Laboratory safety education course integration pushing system and method

Abstract

The invention relates to the technical field of machine learning and block chains, and discloses a laboratory safety education course integration pushing system and method. A trainee grading and classifying subsystem, a course data grading and classifying subsystem and a wrong question grading and classifying subsystem of the system respectively use a deep learning technology to automatically grade and classify trainees, link addresses corresponding to laboratory safety training data, error-prone question types of the trainees and question answering error rates. And the course grading and classification pushing subsystem automatically pushes personalized link address sets corresponding to laboratory safety training data to different trainees by using an unsupervised learning technology, so that the working efficiency is improved. In addition, the course data grading and classifying subsystem integrates link addresses corresponding to laboratory safety training data from different sources inside and outside the school by adopting a block chain technology, and the related problems of communication bottleneck, single-point failure and other privacy leakage caused by a centralized network topology structure can be avoided.

Description

Technical Field

[0001] This application relates to the fields of machine learning and blockchain technology, and specifically to a laboratory safety education course integrated push system and method. Background Art

[0002] University laboratories integrate theory and practical application across various disciplines, providing crucial support for education, scientific research, innovation, competition, and talent development. Laboratories involve high-risk factors such as hazardous chemicals and high-temperature, high-pressure equipment. Accidents can result in serious injury or death, endangering the lives and health of on-site personnel. Furthermore, laboratory equipment (especially precision instruments) is valuable, and improper operation or inadequate maintenance can lead to equipment damage and increased repair costs. Laboratory safety accidents can interrupt or even terminate scientific research projects and may also pollute the surrounding environment (e.g., leaks of toxic and hazardous chemicals or the spread of biological pathogens). Laboratory safety training and education aims to impart laboratory rules and regulations and codes of conduct to participants, ensuring they are internalized, practiced, and effectively implemented. Laboratory safety training and education can enhance safety awareness and operational standards among laboratory managers and users, cultivate a scientific spirit and a sense of responsibility, strengthen emergency response capabilities, reduce accident losses, and ensure the continuity of scientific research.

[0003] In recent years, most schools have introduced laboratory safety courses and established safety access systems. Traditional university laboratory safety education lacks system support and relies entirely on laboratory safety training management units to manually compile statistics on personnel information and categories and hold offline lectures for publicity and education. Offline training is often restricted by time and location, lacking flexibility, which may prevent some employees from attending or fully participating in training activities. Furthermore, offline training often has a fixed curriculum that may not meet the individual needs of different personnel. Furthermore, offline training involves multiple aspects such as venue arrangement, personnel coordination, and material preparation, making it difficult to manage and prone to organizational and coordination issues.

[0004] Some universities have adopted online safety education platforms, allowing students to learn laboratory safety knowledge and pass exams through these systems. However, existing university laboratory safety education systems often lack intelligence and personalization, and are unable to deliver personalized and customized training content based on the specific needs of trainees. Furthermore, course material data between internal and external training repositories is often isolated from one another. While research has explored integrating course material across knowledge bases by building federated learning technology, this centralized network topology often faces communication bottlenecks. Furthermore, all training participants require a central server to coordinate training, leaving the entire technical approach with the potential for a single point of failure. Summary of the Invention

[0005] This application provides a laboratory safety education course integration and push system to solve the problem in the existing technology that most existing university laboratory safety education systems lack intelligence and personalization, and there is a single point of failure when integrating course materials between different training databases.

[0006] Correspondingly, this application also provides a laboratory safety education course integration push method to ensure the implementation and application of the above system.

[0007] In order to solve the above technical problems, the present application discloses a laboratory safety education course integrated push system, which includes:

[0008] The training personnel grading and classification subsystem is used to grade and classify training personnel using deep learning technology to obtain training personnel grading and classification information;

[0009] The course material grading and classification subsystem is used to integrate the link addresses corresponding to different laboratory safety training materials into the laboratory safety training material database through blockchain technology, and use deep learning technology to grade and classify the link addresses of laboratory safety training materials in the laboratory safety training material database to obtain course grading and classification information;

[0010] The wrong question grading and classification subsystem is used to count the types of questions that trainees are prone to making mistakes and the error rates of their answers. Deep learning technology is used to grade and classify the types of questions that are prone to making mistakes and the error rates of each trainee, and to obtain the graded and classified information of each trainee's wrong questions.

[0011] The course grading and classification push subsystem is used to use unsupervised learning technology to process the grading and classification information of trainers, course grading and classification information, and wrong question grading and classification information, obtain the link address set corresponding to all laboratory safety training materials matching each trainer, and push them to the trainers.

[0012] This application also discloses a method for integrating and pushing laboratory safety education courses, the method comprising:

[0013] Use deep learning technology to classify trainees and obtain training personnel classification information;

[0014] Through blockchain technology, the link addresses corresponding to different laboratory safety training materials are integrated and stored in the laboratory safety training database. Deep learning technology is used to classify the link addresses of laboratory safety training materials in the laboratory safety training database to obtain course classification information.

[0015] Collect statistics on the types of questions that trainees are prone to making mistakes and their error rates, and use deep learning technology to classify each trainee's types of questions that are prone to making mistakes and their error rates, to obtain the classification information of each trainee's wrong questions;

[0016] Unsupervised learning technology is used to process the classification information of trainers, courses and wrong questions, and a set of link addresses corresponding to all laboratory safety training materials matching each trainer is obtained and pushed to the trainers.

[0017] In this application, the training personnel classification subsystem, the course material classification subsystem, and the wrong question classification subsystem all use deep learning technology to automatically classify the identity level categories of the training personnel, the link addresses of the laboratory safety training materials, the types of questions that are easy to make mistakes by the training personnel, and the error rates of the answers. On this basis, the course classification push subsystem uses unsupervised learning technology to automatically push personalized laboratory safety training material link addresses to different training personnel, thereby improving work efficiency. In addition, the course material classification subsystem uses blockchain technology to integrate the link addresses of laboratory safety training materials from different sources inside and outside the school, which can avoid communication bottlenecks, single points of failure, and other privacy leakage-related issues caused by centralized network topology structures.

[0018] Additional aspects and advantages of the present application will be given in the following description, which will become apparent from the following description, or will be understood through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0020] Figure 1 A schematic diagram of the framework of the laboratory safety education course integrated push system provided in the embodiment of this application;

[0021] Figure 2 A schematic diagram of the training personnel classification subsystem framework provided in an embodiment of the present application;

[0022] Figure 3 A flowchart of constructing a mature model for grading and classifying training personnel provided in an embodiment of the present application;

[0023] Figure 4 A flowchart for querying and integrating link addresses of cross-domain training course materials from different laboratory safety training databases provided in an embodiment of the present application;

[0024] Figure 5 A schematic diagram of the hierarchical classification framework for course materials provided for the application embodiment;

[0025] Figure 6 A flowchart of a mature model for grading and classifying course materials provided in an embodiment of the present application;

[0026] Figure 7 A schematic diagram of the framework of the wrong question classification and push subsystem provided in an embodiment of the present application;

[0027] Figure 8 A flowchart of a mature model for grading and classifying wrong questions and pushing them to the application embodiment;

[0028] Figure 9 A schematic diagram of the framework of the subsystem for grading and classifying and pushing training course materials provided in an embodiment of the present application;

[0029] Figure 10 A flowchart of a mature model for grading and classifying course materials and pushing them in accordance with an embodiment of the present application;

[0030] Figure 11 A flowchart of the laboratory safety education course integration and push method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0031] The following describes embodiments of the present application in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application and are not to be construed as limiting the present application.

[0032] It will be understood by those skilled in the art that, unless otherwise stated, the singular forms "a," "an," "said," and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of this application refers to the presence of features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or combinations thereof. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0033] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those skilled in the art in the art to which the present invention pertains. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with their meanings in the context of the prior art and, unless specifically defined as herein, will not be interpreted in an idealized or overly formal sense.

[0034] In response to the technical problems existing in the prior art, the present application provides a laboratory safety education course integrated push system and method, which aims to solve at least one of the technical problems of the prior art.

[0035] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0036] The present application embodiment provides a laboratory safety education course integrated push system, such as Figure 1 As shown in , the system includes:

[0037] The training personnel grading and classification subsystem is used to grade and classify training personnel using deep learning technology to obtain training personnel grading and classification information;

[0038] In the embodiment of the present application, according to the requirements of laboratory safety education and training in colleges and universities, deep learning technology is used to conduct a refined analysis of the training personnel. Under the major categories of professional background (such as chemistry, physics, biology, etc.), the training personnel are automatically graded and classified according to general teachers, professional teachers, general graduate students, professional graduate students, general undergraduates, and professional undergraduates. The laboratory safety education course integration push system can provide different training personnel with the link address of the laboratory safety training materials that match them according to the results of the graded classification, avoiding interference and deviation of the graded classification results of the personnel participating in the training by human factors, while also reducing labor costs and improving management efficiency.

[0039] The course material grading and classification subsystem is used to integrate the link addresses corresponding to different laboratory safety training materials into the laboratory safety training material database through blockchain technology, and use deep learning technology to grade and classify the link addresses of laboratory safety training materials in the laboratory safety training material database to obtain course grading and classification information.

[0040] In the embodiment of the present application, the school's existing experimental safety education resources are integrated through blockchain technology, including the school's own and purchased experimental safety education databases, and experimental safety education resources obtained free of charge through platforms such as Super Star, China University MOOC, and Wisdom Tree. The link addresses of laboratory safety training materials inside and outside the school are classified according to "vertical classification and horizontal classification" to form a hierarchical and classified school-level general and professional-level laboratory safety training database. The method of integrating different laboratory safety training material knowledge bases inside and outside the school based on blockchain technology gets rid of the dependence on the "parameter server". Each participant directly communicates in a P2P manner to realize the integration of laboratory safety training materials. There is no need for the role of "parameter server" and no need for a shared model. Therefore, the original laboratory safety training materials will not be moved to a place outside the target laboratory safety training material link address data storage node, reducing the risk of data leakage during transmission and solving the problem of laboratory safety training material course data integration in cross-organizational scenarios. It can also avoid communication bottlenecks, single point failures, and other privacy leakage issues caused by the central node. At the same time, deep learning technology is used to grade and classify the link addresses of laboratory safety training materials to avoid interference and deviation in the grading and classification results caused by human factors, reducing labor costs and improving management efficiency.

[0041] The wrong question grading and classification subsystem is used to count the types of questions that are easy to make mistakes and the error rates of answering questions by trainees. It uses deep learning technology to grade and classify the types of questions that are easy to make mistakes and the error rates of answering questions for each trainee, and obtains the wrong question grading and classification information for each trainee.

[0042] The course grading and classification push subsystem is used to use unsupervised learning technology to process the grading and classification information of trainers, course grading and classification information, and wrong question grading and classification information, obtain the link address set corresponding to all laboratory safety training materials matching each trainer, and push them to the trainers.

[0043] The laboratory safety education course integration and push system uses deep learning technology. According to the hierarchical classification results of the trainees and the hierarchical and classified school-level general and professional laboratory safety training database, it pushes a set of link addresses corresponding to systematic, professional, and detailed laboratory safety training materials to the trainees with different roles and job responsibilities. At the same time, by analyzing the types of questions that are easy to make mistakes and the error rates of answers for different trainees, it pushes a set of link addresses of learning content that needs to be strengthened to the trainees in a personalized and customized manner, thereby improving the pertinence and effectiveness of laboratory safety education.

[0044] In the embodiment of the present application, the basic information of the training personnel (laboratory managers and laboratory users) participating in laboratory safety education is input into the laboratory safety education course integrated push system. The training personnel grading and classification subsystem and the wrong question grading and classification subsystem will grade and classify the training personnel and the wrong questions according to the input personnel information respectively. The course material grading and classification push subsystem is based on the grading and classification of personnel and the grading and classification of wrong questions, combined with the graded and classified laboratory safety training material link address in the laboratory safety training material library output by the course material grading and classification subsystem to push personalized and customized course materials.

[0045] In the embodiment of the present application, the training personnel classification subsystem, the course material classification subsystem, and the wrong question classification subsystem all use deep learning technology to automatically classify the training personnel, the link addresses of the laboratory safety training materials, the types of easy-to-error questions of the training personnel, and the error rates of the answers. On this basis, the course classification push subsystem uses unsupervised learning technology to automatically push personalized sets of laboratory safety training material link addresses to different training personnel, thereby improving work efficiency. In addition, the course material classification subsystem uses blockchain technology to integrate the link addresses of laboratory safety training materials from different sources inside and outside the school, which can avoid communication bottlenecks, single points of failure, and other privacy leakage-related issues caused by the centralized network topology.

[0046] In an optional embodiment, if Figure 2 As shown, the training personnel classification subsystem includes:

[0047] Personnel Information Collection Module: This module is used to collect basic information about training personnel (laboratory managers and laboratory users). Trainers can use this module to upload their work ID, name, ID number, academy, or other basic information that can be used to uniquely identify the laboratory manager or laboratory user receiving training, either individually or in combination.

[0048] Personnel information matching module: used to search for relevant information from the full database on campus based on the basic information of the training personnel and generate a training personnel information table. The training personnel grading and classification subsystem receives the basic information of laboratory managers and laboratory users collected by the personnel information collection module, and links the full database on campus (including all on-campus data of faculty and staff, such as basic information, academic qualifications and titles, teaching and scientific research, etc.; including all on-campus data of graduate students and undergraduates, such as basic information, student status and academic studies, course majors, etc.) through the personnel information matching module to search for the basic information, teaching tasks, scientific research directions, job responsibilities and other information of teachers among laboratory managers and laboratory users, and the basic information, professional courses, scientific research directions, colleges, professional directions and other information of graduate students and undergraduates, to form the information table of teacher laboratory managers and teacher laboratory users, the information table of graduate laboratory managers and graduate laboratory users, and the information table of undergraduate laboratory managers and undergraduate laboratory users.

[0049] Digital processing module: used to extract text information from the information table of training personnel (information table of teacher laboratory managers and teacher laboratory users, information table of graduate laboratory managers and graduate laboratory users, and information table of undergraduate laboratory managers and undergraduate laboratory users) using methods such as Word2Vec, TF-IDF, BERT sentence vectors, and standardize the numerical sequence using methods such as Z-score, and then perform normalization processing and feature vector splicing to obtain the information sequence and feature vector of the training personnel (information sequence and feature vector of teacher training personnel and information sequence and feature vector of graduate training personnel, information sequence and feature vector of undergraduate training personnel).

[0050] It is also used to classify and digitally label the information of all training personnel (teachers, graduate students, undergraduates), and use the digital labels as the classification labels of the corresponding training personnel. The digital labels of all training personnel are divided into teacher category (100) and student category (200) according to their identity. The student category (200) is further divided into undergraduate (201) and graduate student (202). According to their major, they are divided into general level (1000) and professional level (2000). The professional level (2000) is further divided into biology (2001), chemistry (2002), physics (2003), etc.

[0051] AI processing module: a classifier for training the hierarchical classification model of trainees according to the information sequence, feature vector and hierarchical classification label of the trainees, and obtaining a mature model of hierarchical classification of trainees. In the embodiment of the present application, the information sequence and hierarchical classification label of all trainees (teacher laboratory managers and teacher laboratory users, graduate laboratory managers and graduate laboratory users and undergraduate laboratory managers and undergraduate laboratory users) are divided into three groups; the first group is used to train the pre-constructed hierarchical classification model of trainees, the second group is used to test the training accuracy and continuously adjust the hierarchical classification model of trainees to make the test accuracy meet the requirements, and the third group is used to actually measure the classification accuracy of the hierarchical classification model of trainees. Specifically as Figure 3 As shown in:

[0052] The first set of information sequences, feature vectors, and hierarchical classification labels are input into the classifier of the hierarchical classification model for training personnel for classification, and the parameters of the input node number, number of layers, and number of hidden layer nodes of the classifier are adjusted. Classifiers include, but are not limited to, neural networks and convolutional neural networks (CNNs). Specifically, the model is judged using indicators such as precision, accuracy, F-value, and business requirements (e.g., qualifications and experience requirements for laboratory personnel in special fields (biosafety, chemical testing, etc.)), and the parameters of the hierarchical classification model for training personnel are adjusted to determine the hierarchical classification model for training personnel.

[0053] The second set of information sequences, feature vectors and classification labels are input into the trained training personnel classification model, and the output results of the model are evaluated. The personnel classification model is evaluated through indicators such as precision, accuracy, F value, business requirements (for example, the qualification requirements and experience requirements of laboratory personnel in special fields (biosafety, chemical testing, etc.)), and the parameters of the training personnel classification model are adjusted to evaluate and correct the training personnel classification model to obtain a preliminary mature training personnel classification model.

[0054] The third set of information sequences, feature vectors and hierarchical classification labels are input into the preliminary mature model of hierarchical classification of trainees, and the output results of the hierarchical classification model of trainees are evaluated to obtain actual measurement values, and further the actual measurement accuracy of the hierarchical classification model of trainees is obtained. When the actual measurement accuracy is greater than or equal to a threshold (for example, 95%), it is considered that the training can be terminated and the final mature model of hierarchical classification of trainees is determined.

[0055] Personnel grading and classification module: used to perform grading and classification of training personnel using a mature model for grading and classification of training personnel. In the embodiment of the present application, the work number, name, ID number, college or other basic information of the training personnel (teacher laboratory managers and teacher laboratory users, graduate laboratory managers and graduate laboratory users, and undergraduate laboratory managers and undergraduate laboratory users) is input, and the information sequence of teacher laboratory managers and teacher laboratory users, graduate laboratory managers and graduate laboratory users, and undergraduate laboratory managers and undergraduate laboratory users is formed through the personnel information matching module and the digital processing module. According to the mature model of training personnel classification, the generated information sequences of laboratory managers and laboratory users are batch processed to realize the automatic classification of laboratory managers and laboratory users participating in the training. The levels and categories are: general level (1000), professional level (2000), professional level (2000) is further divided into different majors such as biology (2001), chemistry (2002), physics (2003), etc., teacher category (100), student category (200), student category (200) is further divided into undergraduate (201) and graduate student (202). It turns out that there are general-level teachers (1000100), general-level graduate students (1000202), general-level undergraduate students (1000201), professional-level teachers (2000100) are further divided into biology teachers (2001100), chemistry teachers (2002100), and physics teachers (2003100), professional-level graduate students (2000202) are further divided into biology graduate students (2001202), chemistry graduate students (2002202), and biology graduate students (2001202), professional-level undergraduate students (2000201) are further divided into biology undergraduate students (2001201), chemistry undergraduate students (2002201), and physics undergraduate students (2003201), etc.

[0056] The personnel grading and classification module can perform real-time and automatic grading and classification of newly trained teacher laboratory managers and teacher laboratory users, graduate laboratory managers and graduate laboratory users, and undergraduate laboratory managers and undergraduate laboratory users.

[0057] In an optional embodiment, the course material grading and classification subsystem includes:

[0058] The on-campus and off-campus course collection and integration module is used to respond to the query request of the query terminal, query the target laboratory safety training material link address data repository node through each DSP node of the blockchain, merge the query results of each DSP node and return them to the query terminal, and store them in the laboratory safety training material library at the same time; the target laboratory safety training material link address data repository node stores the link address of the corresponding laboratory safety training material;

[0059] The on-campus and off-campus course grading and classification module is used to grade and classify the link addresses of all laboratory safety training materials queried by the terminal device through the course material grading and classification model to obtain course grading and classification information.

[0060] In the embodiment of the present application, the link address for integrating the laboratory safety training materials in the school's internal training database and the school's external training database is provided. The school's internal training database includes general-level training materials and professional-level training materials: the general-level training materials include: the school has purchased, established, and used the relevant laws and regulations on laboratory safety and environmental protection, industry standards, management systems, laboratory equipment personal protection, key points for the use of emergency self-rescue facilities, and general-level courses such as daily laboratory safety knowledge in the knowledge base; professional-level training includes: the school has purchased, established, and used the knowledge base containing experimental operating procedures and operating guidelines in the fields of biology, chemistry, physics, safety of major equipment use, accident cases, common operating errors, full-process control guidelines for hazardous chemicals, safety management and operating specifications for special equipment and facilities, standard operating procedures for various high-risk instruments (such as high-pressure gas cylinders, precision analytical instruments, etc.), animal ethics, and professional safety and technical knowledge of subject characteristics.

[0061] The school's external (off-campus) training resource library includes a general-level training resource library and a professional-level training resource library: General-level training resources include external links to general-level laboratory safety training resources collected from platforms such as China National Science Foundation Public Knowledge Base, China University MOOC Public Knowledge Base, Wisdom Tree Public Knowledge Base, Study Strong Country Public Knowledge Base, and NetEase Open Course Public Knowledge Base, including relevant laws and regulations on laboratory safety and environmental protection, industry standards, management systems, laboratory personal protection, key points for the use of emergency self-rescue facilities, daily laboratory safety knowledge, etc.; Professional-level training resources include external links to professional-level laboratory safety training resources in biology, chemistry, physics, etc., including experimental operating procedures and operating guidelines, safety of major equipment use, accident cases, common operating errors, full-process control guidelines for hazardous chemicals, safety management and operating specifications for special equipment and facilities, standard operating procedures for various high-risk instruments (such as high-pressure gas cylinders, precision analytical instruments, etc.), animal ethics, and subject-specific professional safety and technical knowledge.

[0062] like Figure 4 As shown, according to the requirements of the school laboratory safety training materials, the user connects to the laboratory safety training materials query integration service blockchain (DSP_BC) through the query terminal QT (QueryTerminal), and queries and integrates the link address information of different laboratory safety training materials through the laboratory safety training materials link address joint query request. The link address data repository node DR of different target laboratory safety training materials k (Including A-DD, B-DD, C-DD, ...) Provide corresponding authorized laboratory safety training materials link address query service for each blockchain node (DSP node). A-DD, B-DD, C-DD, ... respectively represent the school's purchased and self-built on-campus storage knowledge bases and the Chaoxing public knowledge base, China University MOOC public knowledge base, Zhihuishu public knowledge base, Xuexi Qiangguo public knowledge base, NetEase Open Class public knowledge base and other laboratory safety training materials link address data storage node DR k The query results of all laboratory safety training material link addresses obtained by the query terminal are integrated and stored in the laboratory safety training material database.

[0063] In the embodiment of the present application, the on-campus and off-campus course grading and classification module uses a course material grading and classification model to grade and classify the link addresses of all laboratory safety training materials queried by the terminal device to obtain corresponding course grading and classification information.

[0064] In an optional embodiment, the on-campus and off-campus course collection and integration module includes:

[0065] The query request initiation module is used to query the terminal to access the blockchain through the communication interface and initiate a joint query request for the link address of the laboratory safety training materials to the blockchain.

[0066] In the query request initiation module, the query terminal (Query Terminal, QT) accesses the laboratory safety training material query integration service blockchain (Distributed Service Platform Blockchain, DSP-BC) through the communication interface. QT constructs a joint query and integration request for the link address of the laboratory safety training material (for example, querying and integrating the link addresses corresponding to the hazardous chemicals full-process control guide course materials in the knowledge base that the school has purchased, established, and used, as well as in public knowledge bases such as China National Knowledge Infrastructure, China University MOOC, Wisdom Tree, Study Strong Country, and NetEase Open Class). In this process, QT calls the smart contract pre-deployed on the DSP-BC network, which contains query format detection, permission verification, and automatic routing rules for data requests. After receiving the SQL query statement submitted by QT, the smart contract automatically performs request legitimacy verification, records the request information, and generates a corresponding event log, providing tamper-proof credentials on the chain for subsequent processing.

[0067] The data access module is used to split the laboratory safety training materials link address joint query request into laboratory safety training materials link address sub-queries for each target laboratory safety training materials link address data repository node through the splitting logic automatically triggered by the smart contract when the preset format and permission requirements of the laboratory safety training materials link address joint query request are verified and passed, and each DSP node sends the laboratory safety training materials link address sub-queries to the target laboratory safety training materials link address data repository node.

[0068] The query result acquisition module is used to perform laboratory safety training material matching and link address query operations in the target laboratory safety training material link address data repository node according to the laboratory safety training material link address sub-query when the DSP node's authority verification is passed, and format the laboratory safety training material link address sub-query result obtained by the query and send it back; among which, the authority verification of the DSP node is carried out through a smart contract.

[0069] The query result verification and integration module is used to merge the laboratory safety training materials link address sub-query results of each DSP node into a laboratory safety training materials link address joint query result, and after the laboratory safety training materials link address joint query result passes the verification of the smart contract, the laboratory safety training materials link address joint query request is returned to the query terminal.

[0070] In an optional embodiment, the data access module includes:

[0071] The query request receiving and verification unit is used to receive the joint query request for the link address of laboratory safety training materials through the blockchain DSP_BC, use each DSP node to call the verification logic built into the smart contract, and automatically determine whether the joint query request for the link address of laboratory safety training materials meets the preset format and permission requirements; only the joint query request for the link address of laboratory safety training materials that passes the smart contract verification will be further received and processed to ensure the automation and security of the system operation.

[0072] The DSP node authentication and smart contract registration unit is used to authenticate each other through the distributed identity authentication protocol, CFL authentication system and alliance chain digital certificate authentication mechanism when the joint query request of the laboratory safety training material link address meets the preset format and permission requirements. The smart contract will register the authentication information on the alliance chain; the authentication information includes the DSP node identity and the address index of the target laboratory safety training material link address data repository node (such as school, Super Star, China University MOOC, Wisdom Tree, Study Strong Country, NetEase Open Class, etc.); registering the authentication information on the alliance chain also ensures the traceability and security of cross-organizational course resource queries.

[0073] The transaction generation and smart contract template calling unit is used by the DSP node to use the private key to jointly query the laboratory safety training material link address and the data repository node DR of each related target laboratory safety training material link address. k (including A-DD, B-DD, C-DD, ...) to perform digital signatures, generate query transactions, and construct transaction orders. In the process of constructing transaction orders, the preset smart contract template is called to automatically process the query transaction format, generate the query transaction ID and contract terms (for example, data confidentiality, query validity, resource matching rules, etc.), and ensure that all subsequent operations strictly comply with the contract provisions.

[0074] The transaction propagation unit is used for other DSP nodes to receive the transaction order broadcast by the DSP node that initiated the transaction order through the P2P communication protocol, and after calling the verification logic in the smart contract to verify the content of the transaction order and the contract signature, it continues to forward the transaction order to the surrounding nodes according to the predetermined strategy until the transaction order is distributed in the blockchain.

[0075] The query request splitting and smart contract rule-driven unit is used by each DSP node to parse the joint query request for the laboratory safety training material link address in the transaction order. Based on the address index of the different target laboratory safety training material link address data repository nodes stored in the DSP node, the smart contract automatically triggers the splitting logic to split the joint query request into laboratory safety training material link address sub-queries for each target laboratory safety training material link address data repository node. During this process, the smart contract performs data format standardization, dynamic adjustment of splitting rules, and logging, providing a basis for subsequent automated matching.

[0076] Data source request authentication unit, used by each DSP node to request the target laboratory security training material link address data repository node DR k (including A-DD, B-DD, C-DD,...) sends a sub-query of the laboratory safety training material link address, and completes the sub-query of the laboratory safety training material link address after automatically verifying the legality of the authorization information and access rights of both parties through the pre-established smart contract between the DSP node and the target laboratory safety training material link address data repository node.

[0077] In an optional embodiment, the query result acquisition module includes:

[0078] Data access rights verification and automatic execution unit, used for target laboratory security training material link address data repository node DR k Based on the local authorization list (including A-DD, B-DD, C-DD, ...), the received DSP node is verified through the built-in permission control module in the smart contract. After verification, the laboratory safety training material matching and link address query operations are automatically executed. The matching and laboratory safety training material link address sub-query results are formatted and sent back to the DSP node. All operations are recorded in the contract, ensuring that the entire permission verification process is open and transparent.

[0079] Each DSP node is linked to the data repository node DR of each target laboratory safety training material link address k (including A-DD, B-DD, C-DD, ...) Get a set of laboratory safety training materials link address sub-query results;

[0080] In an optional embodiment, the query result verification and integration module includes:

[0081] The proof-of-work calculation unit is used to merge the laboratory safety training material link address sub-query results of each DSP node into a laboratory safety training material link address joint query result through a predetermined data merging algorithm in the DSP node, and package the laboratory safety training material link address joint query result, timestamp, previous block hash value and random number into a candidate block, and use the consensus mechanism to complete the consensus calculation of the candidate area.

[0082] The cross-node query result verification unit is used to verify the query return transaction containing the timestamp and the link address of the laboratory safety training materials after the DSP node broadcasts it to the entire blockchain through the smart contract. Other DSP nodes verify the query return transaction by calling the multi-layer verification logic built into the smart contract, including signature verification, hash value comparison and cross-verification of data integrity, timestamp accuracy and query result correctness, to ensure result consistency and tamper-proofness.

[0083] The result recording and blockchain storage unit is used to package verified query return transactions into legal blocks after being confirmed by the smart contract. After all DSP nodes have confirmed through the consensus mechanism, the legal blocks are appended to the blockchain. The smart contract not only records complete transaction information but also triggers subsequent business logic according to preset rules (such as automatically generating query reports, updating the laboratory safety training database, or initiating subsequent data analysis tasks).

[0084] The blockchain's continuous iteration unit is used to continuously generate new blocks according to a pre-set consensus mechanism, forming a chain-like structure. The blockchain continuously adds new blocks to form a chain-like structure according to the established consensus mechanism, while smart contracts undergo version iterations and dynamic upgrades based on application needs. Various contracts (including query verification, permission control, event triggering, and incentive distribution) can be adjusted in real time on the chain based on security, performance, and decentralization requirements to achieve the optimal balance.

[0085] The query result return unit is used to automatically package the results of the joint query of laboratory safety training material links through a smart contract and return them to the query terminal (QT). The DSP node then calls the smart contract to store all verified joint query results of laboratory safety training material links that meet the conditions in the laboratory safety training database. The smart contract mechanism ensures the automated and transparent recording of query results and the efficient execution of subsequent data traceability queries.

[0086] In an optional embodiment, the proof-of-work calculation unit includes:

[0087] The integrated computing sub-unit is used to record the timestamp, data summary and source information of each laboratory safety training material link address sub-query result through a smart contract, and use the predetermined data merging algorithm in the DSP node (such as merge sort, deduplication, data verification, etc.) to merge the laboratory safety training material link address sub-query results of each DSP node into a laboratory safety training material link address joint query result; wherein, the smart contract assists in verifying data consistency during the merging process.

[0088] Construct candidate block unit: used to package the laboratory safety training material link address joint query results, timestamp, previous block hash value and random number into a candidate block, and the smart contract adds conditional verification logic.

[0089] The consensus calculation unit is used to complete the consensus calculation of candidate blocks using a consensus mechanism. It monitors the consensus calculation process in real time through smart contracts, recording transactions and computing power contributions. The consensus mechanism can be any of Proof-of-Work (PoW), Delegated Proof-of-Stake (DPoS), Practical Byzantine Fault Tolerance (PBFT), and Proof-of-Authority (PoA). During the consensus calculation phase, using PoW as an example, DSP nodes frequently adjust random numbers to meet the preset difficulty target. Smart contracts monitor this process in real time, recording transactions and computing power contributions.

[0090] The block broadcast and incentive unit is used to broadcast candidate blocks through the blockchain. After verification by more than half of the DSP nodes based on smart contracts, the candidate blocks are added to the blockchain, and incentive calculation and distribution are automatically triggered. Digital currency rewards are distributed to the DSP nodes that generated the candidate blocks. During this process, the smart contract ensures that the distribution records are open and accurate.

[0091] For example, the present application embodiment provides an implementation method for collecting laboratory safety training materials from different sources through blockchain in an on-campus and off-campus course collection and integration module, as follows:

[0092] First define the following symbols and modules:

[0093] QT: User Terminal (Query Terminal);

[0094] DSP-BC: A blockchain-based service for querying and integrating safety training materials from different laboratories;

[0095] DSP n : Each DSP node of the DSP-BC blockchain that provides query and integration services for safety training materials in different laboratories;

[0096] DR k: Each target laboratory safety training material link address data repository node, which stores the corresponding laboratory safety training materials and the link address of the corresponding laboratory safety training materials. A-DD, B-DD, C-DD, ... represent DR k The address index of

[0097] Σ: digital signature operation;

[0098] ⊕: transaction generation and packaging operations;

[0099] POW: Proof-of-Work mechanism;

[0100] Δ: data query request decomposition operation;

[0101] The process of adding new blocks in a loop;

[0102] SC: smart contract module;

[0103] L_auth: local authorization list;

[0104] L_ally: Consortium chain, used to record authentication results and contract status;

[0105] V_req(·): request verification function;

[0106] V_perm(·): permission verification function;

[0107] DataMerge(·): data integration and merging operations;

[0108] 1. Joint query request for laboratory safety training materials link address initiated

[0109] The user initiates a joint query request for the link address of the laboratory safety training materials to the blockchain (DSP-BC) through the query terminal (QT):

[0110] QT initiated:

[0111] QT constructs the joint query request Q for the link address of the laboratory safety training materials, which is defined as follows:

[0112] Q={SQL,DR}

[0113] Among them, SQL represents the SQL statement used to search for and match the link address of laboratory safety training materials such as "Guidelines for the Whole Process Control of Hazardous Chemicals"; DR = {DR1, DR2, ..., DR k} represents the set of target laboratory safety training material link address data repository nodes (such as the school's internal knowledge base, Super Star, China University MOOC, Wisdom Tree, Study Strong Nation, NetEase Open Course, etc.). M represents the set of laboratory safety training material link addresses in the target laboratory safety training material link address data repository nodes DR, that is, M = {m|m is a link address record of laboratory safety training materials}, for each m∈M, its attribute m.title represents the title of the record. {m∈M|"Guidelines for the Full Process Control of Hazardous Chemicals" is a substring of m.title}; pre-deploy SC1 (smart contract-request processing contract) to pre-verify the request format and permissions;

[0114] process:

[0115] (1) QT→DSP-BC: Submit the encapsulated request;

[0116] (2) SC1 execution:

[0117] Automatically detect SQL format;

[0118] Verify QT identity and access rights (call V_req(Q));

[0119] Generate on-chain event logs (tamper-proof);

[0120] Symbol Description:

[0121] QT→[Q]DSP-BC

[0122] SC1: Q→{V_req(Q)=True}? Continue: Reject

[0123] The request is formatted, pre-checked for permissions, and then sent to the DSP-BC network via a secure communication protocol.

[0124] 2. Query request reception and verification

[0125] Each DSP n After receiving the encapsulated Q, the node automatically performs preliminary verification through SC1 and calls V_req(Q) to ensure that the format and permissions meet the requirements.

[0126] DSP-BC receives the request and serves as a unified entry point:

[0127]

[0128] Verify whether the content request format meets the preset standards and whether QT's identity and access rights meet the authorization requirements. Only when V_req(Q)=True, the query request enters the next processing stage.

[0129] 3. DSP node certification and smart contract registration

[0130] The certification process is divided into the following stages:

[0131]

[0132] (1) Certificate application and registration

[0133] Step 1: The DSP node submits a registration application to the certification authority (CA), providing proof of identity and qualifications;

[0134] Step 2: Generate digital certificate C after CA verification dig , which includes the node's unique identification ID, public key PK, validity period T_valid, scope of authority, and additional authentication data (such as multi-party consensus score, historical reputation, etc.);

[0135] Step 3: CA checks certificate C dig Sign and write the certificate summary and signature result into the alliance chain L_ally.

[0136] (2) Mutual authentication starts

[0137] Step 4: DSP node A establishes a communication connection with DSP node B and exchanges digital certificates C dig _a and C dig _b;

[0138] Step 5: Both parties query the other party’s certificate status on the alliance chain L_ally (verify CA signature, check revocation list, and compare additional authentication data);

[0139] Step 6: If the certificates are valid and the credit score meets the requirements, the dynamic challenge phase begins.

[0140] (3) Dynamic Challenge-Response Process

[0141] Step 7: DSP node A generates a random number R1 and sends it to B;

[0142] Step 8: DSP node B uses private key SK_b to sign R1 to generate S1, and adds timestamp T1 and returns it to A;

[0143] Step 9: DSP node A uses C dig The public key PK_b in _b verifies S1, ensuring that R1 matches S1;

[0144] Step 10: At the same time, DSP node B initiates a reverse challenge, sending a random number R2, DSP node A signs and generates S2, and DSP node B verifies S2;

[0145] Step 11: After the two-way verification is passed, the two parties establish a secure communication channel.

[0146] (4) Smart Contract SC2 Joins

[0147] Step 12: Both parties will automatically organize and package the authentication success information (such as node identification, authentication time T_auth, transaction summary Hash (T_auth|ID)) by SC2 and submit it to the alliance chain L_ally in a digitally signed manner to form a credibility certificate.

[0148] (5) Index records

[0149] Each DSP n The node also records the corresponding target laboratory safety training materials link address data repository node DR k Address Index:

[0150] {DR1,DR2,…,DR k}Address Index

[0151] That is: each DSP node maintains its corresponding target laboratory safety training materials link address data repository node DR k Address indexes (such as A-DD, B-DD, ...) ensure the traceability and security of cross-organizational query operations.

[0152] Index(DSP i )={A-DD,B-DD,…}

[0153] 4. Transaction generation and smart contract template call

[0154] After authentication, the DSP node uses the private key SK to query the laboratory safety training material link address joint query request Q and the target laboratory safety training material link address data repository node (DR k )Address index (A-DD, B-DD, ...) is digitally signed to generate query transaction Tx:

[0155]

[0156] Call the preset smart contract SC3 (transaction generation contract), perform format verification on the signed transaction Tx, generate the transaction ID, and embed the contract terms. Then, perform transaction generation and packaging operations on multiple Tx to generate T_order:

[0157] T_order=⊕{Tx1,Tx2,…,Tx k}via SC3.

[0158] 5. Transaction Propagation

[0159] The DSP node that initiates the transaction order broadcasts T_order to the entire network through the P2P network (such as the Gossip protocol). Each receiving node verifies the transaction content through SC3 and forwards it, and continues to forward T_order to surrounding DSP nodes until the entire network is covered:

[0160]

[0161] 6. Query Request Splitting

[0162] Each DSP n The node resolves the link address of the laboratory safety training materials and the joint query request Q, based on the storage index Index (DSP i ) Split Q into different target laboratory safety training materials link address data repository node DR through the split operation Δ k (such as A-DD, B-DD, ...) Laboratory safety training materials link address subquery {q1,q2,...,q k}, the smart contract SC4 (split contract) automatically formats and triggers rules according to the storage index:

[0163]

[0164] Δ(Q)={q1,q2,…,q k via SC4

[0165] M DRk = {m∈M|m is stored in DR k}

[0166] q j ={m∈M DRj "Guidelines for the Full Process Control of Hazardous Chemicals" is a substring of m.title}

[0167] 7. Data source request authentication

[0168] The DSP node sends a subquery q for the link address of the laboratory safety training materials to the target laboratory safety training materials link address data repository node DR (such as A-DD, B-DD, ...). Both parties perform identity authentication (same as the authentication process above), and call a dedicated smart contract SC5 to realize the authentication record to the alliance chain L_ally, and then verify the permissions through the local authorization list:

[0169] q j Certification via SC5

[0170] If DSP i ∈ authorization list, then DR k → Query Service

[0171] Identity mutual recognition process:

[0172] (1) Digital certificate application and registration

[0173] Step 1.1: Submit registration application

[0174] DSP nodes and DR k Nodes submit registration applications to the certification authority (CA) separately, providing their own identity information and qualification certificates.

[0175] Step 1.2: Generate Certificate

[0176] CA verifies the submitted materials and generates a digital certificate C dig ,The certificate contents include: node unique identification ID, public key PK, validity period T_valid, scope of authority, and additional authentication data (such as multi-party consensus ,score, historical reputation, etc.).

[0177] Step 1.3: Certificate Signing and Recording

[0178] CA uses its private key to create a digital certificate C dig After signing, the certificate summary and signature result will be written into the alliance chain L_ally, providing a credible basis for subsequent authentication.

[0179] (2) Mutual authentication starts

[0180] Step 2.1: Certificate Exchange

[0181] DSP nodes and DR k The nodes establish initial communication and both parties exchange their respective digital certificates C dig .

[0182] Step 2.2: Certificate Status Verification

[0183] Both parties access the consortium chain L_ally to verify the status of the other party's digital certificate. The verification content includes: whether it has a valid CA signature, whether it is on the revocation list, and whether the additional authentication data matches the preset requirements.

[0184] Step 2.3: Prepare to enter the dynamic challenge

[0185] If both parties' certificates are verified successfully, they will proceed to the next step of dynamic challenge to ensure the authenticity of their identities.

[0186] (3) Dynamic challenge-response process

[0187] Step 3.1: Initiate a challenge (DSP node → DR k node)

[0188] The DSP node generates a random number R1 and sends R1 as a challenge to the DR. k node.

[0189] Step 3.2: Response Challenge (DR k Node processing R1)

[0190] DR k The node uses its own private key SK to digitally sign the received R1 to generate signature S1. After appending the current timestamp T1, the signature S1 and T1 are returned to the DSP node.

[0191] Step 3.3: Verify Response (DSP Node Verification S1)

[0192] DSP nodes utilize DR k The public key PK in the node certificate verifies S1 to ensure that R1 matches S1.

[0193] Step 3.4: Initiate a reverse challenge (DR k Node → DSP Node)

[0194] DR k The node generates a random number R2 and sends it to the DSP node as a reverse challenge.

[0195] Step 3.5: Respond to the reverse challenge (DSP node processes R2)

[0196] The DSP node signs R2 with its own private key to generate signature S2. After adding timestamp T2, S2 and T2 are returned to DR. k node.

[0197] Step 3.6: Verify the reverse response (DR k Node Verification S2)

[0198] DR k The node verifies S2 using the public key in the DSP node certificate to ensure that R2 matches S2.

[0199] (4) Consensus record of certification results

[0200] Step 4.1: Organize authentication information

[0201] After both parties pass the dynamic challenge, they organize the authentication result information into the following key data: both parties' node identification (ID), authentication time (T_auth), challenge random number and corresponding digital signature summary (such as Hash (T_auth|ID|R1|S1) etc.).

[0202] Step 4.2: Call smart contract to record authentication

[0203] Both parties jointly invoke a dedicated smart contract, SC5, to submit the aforementioned authentication record to the consortium blockchain, L_ally. This record is digitally signed, ensuring that the authentication information cannot be tampered with and providing a basis for subsequent permission verification and auditing.

[0204] (5) Establish a secure communication channel

[0205] Step 5.1: Verify the authentication result

[0206] When SC5 successfully records the authentication results, both parties confirm that the authentication process has passed.

[0207] Step 5.2: Establish a secure channel

[0208] DSP nodes and DR k The node establishes an encrypted, trusted, and secure communication channel based on the authentication information, providing security for subsequent data transmission and laboratory safety training material link address sub-query.

[0209] (6) Local authorization verification

[0210] Step 6.1: Call the local authorization list

[0211] After establishing a secure communication channel, both parties check permissions based on their respective local authorization lists L_auth.

[0212] Step 6.2: Permission Verification

[0213] Only in local authorization verification V_perm (DSP, DR k )=True, DR k The node allows the DSP node to continue to perform subsequent data query operations.

[0214] 8. Data access rights verification

[0215] Target laboratory safety training materials link address data repository node DR k (such as A-DD, B-DD, ...), according to the local authorization list L_auth, the received DSP node request is verified. Only when V_perm (DSP i , DR k )=True, the target laboratory security training material link address data repository node executes the query operation and returns the corresponding course material link address data. At the same time, SC6 (authority contract) records the authority verification log.

[0216] V_perm(DSP,DR)→{True}via SC6

[0217] 9. Proof of Work (PoW) and Consensus Calculation

[0218] For each subquery of the link address of the laboratory safety training materials, the DSP node uses the POW mechanism (i.e. solving mathematical problems) to query the link address of the course materials:

[0219]

[0220] The results R of the sub-query courses link addresses of each laboratory safety training material link address are aggregated and analyzed to form the final query result R 联合 At the same time, DSP nodes gain the power to create new blocks and may receive digital currency rewards:

[0221]

[0222] Specifically:

[0223] Each DSP node receives data from each target laboratory security training material link address data repository node DR k Get a set of laboratory safety training materials link address sub-query link address results R k , R_total is the query result of the link addresses corresponding to all laboratory safety training materials in the final set of target laboratory safety training material data repositories (such as the school's internal knowledge base, Chaoxing, China University MOOC, WisdomTree, Xuexi Qiangguo, NetEase Open Courses, etc.):

[0224] R_total={R1,R2,…,R k}

[0225] (1) Integrated computing:

[0226] The DSP node uses a predetermined data merging algorithm (such as merge sort, deduplication, and data verification) to generate a unified query result R_final and calls the smart contract SC7 to assist in checking data consistency:

[0227] R_final=DataMerge(R_total)via SC7

[0228] (2) Construct candidate blocks:

[0229] DSP node will R_final, timestamp T, previous block hash H(B (i-1) ) and the random number Nonce constitute the candidate block B_candidate:

[0230] B_candidate={T,R_final,H(B (i-1) ),Nonce}

[0231] (3) Consensus calculation (taking PoW as an example):

[0232] The DSP node repeatedly adjusts the Nonce and calculates the hash value H(B_candidate) until the network difficulty target D is met, i.e.:

[0233] Find Nonce s.t.H(B_candidate)<Target(D)

[0234] where Target(D) is the dynamic target value corresponding to the difficulty D (e.g., requiring the first k bits of H(B_candidate) to be 0). SC7 simultaneously records the consensus calculation process in real time and monitors the contributed computing power.

[0235] 10. Block Broadcasting and Incentives

[0236] When the DSP node successfully generates a candidate block B* (satisfying H(B*)<Target(D)), the node immediately sends B* and its hash value H(B*) to the entire network through a broadcast protocol (such as the Gossip protocol). After verifying B*, all DSP nodes in the network add it to the blockchain after consensus confirmation:

[0237]

[0238] At the same time, SC8 automatically triggers the incentive calculation and distribution. The node that generates B* receives a digital currency reward Reward, and its calculation can be expressed as:

[0239] Reward=f(Difficulty,NetworkState)via SC8

[0240] 11. Verification of Cross-Node Query Results [[ID=z8]]

[0241] After completing the block consensus, the DSP node broadcasts the transaction Tx_verified containing the combined query result R_final of the timestamp T and the link address of the laboratory safety training materials to the DSP-BC network. Other DSP n nodes call SC9 (verification contract) to perform multi-level verification on Tx_verified, including: digital signature verification, hash value comparison, accuracy check of the timestamp T, and cross-verification of application-specific query logic, to ensure the consistency and correctness of the course resource query results across the network:

[0242] Complete network verification

[0243]

[0244] SC9:Validate(Tx_verified)

[0245] 12. Result Recording and Blockchain Storage

[0246] The verified Tx_verified is packaged into a legal block B i , whose structure is defined as:

[0247] B i ={T i ,R_final,H(B i-1 ),Nonce,…}

[0248] After all DSP nodes confirm through the consensus mechanism, B i Append to the blockchain data structure. At the same time, SC 10 (Storage contract) records all transaction records for easy subsequent tracing.

[0249]

[0250] The addition of blockchain not only ensures the immutability of data, but also triggers the next round of competition for new block generation.

[0251]

[0252] 13. Blockchain Continuous Iteration and Query Results Return

[0253] The blockchain system continuously generates new blocks based on a preset consensus mechanism (such as PoW, DPoS, PBFT, PoA, etc.), forming a chain structure:

[0254]

[0255] Different types of blockchains can choose different consensus algorithms and data verification rules based on actual application requirements to achieve the best balance between security, performance and decentralization.

[0256] And R 联合 →Database Storage

[0257] Finally, the laboratory safety training material link address joint query integration result R_final is returned to the query terminal QT and stored in the laboratory safety training material database, completing the link address integration process corresponding to this cross-domain course material.

[0258] R_final--[SC 11 ]-->QT&Laboratory Safety Training Database

[0259] The specific implementation process of the above-mentioned on-campus and off-campus course collection and integration module in the embodiment of this application has the following advantages through smart contract technology:

[0260] (1) Automated verification and processing

[0261] Utilize contract modules such as SC1 to SC9 to automatically complete request, transaction, and permission verification, reducing manual intervention and improving response speed.

[0262] (2) Transparent on-chain records

[0263] Each key node (authentication results, transaction generation, permission verification, consensus process and incentive distribution) is recorded by the smart contract to the alliance chain L_ally, ensuring that the entire process cannot be tampered with and cannot be traced.

[0264] (3) Contract-driven business logic

[0265] Each stage uses preset contract templates (such as transaction generation contract SC3, split contract SC4, verification contract SC9, storage contract SC 10 and result contract SC 11 ) to automatically trigger business logic and improve the operating efficiency of the overall system.

[0266] (4) Dynamic control and subsequent expansion

[0267] The SC module supports version iteration and dynamic upgrades, and can be adjusted in real time based on security, performance, decentralization and other requirements to ensure that the system always maintains the optimal balance during operation.

[0268] In the embodiment of the present application, the course material classification framework is as follows Figure 5 As shown, the laboratory safety training database includes the first and second levels. The first level is the school's general training database for all teachers and students, and the second level is the school's professional training database for professional teachers and professional students. Through the first and second levels, the hierarchical classification of course materials for all staff, the entire process, and all aspects is achieved. Among them, the school's general training database includes links to general laboratory safety training materials, which are used to impart basic safety knowledge and cultivate basic safety skills. It contains links to materials on basic laboratory safety and application. Specifically, it includes links to general laboratory safety training materials collected from knowledge bases purchased, established, and used by the school, as well as from external platforms such as the Super Star Public Knowledge Base, China University MOOC Public Knowledge Base, Wisdom Tree Public Knowledge Base, Study Strong Country Public Knowledge Base, and NetEase Open Class Public Knowledge Base, including relevant laws and regulations on laboratory safety and environmental protection, industry standards, management systems, laboratory personal protection equipment, key points for the use of emergency self-rescue facilities, and daily laboratory safety knowledge.

[0269] The school's professional training database includes links to professional laboratory safety training materials, which are used to impart professional safety knowledge and cultivate professional safety skills. It contains materials on basic laboratory safety and application, as well as materials at various professional levels (such as biology, chemistry, physics, etc.). Specifically, it includes links to professional laboratory safety training materials in the fields of biology, chemistry, physics, etc., collected from the school's purchased, established, and used knowledge bases, as well as from external platforms such as the Super Star Public Knowledge Base, China University MOOC Public Knowledge Base, Wisdom Tree Public Knowledge Base, Study Strong Nation Public Knowledge Base, and NetEase Open Class Public Knowledge Base, including documents such as experimental operating procedures and operating guidelines, safety of major equipment use, accident cases, common operating errors, full-process control guidelines for hazardous chemicals, safety management and operating specifications for special equipment and facilities, standard operating procedures for various high-risk instruments (such as high-pressure gas cylinders, precision analytical instruments, etc.), animal ethics, and subject-specific professional safety technical knowledge.

[0270] Mark the link addresses of the general and professional-level laboratory safety training materials corresponding to teachers, graduate students, and undergraduates. Teachers and students are divided into undergraduates and graduate students, general level and professional level. General level is divided into general compulsory and general elective. Professional level is divided into professional compulsory and professional elective. Professional level is divided into biology, chemistry, physics, ... and other different fields. Biology is divided into biology compulsory and biology elective. Chemistry is divided into chemistry compulsory and chemistry elective. Physics is divided into physics compulsory and physics elective, ... and so on.

[0271] In an optional embodiment, the on-campus and off-campus course grading and classification module includes:

[0272] Digital processing module: used to digitally encode the link addresses of all laboratory safety training materials (general and professional training materials for teachers, graduate students, and undergraduates) queried by the terminal device, and to classify and mark the link addresses of all laboratory safety training materials using a digital hierarchical coding method to obtain the digital codes of the link addresses of the laboratory safety training materials and the digital sequences and digital marks of the hierarchical classification of the link addresses of the laboratory safety training materials.

[0273] Specifically, the link addresses of all laboratory safety training materials are coded in the following manner to obtain the digital codes of the corresponding laboratory safety training material link addresses:

[0274] For example:

[0275] Knowledge base source code: the school has purchased, established, and used the knowledge base (10), the Chaoxing public knowledge base outside the school (11), the China University MOOC public knowledge base outside the school (12), the WisdomTree public knowledge base outside the school (13), the Xuexi Qiangguo public knowledge base outside the school (14), and the NetEase Open Course public knowledge base outside the school (15);

[0276] Laboratory safety training material type codes: relevant laws and regulations on laboratory safety and environmental protection (100), industry standards (101), management system (102), laboratory equipment personal protection (103), key points for the use of emergency self-rescue facilities (104), daily laboratory safety knowledge (105), experimental operation procedures and operation guides (106), safety of major equipment use (107), accident cases (108), common operating errors (109), protective equipment use guides (110), full-process control guides for hazardous chemicals (111), safety management and operation specifications for special equipment and facilities (112), standard operating procedures for various high-risk instruments (such as high-pressure gas cylinders, precision analytical instruments, etc.) (113), animal ethics (114), subject-specific professional safety and technical knowledge documents (115);

[0277] The coding rule for the link address of the laboratory safety training materials course is: knowledge base source code + laboratory safety training materials type code + N (N is an integer, N>=10) digit serial number.

[0278] The digital codes of the laboratory safety training materials link addresses were Z-score standardized and normalized to obtain the feature vectors.

[0279] Specifically, all the links to the laboratory safety training materials are classified and categorized according to the following method to obtain the corresponding classification labels for the links to the laboratory safety training materials:

[0280] Teacher category (100), student category (200), students (200) are further divided into undergraduates (201), postgraduates (202), general education level is marked as (1000), general education level is further divided into general education compulsory (1001), general education elective (1002), professional level is marked as (2000), professional level is further divided into professional compulsory (2100), professional elective (2200), professional level (2000) is further divided into biology (2001), chemistry (2002), physics (2003), ..., and other different fields, biology is divided into biology compulsory (2101), biology elective (2201), chemistry is divided into chemistry compulsory (2102), chemistry elective (2202), physics is divided into physics compulsory (2103), physics elective (2203), ..., and so on.

[0281] General-level teacher courses (1001000), professional-level teacher courses (1002000), professional-level teacher courses are divided into different fields such as biology (1002001), chemistry (1002002), physics (1002003)...

[0282] General graduate courses (2021000), professional graduate courses (2022000), professional graduate courses are divided into different fields such as biology (2022001), chemistry (2022002), physics (2022003)...

[0283] There are general undergraduate courses (2011000) and professional undergraduate courses (2012000). Professional undergraduate courses are further divided into different fields such as biology (2012001), chemistry (2012002), physics (2012003)...

[0284] AI processing module: used to train the course material classification model based on the digitally encoded feature vectors and hierarchical classification labels of the laboratory safety training material link addresses, and obtain a mature model of course material classification. In the embodiment of the present application, the feature vectors and hierarchical classification labels of the digitally encoded link addresses of all laboratory safety training materials are divided into three groups; the first group is used to train the pre-constructed course material classification model, the second group is used to test the training accuracy and continuously adjust the course material classification model to meet the test accuracy requirements, and the third group is used to actually measure the classification accuracy of the course material classification model. Specifically, Figure 6 As shown:

[0285] 1) Input the first set of feature vectors and classification labels into the classifier of the course material classification model for classification, and adjust the classifier's input node number, number of layers, and number of hidden layer nodes. The classifier includes neural networks, CNNs, etc. Specifically, the course material classification model is judged using indicators such as precision, accuracy, F-value, and business requirements (classification standards for on-campus and off-campus course materials, general education level, professional level), and the parameters of the course material classification model are adjusted to determine the course material classification model.

[0286] 2) Input the second set of feature vectors and classification labels into the trained course material classification model, evaluate the output of the model, and judge the model using indicators such as precision, accuracy, F-score, and business requirements (on-campus and off-campus course material classification standards, general education level, professional level), adjust the parameters of the model, evaluate, and modify the model to obtain a preliminary mature model for course material classification.

[0287] 3) Inputting the third set of feature vectors and classification labels into the preliminary course material classification maturity model, evaluating the model's output to obtain actual measurement values, and further determining the model's actual measurement accuracy. When the actual measurement accuracy is greater than or equal to a threshold (e.g., 95%), training is terminated, and the final course material classification maturity model is determined.

[0288] 4) Repeat steps 2) and 3) until the conditions are met, and obtain a mature model for grading and classifying course materials.

[0289] The on- and off-campus course grading and classification module is used to automatically categorize the links corresponding to all laboratory safety training materials based on the mature model for grading and classifying course materials. When the links corresponding to newly stored laboratory safety training materials are obtained through the blockchain, the links corresponding to laboratory safety training materials processed by the digital processing module can be automatically graded and classified.

[0290] In an optional embodiment, if Figure 7 As shown, the wrong question grading and classification subsystem includes:

[0291] Personnel Information Collection Module: used to collect basic information of training personnel (laboratory managers and laboratory users). Through the personnel information collection module, trainers can upload their work number, name, ID number, college, or other basic information that can be used to uniquely identify the laboratory managers and laboratory users receiving training, either individually or in combination.

[0292] Wrong exercise collection module: used to collect information such as titles and types of wrong exercises of different trainees through the system database;

[0293] Incorrect question type and error rate statistics module: used to count the types of easy-to-make mistakes and error rates of questions answered by different trainees through the system;

[0294] Digital processing module: used to extract data that uniquely identifies the trainees' identity information, such as their work number, student number, and ID number, and perform standardization and normalization processing on them; digitize the types of questions that are easy to make mistakes for the trainees (including digital layered coding methods, etc.); perform Z-score standardization and normalization processing; standardize the error rate of answers (including Z-score standardization) and perform minimum and maximum normalization processing; obtain a digital sequence of the trainees' basic information, the types of questions that are easy to make mistakes, and the error rate of answers; and then further splice them into a feature vector and mark them.

[0295] Specifically, the types of easy-to-error questions are numerically labeled and classified in the following way to obtain the corresponding numerical sequences, feature vectors, and classification labels:

[0296] The easy-to-make mistakes of the trainees are first classified into general knowledge (1000) and professional knowledge (2000). The professional knowledge is then classified into biology (1002001), chemistry (1002002), physics (1002003), etc. Each professional knowledge classification is further classified in turn. For example, biology is divided into zoology (1002001001), botany (1002001002), microbiology (1002001003), cell biology (1002001004), etc., and chemistry (1002002) is divided into inorganic chemistry (1002001005). 02001), Organic Chemistry (1002002002), Analytical Chemistry (1002002003), Polymer Chemistry (1002002004), etc. Physics (1002003) is divided into Mechanics (1002003001), Electromagnetism (1002003002), Optics (1002003003), Heat and Thermodynamics (1002003004), Atomic and Molecular Physics (1002003005), etc., and then Z-score standardization and normalization are performed to obtain the numerical sequence of the types of questions that are easy to make mistakes for the trainees.

[0297] The error rates of easy-to-fall-through questions are Z-score standardized (StandardScaler) or normalized (MinMaxScaler) to eliminate the dimensionality effect and obtain a numerical sequence of error rates of easy-to-fall-through questions;

[0298] The basic information of the trainees, the numerical sequence of the types of easy-to-make mistakes, and the numerical sequence of the error rate of the answers are concatenated to obtain the feature vector.

[0299] AI processing module: used to train the wrong question grading and classification model based on the basic information digital sequence of the trainees, the digital sequence of the types of easy-to-make mistakes, the digital sequence of the error rate of the answers, the characteristic vectors and the graded classification labels, and obtain a mature model of wrong question grading and classification. In the embodiment of the present application, the digital sequences and graded classification labels of all the types of easy-to-make mistakes and the error rate of the answers are divided into three groups; the first group is used to train the pre-constructed wrong question grading and classification model, the second group is used to test the training accuracy and continuously adjust the wrong question grading and classification model to make the test accuracy meet the requirements, and the third group is used to actually measure the classification accuracy of the wrong question grading and classification model. Specifically, Figure 8 As shown:

[0300] 1) Input the basic information sequence, error-prone question type sequence, error rate sequence, feature vector, and classification label of the first group of trainees into the classifier of the error-prone question classification model for classification, and adjust the classifier's input node number, number of layers, and number of hidden layer nodes. The classifier includes neural networks, CNNs, etc. Specifically, the error-prone question classification model is judged using indicators such as precision rate, accuracy rate, F-value, and business requirements (error-prone question classification standards, general knowledge level, professional level), and the parameters of the error-prone question classification model are adjusted to determine the error-prone question classification model.

[0301] 2) Input the second group of trainees' basic information, the types of questions prone to error, the error rate, the feature vectors, and the classification labels into the trained error question classification model. Evaluate the output of the error question classification model using indicators such as precision, accuracy, F-value, and business requirements (error question classification standards, general knowledge level, professional level), and adjust the parameters of the error question classification model to evaluate and correct the model, thus obtaining a preliminary mature error question classification model.

[0302] 3) Input the basic information numeric sequence, the numeric sequence of the types of questions prone to error, the numeric sequence of the error rate of the questions, the feature vectors, and the classification labels of the third group of trainees into the preliminary mature model for grading and classifying wrong questions. Evaluate the output of the model to obtain the actual measurement value, and further determine the actual measurement accuracy of the model. When the actual measurement accuracy is greater than or equal to a threshold (e.g., 95%), it is considered that the training can be terminated, and the final mature model for grading and classifying wrong questions is determined.

[0303] 4) Repeat steps 2) and 3) until the conditions are met, and obtain a mature model for grading and classifying wrong questions.

[0304] The Error Question Grading and Classification Module is used to automatically categorize the types of errors and error rates of all digitized and standardized trainees based on a mature model for error question grading and classification. When new trainee information is received, the Error Question Grading and Classification Module automatically categorizes the types of errors and error rates of the trainees processed by the Digital Processing Module by matching the information with the trainee's information.

[0305] In an optional embodiment, the course grading and classification push subsystem includes:

[0306] The course grading and classification push module is used to input the trainer grading and classification information of the trainees, the wrong question grading and classification information, and the course grading and classification information corresponding to the link addresses of all laboratory safety training materials into the course material grading and classification push model, and predict the link address set corresponding to the compulsory laboratory safety training materials and the link address set corresponding to the optional laboratory safety training materials for the trainees.

[0307] In an optional embodiment, the course material classification and push model is used to push a set of link addresses corresponding to all required laboratory safety training materials matched in the laboratory safety training material library to each trainee based on the trainee classification information of each trainee;

[0308] The course material grading and classification push model is also used to push a set of link addresses corresponding to all matching elective laboratory safety training materials in the laboratory safety training material library to each trainee based on the grading and classification information of the wrong questions corresponding to each trainee.

[0309] In an optional embodiment, if Figure 9 As shown, the course classification and push module includes:

[0310] Trainer information acquisition module: used to obtain the information sequence, feature vector and classification labels of trainees;

[0311] Course material information acquisition module: used to obtain the digital sequence, feature vector and classification label of the link address corresponding to the general and professional level laboratory safety training materials in the laboratory safety training database;

[0312] Error question information acquisition module: used to obtain the types of easy-to-make mistakes and the numerical sequence, feature vector and classification labels of the error rates of different trainees during exercise training and assessment;

[0313] Digital processing module: used to perform deduplication, missing value filling, outlier detection and numerical normalization on the data obtained by the training personnel information acquisition module, course material information acquisition module and incorrect exercise information acquisition module (including the information sequence, feature vector and hierarchical classification label of the training personnel, the digital sequence, feature vector and hierarchical classification label of the link address of the general and professional laboratory safety training materials in the laboratory safety training database, the types of easy-to-make mistakes generated by different trainees during exercise training and assessment, the digital sequence, feature vector and hierarchical classification label of the error rate of the answer) to form a pre-processed feature vector set for course push. If the feature dimension is too high, PCA or Autoencoder can be used for dimensionality reduction to compress the redundant features into a principal component space of 20-50 dimensions;

[0314] AI processing module: used to divide the course push preprocessing feature vector sets of all trainers into three groups, and perform cluster analysis on the course push preprocessing feature vector sets through unsupervised learning methods based on the three groups of course push preprocessing feature vector sets (including the information sequence and feature vector of the trainers, the digital sequence and feature vector of the link address corresponding to the general and professional level training materials in the laboratory safety training database, the type of easy-to-make mistakes generated during exercise training and assessment, and the digital sequence and feature vector of the error rate of answering questions), train the course material grading and classification push model, and obtain the mature model of course material grading and classification push. Among them, the first group is used to train the course material grading and classification push model, the second group is used to test the training accuracy of the course material grading and classification push model and continuously adjust it to make the test accuracy meet the requirements, and the third group is used to measure the accuracy of the mature course material grading and classification push model. Specifically, Figure 10 As shown:

[0315] The first set of course push preprocessing feature vector sets are input into the unsupervised learning model for course material classification and push for clustering. The unsupervised learning models include K-Means, K-Medoids, DBSCAN, hierarchical clustering, Gaussian mixture model (GMM), etc. The key hyperparameters (such as K value of K-Means, ε / MinPts of DBSCAN, and split threshold of hierarchical clustering) are adjusted, and the validation set is used to calculate the Silhouette Coefficient, DB Index (Davies-Bouldin Index), Dunn Index, Cluster Stability and other indicators. The Silhouette Coefficient is preferred. The model with the highest Dunn index (Coefficient) and the lowest Davies–Bouldin index (DBI) is selected. If multiple approximate solutions appear, their Dunn index (the larger the better) and clustering stability metrics (such as the maximum average Jaccard similarity in 100 self-service samplings) are compared. Ultimately, the algorithm and hyperparameter combination with the highest intra-cluster compactness, inter-cluster separability, and result reproducibility are identified for final tuning. Finally, the course material grading and classification push model is judged using indicators such as precision, accuracy, recall, and F-value. The parameters of the course material grading and classification push model are adjusted to determine the course material grading and classification push model.

[0316] The second set of course push preprocessing feature vector sets is input into the trained course material grading and classification push model, and the output results of the course material grading and classification push model are evaluated. The model is judged by using indicators such as precision, accuracy, recall, F value, and business needs (1. According to the grading and classification results of the trainees, the link address set corresponding to the compulsory laboratory safety training materials matching the level and type of the trainees in the school-level general and professional training material database is pushed to them; 2. According to the different types of easy-to-make mistakes and the error rate of the trainees, the link address set corresponding to the optional laboratory safety training materials matching the easy-to-make mistakes of the trainees in the school-level general and professional training material database is pushed to them for targeted consolidation; 3. Special fields (biosafety, chemical testing, etc.) require laboratory personnel's qualifications and experience requirements, and the link address set corresponding to the optional laboratory safety training materials matching the special fields in the school-level general and professional training material database is pushed to them for targeted consolidation). The parameters of the course material grading and classification push model are adjusted to evaluate and correct the model, and a preliminary mature model for course material grading and classification push is obtained.

[0317] The third set of course push preprocessing feature vector sets is input into the preliminary course material grading and classification push mature model, and the output results of the model are evaluated to obtain the actual measurement value, and the actual measurement accuracy of the model is further obtained. When the actual measurement accuracy is greater than or equal to the threshold (for example, 95%), it is considered that the training can be terminated, and the final course material grading and classification push mature model is determined to obtain the course material grading and classification push mature model.

[0318] Course classification and push module: used to push the matching set of link addresses corresponding to the required laboratory safety training materials and the set of link addresses corresponding to the optional laboratory safety training materials to the trainers according to the mature model of course material classification and push.

[0319] Specifically, the identity information sequence and label of the laboratory managers and laboratory users participating in the training, the digital sequence and label of the training materials in the laboratory safety training database, the digital sequence and label of the types of easy-to-make mistakes and the error rates of the answers for different types of personnel are input, and through the mature model of course material classification and push, the link address set corresponding to the laboratory safety training materials is automatically pushed in real time and in batches to the laboratory managers and laboratory users participating in the laboratory safety education and training, the newly joined teacher laboratory managers and teacher laboratory users, the graduate student laboratory managers and graduate student laboratory users, and the undergraduate student laboratory managers and undergraduate laboratory users. Although the link address set corresponding to the laboratory safety training materials is pushed based on the personnel classification results of the managers and users participating in the training, due to the different classification results of the easy-to-make mistakes and the error rates of the answers for different personnel, even if the system is for personnel of the same level and type (for example, all the participants in the training are biology teachers (2001100)), the link address set corresponding to all laboratory safety training materials matched with each training personnel will be different, and each has its own focus (pushing the required and the required courses within the scope of the general education teacher courses (1001000)). While the required laboratory safety training materials for professional-level teachers in biology (1002001) are consistent, the grading and classification results for easily-error-prone questions and error rates may vary among different trainees. Based on the different easily-error-prone questions and error rates, as well as the qualifications and experience requirements in specific fields, the system will push different sets of link addresses for elective courses in the general education teacher course (1001000) and elective laboratory safety training materials for professional-level teachers in biology (1002001). This allows for personalized and customized push of laboratory safety training materials tailored to the characteristics of different trainees.

[0320] The corresponding relationship between pushed courses and training participants is as follows:

[0321] General Education Teacher Courses (1000100): Pushes a collection of link addresses for laboratory safety training materials for required courses within the General Education Teacher Courses (1001000) and some elective courses within the General Education Teacher Courses (1001000);

[0322] General Graduate Courses (1000202): Push links to laboratory safety training materials for required courses within the General Graduate Courses (2021000) and some elective courses within the General Graduate Courses (2021000).

[0323] General Education Undergraduate Courses (1000201): Pushes links to laboratory safety training materials for required courses within the General Education Undergraduate Courses (2011000) and some elective courses within the General Education Undergraduate Courses (2011000).

[0324] Professional-level teachers (2000100) are further divided into biology teachers (2001100), chemistry teachers (2002100), and physics teachers (2003100). The push scope is as follows:

[0325] Biology Teachers (2001100): Push a collection of links to laboratory safety training materials for required courses within the General Education Teachers Courses (1001000) and some elective courses within the General Education Teachers Courses (1001000). Also push a collection of links to laboratory safety training materials for required courses within the Teacher Biology Courses (1002001) within the Professional Education Teachers Courses (1002000) and some elective courses within the Teacher Biology Courses (1002001).

[0326] Chemistry Teachers (2002100): Push a collection of link addresses corresponding to laboratory safety training materials for the required courses within the General Education Teachers Courses (1001000) and some elective courses within the General Education Teachers Courses (1001000), as well as a collection of link addresses corresponding to laboratory safety training materials for the required courses within the Teacher Chemistry Course (1002002) within the Professional Education Teachers Courses (1002000) and some elective courses within the Teacher Chemistry Course (1002002).

[0327] Physics Teachers (2003100): Push a collection of link addresses corresponding to laboratory safety training materials for required courses within the General Education Teachers Courses (1001000) and some elective courses within the General Education Teachers Courses (1001000), as well as a collection of link addresses corresponding to laboratory safety training materials for required courses within the Teacher Physics Courses (1002003) within the Professional Education Teachers Courses (1002000) and some elective courses within the Teacher Physics Courses (1002003).

[0328] …

[0329] And so on;

[0330] Professional graduate students (2000202) are further divided into biology graduate students (2001202), chemistry graduate students (2002202), and biology graduate students (2001202). The push scope is as follows:

[0331] Biology graduate students (2001-202): Push a collection of link addresses corresponding to laboratory safety training materials for required courses within the scope of general graduate courses (2021000) and some elective courses within the scope of general graduate courses (2021000), as well as a collection of link addresses corresponding to laboratory safety training materials for required courses within the scope of graduate biology (2022001) within the scope of professional graduate courses (2022000) and some elective courses within the scope of graduate biology (2022001);

[0332] Graduate students in Chemistry (2002202): Push a collection of links to laboratory safety training materials for required courses within the General Education Graduate Courses (2021000) and some elective courses within the General Education Graduate Courses (2021000). Also, push a collection of links to laboratory safety training materials for required courses within the Graduate Chemistry (2022002) and some elective courses within the Graduate Chemistry (2022002) within the Professional Graduate Courses (2022000).

[0333] Physics graduate students (2001-202): Push a collection of links to laboratory safety training materials for required courses within the General Education Graduate Courses (2021000) and some elective courses within the General Education Graduate Courses (2021000). Also push a collection of links to laboratory safety training materials for required courses within the Graduate Physics (2022003) and some elective courses within the Graduate Physics (2022003) within the Professional Graduate Courses (2022000).

[0334] …

[0335] And so on;

[0336] Professional undergraduate students (2000201) are further divided into biology undergraduate students (2001201), chemistry undergraduate students (2002201), and physics undergraduate students (2003201). The push scope is as follows:

[0337] Biology undergraduates (2001-201): Push links to laboratory safety training materials for required courses within the General Education Undergraduate Courses (2011-000) and some elective courses within the General Education Undergraduate Courses (2011-000). Also push links to laboratory safety training materials for required courses within the Undergraduate Biology Courses (2012-001) and some elective courses within the Undergraduate Biology Courses (2012-001) within the Professional Undergraduate Courses (2012-000).

[0338] Undergraduate Chemistry (2002201): Push links to laboratory safety training materials for required courses within the General Education Undergraduate Courses (2011000) and some elective courses within the General Education Undergraduate Courses (2011000). Also, push links to laboratory safety training materials for required courses within the Undergraduate Chemistry (2012002) and some elective courses within the Undergraduate Chemistry (2012002) course within the Professional Undergraduate Courses (2012000).

[0339] Physics undergraduates (2003201): Push links to laboratory safety training materials for required courses within the General Education undergraduate curriculum (2011000) and some elective courses within the General Education undergraduate curriculum (2011000). Also push links to laboratory safety training materials for required courses within the Undergraduate Physics (2012003) and some elective courses within the Undergraduate Physics (2012003) within the Professional Undergraduate curriculum (2012000).

[0340] …

[0341] And so on.

[0342] Based on the above, the laboratory safety education course integration and push system proposed in the embodiment of the present application includes a training personnel grading and classification subsystem, a course material grading and classification subsystem, a wrong question grading and classification subsystem, and a course material grading and classification push subsystem. The training personnel grading and classification subsystem, the course material grading and classification subsystem, the wrong question grading and classification subsystem, and the course material grading and classification push subsystem all include AI processing modules to respectively construct a training personnel grading and classification maturity model, a course material grading and classification maturity model, a wrong question grading and classification maturity model, and a course material grading and classification push maturity model. Among them, the course material grading and classification subsystem uses blockchain technology to integrate resources such as the link address collection corresponding to the general education and professional education courses in different safety education resource databases inside and outside the school. After completing the training of mature models for grading and classification of training personnel, grading and classification of course materials, grading and classification of wrong questions, and grading and classification push of course materials, unsupervised learning technology can be used to push a set of link addresses corresponding to laboratory safety training materials according to the personnel level and category, type of easy-to-make mistakes, error rate of answers, and actual professional direction of the training personnel (laboratory managers and laboratory users participating in laboratory safety education and training). This reduces the manual workload and human errors that laboratory safety training management units need to perform in organizing training personnel materials, organizing training course materials, grading and classifying laboratory managers and laboratory users participating in training, and the manual workload of matching link addresses corresponding to different laboratory safety training materials for laboratory managers and laboratory users participating in training. The personnel information that originally required manual querying of a large number of system databases or multiple tables to match is now automatically completed. The external general-level laboratory safety training materials that originally required manual querying of a large number of purchased, established, and used knowledge bases, as well as the Super Star Public Knowledge Base, China University MOOC Public Knowledge Base, Wisdom Tree Public Knowledge Base, Study Strong Nation Public Knowledge Base, NetEase Open Class Public Knowledge Base, etc., are automatically associated and searched and integrated to form a hierarchical and classified laboratory safety training material database. Then, the training personnel are automatically classified according to the obtained information, and the classified training personnel are associated with the hierarchical and classified laboratory safety training material database and the hierarchical classification information of different error-prone question types and error rates of different training personnel. The corresponding link address set of the corresponding laboratory safety training materials is automatically pushed to each training personnel. While ensuring that the required laboratory safety training materials can be accurately pushed to each participant, it can also push targeted, personalized, and customized link address sets of optional laboratory safety training materials to different participants based on their error rates and types of wrong questions, thus achieving the goal of systematic, refined, professional, and personalized safety education and training, and solving the problems brought about by the traditional "one-size-fits-all" training model of laboratory safety education.At the same time, the system has achieved full automation of the grading and classification of training participants and the recommendation of training courses. It can grade and classify newly-joined teacher laboratory managers and teacher laboratory users, graduate laboratory managers and graduate laboratory users, and undergraduate laboratory managers and undergraduate laboratory users in real time, automatically, and in batches. It can also automatically grade and classify newly-entered laboratory safety training materials obtained through the blockchain. Then, based on the error rates of different trainees, the grading and classification results of wrong question types, and the qualifications and experience requirements in special fields, the system pushes targeted, personalized, and customized link address sets corresponding to graded and classified laboratory safety training materials to the trainees who have been graded and classified. This reduces the workload of laboratory safety training management units in organizing training and allocating training courses each time, and improves work efficiency.

[0343] Based on the same principle as the method provided in the embodiment of the present application, the embodiment of the present application also provides a method for integrating and pushing laboratory safety education courses, such as Figure 11 As shown, the method includes the following steps:

[0344] Step 1101: Use deep learning technology to classify the trainees and obtain classification information of the trainees;

[0345] Step 1102: Using blockchain technology, link addresses corresponding to different laboratory safety training materials are integrated and stored in a laboratory safety training database. Deep learning technology is used to classify the link addresses of the laboratory safety training materials in the laboratory safety training database to obtain course classification information.

[0346] Step 1103: Count the types of questions that are easy to make mistakes and the error rates of the questions answered by the trainees, and use deep learning technology to classify the types of questions that are easy to make mistakes and the error rates of the questions answered by each trainee, so as to obtain the classification information of the wrong questions answered by each trainee;

[0347] Step 1104, using the unsupervised learning algorithm in the machine learning technology to process the training personnel classification information, course classification information and wrong question classification information, obtain the link address set corresponding to all laboratory safety training materials matching each training personnel, and push it to the corresponding training personnel.

[0348] In the present application, deep learning technology is used to automatically classify the training personnel, laboratory safety training materials, and the types of easy-to-error questions and error rates of the training personnel. On this basis, an unsupervised learning algorithm is used to automatically push the link address set corresponding to the personalized laboratory safety training materials to different training personnel, thereby improving work efficiency. In addition, the use of blockchain technology to integrate the link address set corresponding to laboratory safety training materials from different sources inside and outside the school can avoid communication bottlenecks, single points of failure, and other privacy leakage related issues caused by the centralized network topology.

[0349] The laboratory safety education course integrated push method provided in the embodiment of the present application can achieve Figures 1 to 10 To avoid repetition, the various processes implemented in the system embodiment will not be described here.

[0350] The laboratory safety education course integration and push method of the embodiment of the present application can realize the laboratory safety education course integration and push system provided by the embodiment of the present application. The implementation principle is similar. The steps in the laboratory safety education course integration and push method in each embodiment of the present application correspond to the actions performed by each module and unit in the laboratory safety education course integration and push system in each embodiment of the present application. For the detailed functional description of each step of the laboratory safety education course integration and push method, please refer to the description in the corresponding laboratory safety education course integration and push system shown in the previous text, which will not be repeated here.

[0351] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the above-mentioned disclosed concepts. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A laboratory safety education course integrated push system, characterized by: The system comprises: The training personnel grading and classification subsystem is used to grade and classify training personnel using deep learning technology to obtain training personnel grading and classification information; The course material grading and classification subsystem is used to integrate the link addresses corresponding to different laboratory safety training materials into the laboratory safety training material database through blockchain technology, and use deep learning technology to grade and classify the link addresses of laboratory safety training materials in the laboratory safety training material database to obtain course grading and classification information; The wrong question grading and classification subsystem is used to count the types of questions that trainees are prone to making mistakes and the error rates of their answers. Deep learning technology is used to grade and classify the types of questions that are prone to making mistakes and the error rates of each trainee, and to obtain the graded and classified information of each trainee's wrong questions. The course grading and classification push subsystem is used to use unsupervised learning technology to process the training personnel grading and classification information, the course grading and classification information, and the wrong question grading and classification information, obtain the link address set corresponding to all laboratory safety training materials matching each trainer, and push them to the corresponding trainer.

2. The laboratory safety education course integrated push system according to claim 1 is characterized in that: The course material grading and classification subsystem includes: The on-campus and off-campus course collection and integration module is used to respond to the query request of the query terminal, query the target laboratory safety training material link address data storage node through each DSP node of the blockchain, merge the query results of each DSP node and return them to the query terminal, and store them in the laboratory safety training material library at the same time; the target laboratory safety training material link address data storage node stores the link address of the corresponding laboratory safety training material; The on-campus and off-campus course grading and classification module is used to grade and classify the link addresses of all laboratory safety training materials queried by the terminal device through the course material grading and classification model to obtain the course grading and classification information.

3. The laboratory safety education course integrated push system according to claim 2 is characterized in that: The on-campus and off-campus course collection and integration module includes: A query request initiating module, configured to query a terminal to access the blockchain through a communication interface and initiate a joint query request for a link address of laboratory safety training materials to the blockchain; A data access module is configured to, when a preset format and permission requirements of the laboratory safety training material link address joint query request are verified to be passed, split the laboratory safety training material link address joint query request into laboratory safety training material link address sub-queries for each target laboratory safety training material link address data repository node through a splitting logic automatically triggered by a smart contract, and each DSP node sends the laboratory safety training material link address sub-query to the target laboratory safety training material link address data repository node; A query result acquisition module is configured to, if the authority verification of the DSP node passes, perform a laboratory safety training material matching and link address query operation on the target laboratory safety training material link address data repository node according to the laboratory safety training material link address subquery, and format the laboratory safety training material link address subquery result obtained by the query and send it back; wherein the authority verification of the DSP node is performed through a smart contract; The query result verification and integration module is used to merge the laboratory safety training material link address sub-query results of each DSP node into a laboratory safety training material link address joint query result, and after the laboratory safety training material link address joint query result passes the verification of the smart contract, the laboratory safety training material link address joint query request is returned to the query terminal.

4. The laboratory safety education course integrated push system according to claim 3 is characterized in that: The data access module includes: A query request receiving and verification unit is used to receive the joint query request for the link address of the laboratory safety training materials through the blockchain, use each DSP node to call the verification logic built into the smart contract, and automatically determine whether the joint query request for the link address of the laboratory safety training materials meets the preset format and permission requirements; The DSP node authentication and smart contract registration unit is used to, when the joint query request for the laboratory safety training material link address meets the preset format and permission requirements, all DSP nodes authenticate each other through the distributed identity authentication protocol, CFL authentication system and alliance chain digital certificate authentication mechanism, and the smart contract registers the authentication information on the alliance chain; the authentication information includes the DSP node identity and the address index of the target laboratory safety training material link address data repository node; The transaction generation and smart contract template calling unit is used for the DSP node to use the private key to digitally sign the joint query request of the laboratory safety training material link address and the data storage node of each related target laboratory safety training material library link address, generate a query transaction, and construct a transaction order; wherein, in the process of constructing the transaction order, the preset smart contract template is called to automatically process the query transaction format, generate the query transaction ID and contract terms; The transaction propagation unit is used for other DSP nodes to receive the transaction order broadcasted by the DSP node that initiated the transaction order through the P2P communication protocol, and after calling the verification logic in the smart contract to verify the content of the transaction order and the contract signature, to continue forwarding the transaction order to surrounding nodes according to a predetermined strategy until the transaction order is completely distributed in the blockchain; The query request splitting and smart contract rule driving unit is used for each DSP node to parse the laboratory safety training material link address joint query request in the transaction order, and use the smart contract to automatically trigger the splitting logic according to the address index of different target laboratory safety training material link address data storage repository nodes stored in the DSP node, and split the laboratory safety training material link address joint query request into laboratory safety training material link address sub-queries for each target laboratory safety training material link address data storage repository node. The data source request authentication unit is used for each DSP node to send the laboratory safety training material link address sub-query to the target laboratory safety training material link address data storage repository node, and complete the laboratory safety training material link address sub-query after automatically verifying the legitimacy of the authorization information and access rights of both parties between the DSP node and the target laboratory safety training material link address data storage repository node through a pre-established smart contract.

5. The laboratory safety education course integrated push system according to claim 3 is characterized in that: The query result acquisition module includes: The data access permission verification and automatic execution unit is used for the target laboratory safety training material link address data repository node to verify the permissions of the received DSP node based on the local authorization list through the permission control module built into the smart contract, and automatically execute the laboratory safety training material matching and link address query operations after the verification is passed, and format the matching and laboratory safety training material link address sub-query results and send them back to the DSP node.

6. The laboratory safety education course integrated push system according to claim 3 is characterized in that: The query result verification and integration module includes: The proof-of-work calculation unit is used to merge the laboratory safety training material link address sub-query results of each DSP node into a laboratory safety training material link address joint query result through a predetermined data merging algorithm in the DSP node, and package the laboratory safety training material link address joint query result, timestamp, previous block hash value and random number into a candidate block, and use a consensus mechanism to complete the consensus calculation of the candidate block; A cross-node query result verification unit is used to verify the query return transaction containing the timestamp and the link address of the laboratory safety training materials to the entire blockchain through the smart contract. Other DSP nodes verify the query return transaction by calling the multi-layer verification logic built into the smart contract, including signature verification, hash value comparison and cross-verification of data integrity, timestamp accuracy and query result correctness; The result recording and blockchain storage unit is used to package the verified query return transaction into a legal block after confirmation by the smart contract, and append the legal block to the blockchain after confirmation by all DSP nodes through the consensus mechanism; The blockchain continuous iteration unit is used to continuously generate new blocks according to the preset consensus mechanism to form a chain structure; The query result returning unit is used to automatically package the joint query result of the laboratory safety training material link address through the smart contract and return it to the query terminal, and the DSP node calls the smart contract to store the verified joint query results of all laboratory safety training material link addresses that meet the conditions into the laboratory safety training material database.

7. The laboratory safety education course integrated push system according to claim 6 is characterized in that: The proof-of-work calculation unit includes: The integrated computing sub-unit is used to record the timestamp, data summary and source information of each laboratory safety training material link address sub-query result through the smart contract, and merge the laboratory safety training material link address sub-query results of each DSP node into a laboratory safety training material link address joint query result using a predetermined data merging algorithm within the DSP node; wherein, the smart contract assists in verifying the data consistency during the merging process; Constructing a candidate block sub-unit: used to package the laboratory safety training material link address, the query result, the timestamp, the hash value of the previous block, and the random number into a candidate block, and the smart contract adds conditional verification logic; A consensus calculation subunit, configured to complete the consensus calculation of the candidate block using a consensus mechanism, monitor the consensus calculation process in real time through smart contracts, and record transaction status and computing power contributions; the consensus mechanism is any one of Proof of Work (PoW), Delegated Proof of Stake (DPoS), Practical Byzantine Fault Tolerance (PBFT), and Proof of Authority (PoA); The block broadcast and incentive sub-unit is used to broadcast the candidate area through the blockchain. After passing the verification triggered by more than half of the DSP nodes based on the smart contract, the candidate area is added to the blockchain, and the incentive calculation and distribution are automatically triggered to distribute digital currency rewards to the DSP nodes that generated the candidate area.

8. The laboratory safety education course integrated push system according to claim 1 is characterized in that: The course grading and classification push subsystem includes: The course grading and classification push module is used to input the trainer grading and classification information, the wrong question grading and classification information, and the course grading and classification information corresponding to the link addresses of all laboratory safety training materials into the course material grading and classification push model, and predict the link address set corresponding to the trainer's compulsory laboratory safety training materials and the link address set corresponding to the optional laboratory safety training materials.

9. The laboratory safety education course integrated push system according to claim 8 is characterized in that: The course material classification and push model is used to push a set of link addresses corresponding to all required laboratory safety training materials matched in the laboratory safety training material library to each trainee according to the trainee classification information of each trainee; The course material grading and classification push model is also used to push a set of link addresses corresponding to all matching optional laboratory safety training materials in the laboratory safety training material library to each trainee based on the grading and classification information of the wrong questions of each trainee.

10. A laboratory safety education course integrated push method, characterized in that: The method comprises: Use deep learning technology to classify trainees and obtain training personnel classification information; Through blockchain technology, the link addresses corresponding to different laboratory safety training materials are integrated and stored in the laboratory safety training database. Deep learning technology is used to classify the link addresses of laboratory safety training materials in the laboratory safety training database to obtain course classification information. Collect statistics on the types of questions that trainees are prone to making mistakes and their error rates, and use deep learning technology to classify each trainee's types of questions that are prone to making mistakes and their error rates, to obtain the classification information of each trainee's wrong questions; Unsupervised learning technology is used to process the training personnel classification information, the course classification information and the wrong question classification information to obtain a set of link addresses corresponding to all laboratory safety training materials matching each training personnel, and push them to the corresponding training personnel.

Images