safety and environmental concerns.

[0024] It receives security and environmental protection questions submitted by users through multiple terminals based on natural language, and authenticates the user's identity and permissions.

[0025] User identity and permissions are authenticated based on the user's enterprise Active Directory (AD) account to ensure that information is not accessed without authorization and to prevent information leakage or accidental operation.

[0026] This application sets up a three-tiered access control system based on job positions: ordinary employees can only access the relevant regulations for their own positions. For example, production operators can view equipment safety operation and hazardous chemical requisition regulations; administrative staff can view environmental protection requirements for office areas and waste classification standards; safety administrators can view all knowledge and operation logs; and system administrators have configuration permissions, such as knowledge base field configuration, search weight adjustment, batch import of user permissions, and model training parameter settings.

[0027] Based on user permissions, irrelevant knowledge can be filtered out. For example, a regular operator cannot see the audit report of senior management, thus narrowing the search scope and making the search response faster and the results more accurate. This directly supports the goal of responding within 2 seconds, meets the needs of real-time consultation at the work site, and also prevents the user experience from being worsened by seeing irrelevant information.

[0028] Furthermore, all queries are linked to specific individuals and roles, allowing security administrators to quickly pinpoint who queried what and when in the event of an incident or violation. This reduces the risk of system misuse and avoids management chaos caused by open information.

[0029] Multiple terminals include mobile devices or PCs, but users can only log in to one terminal device to make inquiries. Changing devices requires verification by the original device or approval by the security administrator. When using mobile devices, encrypted transmission is used to avoid data leakage.

[0030] This step utilizes a two-way mapping matrix of job tags and knowledge tags. The welding operator tag is associated with knowledge tags such as welding operation safety and environmentally friendly treatment of welding slag. When a user queries a question, knowledge units without associated tags are directly blocked, thus narrowing the search scope. The permission mapping relationship of commonly used jobs is preloaded locally to reduce database query time. Combined with subsequent search strategies, this ensures that more than 95% of query response times are ≤2 seconds, meeting the needs of scenarios with no tolerance for network latency, such as workshops and construction sites.

[0031] S2: Determine whether the safety and environmental protection questions asked by users fall under the scope of safety and environmental protection regulations.

[0032] Anonymize internal enterprise safety and environmental protection documents, historical Q&A records, and professional terminology databases to obtain safety and environmental protection text data. The base language model is trained and fine-tuned based on text data in the field of safety and environmental protection to form a semantic parsing model for safety and environmental protection. The safety and environmental protection semantic parsing model parses user-asked questions about safety and environmental protection and obtains the parsing results. The analysis results are categorized using preset prompt words and classification rules to determine whether the safety and environmental protection issues raised by the current user fall under the scope of safety and environmental protection regulations.

[0033] The following is an example of the preset prompt word classification rules: Criteria for identifying problems within the field: It includes core safety and environmental protection terms, such as safe operation, environmental compliance, emergency response, regulations and provisions, emission compliance, and protective measures. This involves enterprise safety and environmental protection related processes, such as approval, inspection, rectification, and reporting; Inquire about the rules, regulations, operating procedures, and prohibited items.

[0034] Criteria for determining issues outside the domain: This includes content related to non-safety matters such as salaries, personnel, and equipment maintenance, or content related to non-environmental matters such as administration and logistics. Casual questions without a clear focus on safety or environmental protection, such as "How's the weather today?" External regulatory consulting that goes beyond the scope of corporate management.

[0035] If the current user's safety and environmental protection issue is determined to fall under the scope of safety and environmental protection regulations, then proceed to S3.

[0036] If it is determined that the safety and environmental protection issues raised by the user do not fall within the scope of safety and environmental protection regulations, such as Figure 3 As shown, a general response model will be invoked to answer non-safety and environmental protection questions.

[0037] S3: The statement rewriting model rewrites the safety and environmental protection questions raised by users to obtain instruction statements.

[0038] A large number of real, conversational user questions were collected from historical customer service logs, safety inspection records, employee training questionnaires, incident report questions, and external industry Q&A examples. An annotation team composed of safety experts, knowledge base administrators, and frontline employee representatives was formed. Among them, safety experts accounted for 40% and were responsible for the accuracy of professional terminology; knowledge base administrators accounted for 30% and were responsible for the matching of clauses; and frontline employee representatives accounted for 30% and were responsible for the adaptation of conversational language. The annotation team annotated and rewrote each original question to obtain the original question and the corresponding rewritten question. For example, the original question was "Where should the workshop wastewater be discharged?" and the rewritten instruction statement was "Industrial wastewater from the production workshop, including the environmentally friendly discharge path, treatment process, and compliance standards for cleaning wastewater and cooling wastewater."

[0039] Each original problem and its corresponding rewritten problem are formatted into instruction samples for model training. The base text generation model is trained based on instruction samples, enabling it to learn the mapping relationship between the original question and the corresponding rewritten question, thus obtaining the sentence rewriting model.

[0040] The base text generation model needs to be able to generate text, follow instructions, and understand context.

[0041] The statement rewriting model will transcribe the user's question into an instruction statement.

[0042] By rewriting user-input natural language questions using a statement rewriting model, the questions become more standardized, complete, and include more key terms, enabling subsequent knowledge base retrieval steps to find answers more quickly and accurately. For example, "How to handle an oil leak?" can be transformed into "Emergency Response Procedures for Oil Spills in Production Areas."

[0043] During the application of the sentence rewriting model, the original question and its corresponding rewritten question are obtained, as well as the final answer of subsequent retrieval and question answering. For the original question and its corresponding rewritten question where the final answer is inaccurate or the user feedback is poor, the analysis is backtracked to determine whether there was a problem with the original question labeling and rewriting. Batch backtracking according to the set time, filtering rewritten samples with an answer dissatisfaction rate ≥30%; If a single rewritten sample receives ≥5 instances of inaccurate user feedback, the process will immediately begin the backtracking phase. Verify the rewritten question's match with the original intent. For example, if a user asks "What should I do if there's an oil leak?", does rewriting it as "Emergency response to oil leaks in production areas" cover the intent? Investigate whether the knowledge base search results are biased due to the rewritten statements not matching the keywords; If the issue is with annotation, re-annotate and rewrite; if it's a scenario the model hasn't learned, add it to the instruction sample library and use small-batch fine-tuning to ensure that rewriting the statement and iterating the model does not affect the existing performance.

[0044] S4: Based on the instruction statement, use a hybrid retrieval strategy to search for knowledge base content and obtain the search results; Extract text information from enterprise safety and environmental protection related documents such as "Safety Management System" and "Environmental Protection Operation Specifications"; establish a hierarchical relationship of the text information into a knowledge base based on the triple structure of clause number, applicable scenario, and core content to obtain semantic tags and hierarchical position information; form several independent safety rule units based on the semantic tags and hierarchical position information, and build a knowledge base based on the several independent safety rule units.

[0045] A hybrid retrieval strategy is employed for knowledge base retrieval. Keyword retrieval is used for specialized terms to ensure accuracy, while semantic vector retrieval is used for ambiguous expressions to improve recall. The weights of the two strategies are dynamically allocated, with a default weighting of 40% for keywords and 60% for semantic vectors. Once the retrieval hit rate consistently exceeds 90% after testing, the safety and environmental protection knowledge base is considered complete.

[0046] The system searches for content in the safety and environmental protection knowledge base based on command statements. If a match is found, the searched content is extracted; otherwise, the user is notified that no relevant content can be found.

[0047] A single rule unit is considered to have a match rate of ≥80% with the question. Multiple match results are sorted by match rate + relevance to the applicable scenario. No match found: No safety and environmental protection regulations related to your question were found. We suggest you provide the following information and try again: Specific work scenarios, such as production workshops or warehouses; The materials or equipment involved, such as oil or wastewater treatment equipment, The issue is also recorded in the missing issues database, triggering the knowledge base administrator to check for any missing clauses.

[0048] S5: The recall summary model summarizes and rewrites the searched knowledge base content and outputs it to the user.

[0049] The recall summary model performs deep semantic understanding and extracts key information based on the search knowledge base. Following the intent of the user's original question, the model logically organizes, deduplicates, and integrates these scattered, sometimes redundant, and even context-deficient clauses to generate a clear, fluent, and directly addressing the user's core needs for a safe and environmentally friendly answer.

[0050] The recall summary model extracts content directly related to the user's intent from multiple hit rule units. For example, if a user asks about hot work approval, the recall summary model only extracts the approval process, required materials, and approval time limit, filtering out irrelevant protective measures clauses. Then, the content is reorganized in the order of process first, then requirements, and core information first, then supplementary information. For example, the approval process is: submit application → safety review → supervisor approval → receive work permit; supplementary requirements: approval time limit ≤ 24 hours, special scenarios require approval from more than 2 safety administrators; duplicate statements in different rule units are deleted. For example, if multiple clauses mention that safety administrator review is required, only one is retained.

[0051] Rule 1: Hot work requires the submission of a "Hot Work Application Form", specifying the work location, time, and risk level; Rule 2: Hot work approval is the responsibility of the safety manager; work can only proceed after approval. Rule Unit 3: Special hot work operations require approval from at least two safety managers, and the approval time limit is ≤24 hours; The following is the approval process and requirements for hot work: 1. Submit a "Hot Work Application Form," specifying the work location, time, and risk level; 2. Routine hot work operations: subject to review by the safety manager; 3. Special hot work operations must be reviewed by two safety managers within a time limit of ≤48 hours. 4. Work will commence after approval.

[0052] The core content of the output answer is bolded; the cited clause number is marked at the end of the answer, and users can click to view the full text of the original clause; users can vote for the answer as satisfied or dissatisfied, and can select reasons for dissatisfaction, such as incomplete answer, vague description, or inaccurate information.

[0053] Monthly analysis of feedback data: if incomplete answers account for ≥20%, optimize the key information extraction algorithm of the recall summary model; if discrepancies with reality account for ≥10%, verify the timeliness and accuracy of the knowledge base terms.

[0054] In summary, this method, through multi-terminal access and natural language interaction technology, enables employees to obtain accurate safety policy answers instantly at the work site, overcoming the delays and distortions inherent in traditional layered information transmission models, ensuring operational standardization, and reducing human error and accident risks from the source. Secondly, relying on a 24 / 7 online personalized Q&A mechanism, it can replace some centralized training, enabling employees to learn on demand in fragmented and scenario-based ways, avoiding problems such as training time conflicts and rigid content, and significantly improving the efficiency of safety knowledge acquisition. Thirdly, the system automatically handles a large number of repetitive policy inquiries, freeing safety administrators from tedious daily Q&A, allowing them to focus on core management tasks such as on-site inspections, risk analysis, and emergency drills, comprehensively improving the enterprise's safety management level.

[0055] This method also employs a three-tiered permission system and a job-related knowledge filtering mechanism to protect sensitive information while significantly narrowing the search scope, achieving a rapid response within 2 seconds. It ensures that employees only access policy content relevant to their own positions, effectively improving query accuracy and user experience. Using a domain-adaptive semantic parsing and sentence rewriting model, it transforms colloquial and ambiguous questions into standardized queries. A hybrid retrieval strategy combining keywords and semantic vectors balances accuracy and recall, ensuring the reliability and completeness of answers. All query operations are linked to personnel and job information, enabling full-process traceability and providing strong support for incident retrospective analysis and accountability. Simultaneously, through user feedback and data feedback mechanisms, it continuously optimizes semantic parsing, sentence rewriting models, and the knowledge base, constructing a self-iterating intelligent management closed loop. Based on enterprise AD account authentication and hierarchical permission control, it ensures information security and compliant access. During the knowledge base construction phase, original documents are anonymized to prevent sensitive data leakage, fully meeting enterprise security and privacy protection requirements.

[0056] Furthermore, this method improves the accuracy of question and answer, and the combination of hybrid retrieval achieves a system matching accuracy rate of 92%; the rewriting function reduces the time to solve complex problems from 30 minutes to 1 minute, significantly improving work efficiency; the large model based on natural language analysis reduces the cost of manual consultation by 70%, ensuring the timeliness of knowledge; and the instant response to operational questions reduces the risk of non-compliance in pilot departments by 42%, realizing the transformation of safety management from "post-event rectification" to "pre-event prevention".

[0057] The second objective of this invention is to propose an intelligent question-and-answer system for enterprise safety and environmental management, such as... Figure 4 As shown, it includes: Parsing and Judgment Module 100: Used to obtain safety and environmental protection questions raised by users, use a safety and environmental protection semantic parsing model to parse the safety and environmental protection questions raised by users, obtain parsing results, and determine whether the safety and environmental protection questions raised by users belong to the field of safety and environmental protection regulations based on the parsing results; Instruction Statement Module 200: If the safety and environmental protection question raised by the current user belongs to the field of safety and environmental protection regulations, the instruction statement module 200 is used to rewrite the safety and environmental protection question raised by the user using the statement rewriting model to obtain the instruction statement. Search Results Module 300: Used to search for rule units in the safety and environmental protection knowledge base based on command statements, summarize the matched rule units, and obtain the search results; Answer generation module 400: Used by the recall summary model to extract content directly related to the user's intent from the search results and generate safe and environmentally friendly answers.

[0058] This invention proposes an intelligent question-and-answer system for enterprise safety and environmental management. It integrates data on the company's internal safety and environmental regulations, opens a query port for company employees, and realizes functions such as dialogue analysis, knowledge base query, and dialogue return for employees' safety and environmental issues through the application of a large model. It eliminates the traditional cumbersome document query process, improves the efficiency and accuracy of employees' responses to safety and environmental issues, and ensures the safety and smoothness of production operations in the company.

[0059] like Figure 5 As shown, a third objective of this invention is to provide an electronic device comprising a processor 501, a memory 502, and a display screen 503. The memory 502 and the display screen 503 are both connected to the processor 501, such as via a bus 504. Optionally, the electronic device may further include a transceiver 505. It should be noted that in practical applications, the transceiver 505 is not limited to one type, and the structure of this electronic device does not constitute a limitation on the embodiments of this application.

[0060] Processor 501 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 501 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0061] Bus 504 may include a pathway for transmitting information between the aforementioned components. Bus 504 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 504 can be divided into address bus, data bus, control bus, etc.

[0062] The memory 502 may be a ROM (Read Only Memory) or other type of static storage device capable of storing static information and instructions, RAM (Random Access Memory) or other type of dynamic storage device capable of storing information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0063] The memory 502 is used to store application code that executes the solution of this application, and its execution is controlled by the processor 501. The processor 501 is used to execute the application code stored in the memory 502 to implement the content shown in the foregoing method embodiments.

[0064] Figure 5 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.

[0065] A fourth objective of this invention is to provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, performs the aforementioned functions. Figure 1 The illustrated method embodiments include various processes. For example, a memory may include instructions that can be executed by a processor of an electronic device to perform the described method.

[0066] A computer-readable storage medium can be a tangible device that holds and stores instructions used by an instruction execution device. A computer-readable storage medium can be, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof. Specifically, a computer-readable storage medium can be a portable computer disk, a hard disk, a USB flash drive, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), staging random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory stick, floppy disk, optical disk, magnetic disk, mechanical encoding device, or any combination thereof.

[0067] A fifth objective of this invention is to provide a computer program product comprising computer instructions that, when executed by a processor, implement the above-described... Figure 1 The various processes of the method embodiments shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0068] The above content provides a further detailed description of the present invention. It should not be construed that the specific embodiments of the present invention are limited to this. For those skilled in the art, several simple deductions or substitutions can be made without departing from the concept of the present invention, and all such deductions or substitutions should be considered to fall within the scope of protection of the present invention.

Claims

1. An intelligent question-and-answer method for enterprise safety and environmental management, characterized in that, include: The system acquires safety and environmental protection questions raised by users, uses a safety and environmental protection semantic parsing model to parse these questions, obtains the parsing results, and determines whether the current safety and environmental protection questions raised by users fall within the scope of safety and environmental protection regulations based on the parsing results. If the safety and environmental protection question raised by the user falls under the scope of safety and environmental protection regulations, the statement rewriting model is used to rewrite the safety and environmental protection question raised by the user to obtain the instruction statement; The search results are obtained by searching for rule units in the safety and environmental protection knowledge base based on command statements, summarizing the matched rule units; The recall summary model extracts content directly related to user intent from search results to generate safe and environmentally friendly answers.

2. The intelligent question-and-answer method for enterprise safety and environmental protection management according to claim 1, characterized in that, The process involves obtaining user-submitted safety and environmental protection questions, parsing them using a safety and environmental protection semantic parsing model, obtaining parsing results, and determining whether the user-submitted safety and environmental protection question falls under the scope of safety and environmental protection regulations based on the parsing results. The system receives safety and environmental protection questions from users, authenticates the user's identity and permissions, and then uses a safety and environmental protection semantic parsing model to parse the questions and obtain the parsing results. The analysis results are classified according to the preset prompt words and classification rules to determine whether the safety and environmental protection issues raised by the current user fall under the scope of safety and environmental protection regulations. If the safety and environmental protection question raised by the user does not fall under the scope of safety and environmental protection regulations, a general answer model will be invoked to answer the question.

3. The intelligent question-and-answer method for enterprise safety and environmental protection management according to claim 1, characterized in that, If the safety and environmental protection question raised by the user falls under the scope of safety and environmental protection regulations, the statement rewriting model is used to rewrite the user's safety and environmental protection question to obtain instruction statements, including: If the safety and environmental protection issue raised by the user falls under the scope of safety and environmental protection regulations, then the user's safety and environmental protection issue will be input into the statement rewriting model. The statement rewriting model rewrites the security and environmental protection questions raised by the current user based on natural language, and obtains the instruction statement.

4. The intelligent question-and-answer method for enterprise safety and environmental protection management according to claim 1, characterized in that, The process involves searching for rule units in the safety and environmental protection knowledge base based on command statements, summarizing the matched rule units, and obtaining the search results, including: The system searches for rule units in the safety and environmental protection knowledge base based on command statements. When the matching degree between a rule unit and the safety and environmental protection question raised by the user is greater than or equal to a set value, the rule unit is considered a hit rule unit. The search results are obtained by summarizing the rule units that have at least one match. If no matching rule unit is found, the reply will be that no safety and environmental protection regulations related to your question were found.

5. The intelligent question-and-answer method for enterprise safety and environmental protection management according to claim 4, characterized in that, If no rule unit is matched, the system will reply with relevant information, including: If no rule unit is matched, the user is prompted to supplement the information, and the instruction statement is regenerated based on the supplemented information. Based on the regenerated instruction statement, search for rule units in the safety and environmental protection knowledge base to obtain new query results; If no rule unit is found this time, we will reply that no safety and environmental protection regulations related to your question were found, and record the question in the missing question database.

6. The intelligent question-and-answer method for enterprise safety and environmental protection management according to claim 1, characterized in that, The recall summary model extracts content directly related to user intent from the search results to generate safe and environmentally friendly answers, including: The recall summary model extracts key information from the search results to obtain the extracted content; The extracted content is logically organized, deduplicated, and integrated to generate a safe and environmentally friendly answer.

7. An intelligent question-and-answer system for enterprise safety and environmental management, based on the intelligent question-and-answer method for enterprise safety and environmental management as described in any one of claims 1-6, characterized in that, include: The parsing and judgment module is used to obtain the safety and environmental protection questions raised by users, use the safety and environmental protection semantic parsing model to parse the safety and environmental protection questions raised by users, obtain the parsing results, and determine whether the safety and environmental protection questions raised by users fall within the scope of safety and environmental protection regulations based on the parsing results. Instruction Statement Module: If the safety and environmental protection question raised by the current user belongs to the field of safety and environmental protection regulations, the module uses the statement rewriting model to rewrite the user's safety and environmental protection question to obtain instruction statements; Search Results Module: Used to search for rule units in the safety and environmental protection knowledge base based on command statements, summarize the matched rule units, and obtain the search results; Answer generation module: Used by the recall summary model to extract content directly related to the user's intent from the search results and generate safe and environmentally friendly answers.

8. An electronic device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the intelligent question-and-answer method for enterprise safety and environmental management as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the intelligent question-and-answer method for enterprise safety and environmental management as described in any one of claims 1-6.

10. A computer program product, characterized in that, It includes computer instructions that, when executed by a processor, implement an intelligent question-and-answer method for enterprise safety and environmental management as described in any one of claims 1-6.

Images