A job transfer instruction response method and device based on knowledge enhancement and human-computer interaction, equipment and storage medium

Abstract

This application discloses a method, device, equipment, and storage medium for responding to job transfer instructions based on knowledge enhancement and human-computer interaction, involving the interdisciplinary field of artificial intelligence and information security. The instruction access layer generates raw instruction data from various channels; the intelligent parsing layer performs text standardization and OCR recognition on the data, and obtains a knowledge-enhanced representation vector through semantic encoding and knowledge matrix fusion, identifies entities and maps them to standard codes to obtain the job transfer task; the compliance decision layer calculates risk scores based on the task and generates execution control results; the human-computer interaction layer determines the source / target job permission set and permission change set according to the task and results, generates permission execution data, and performs two-factor verification if the execution action is a preset action; the execution audit layer adjusts the permissions and transfers the research personnel's accounts based on the above data and verification results to obtain the completion result, thereby improving the efficiency of transferring research positions and reducing the error rate and security risks of job transfer.

Description

Technical Field

[0001] This invention relates to the intersection of artificial intelligence and information security, and in particular to a method, apparatus, device, and storage medium for responding to job transfer instructions based on knowledge enhancement and human-computer interaction. Background Technology

[0002] Currently, automatically, accurately, and securely converting unstructured and ambiguous personnel reassignment instructions into standardized permission adjustment operations across multiple information systems involves the following issues: First, there's the issue of accurately parsing unstructured and unreliable information at the information flow level. Faced with unstructured carriers (relying on screenshots and text, information is free text or non-standard images, unable to be directly understood and processed by the system), informal channels (using instant messaging tools, information is easily buried, lacking formal process control for guaranteed delivery, receipts, and traceability), and non-standard content (notifications may lack key fields such as precise effective time, new job codes, and specific permission requirements, relying entirely on manual understanding and secondary communication), how can we robustly extract accurate personnel identifiers and source / target organization job identifiers from multimodal (text, image) non-standardized notifications and map them to the system's internal standard codes?

[0003] Secondly, there are issues related to manual intervention and operational risk control at the control flow level. Faced with situations where triggering is entirely manual (the process relies on HR personnel remembering and executing notifications, posing risks of forgetting or delays), execution is entirely manual (system administrators need to complete reading notifications → logging into the system → finding personnel → understanding old / new positions → manually selecting / adjusting a series of permissions, a process highly dependent on personal experience and responsibility), and there is a lack of review and auditing (adjustment operations lack automated verification mechanisms to confirm whether the adjustment results are consistent with the adjustment commands), how can we dynamically determine whether the adjustment complies with established compliance strategies (such as requirements for confidential positions and the principle of separation of financial permissions) before execution, and automatically prevent risks?

[0004] Thirdly, there are issues of system fragmentation and logical deficiencies at the data flow level. Faced with situations where permission mapping relies on memorization (the correspondence between job positions and system permission sets exists only in experience or scattered documents, lacking structured rules), unrelated change records (permission adjustment operation logs are separated from job transfer documents, making it difficult to quickly pinpoint responsibility when problems arise), and a lack of compliance checks (when transferring to confidential, financial, or other special positions, the system cannot automatically check whether the personnel have completed the necessary training, signed specific agreements, or other prerequisites), how can we provide effective visual decision support and error prevention mechanisms for high-risk or batch operations to ensure the traceability and non-repudiation of human-machine collaborative operations?

[0005] The current approach is essentially a "breakpoint" process that relies entirely on manual bridging. The lack of digital connection and automated logic between the command and execution systems leads to inefficiency, high error rates, and significant security risks.

[0006] As can be seen from the above, how to improve the efficiency of reassigning research positions and reduce the error rate and safety risks of reassignment in the process of responding to reassignment orders is an urgent problem to be solved. Summary of the Invention

[0007] In view of this, the purpose of this invention is to provide a method, apparatus, device, and storage medium for responding to job reassignment instructions based on knowledge enhancement and human-computer interaction, which can improve the efficiency of reassigning research positions during the job reassignment instruction response process. The specific solution is as follows: Firstly, this application provides a method for responding to job transfer instructions based on knowledge enhancement and human-computer interaction, applied to a research position transfer system including an instruction access layer, an intelligent parsing layer, a compliance decision-making layer, a human-computer interaction layer, and an execution audit layer. The method includes: The original instruction data is generated using the instruction access layer and based on the job transfer instructions from each channel; The original instruction data is normalized and image OCR is performed using an intelligent parsing layer to obtain a text sequence. Then, the text sequence is semantically encoded and fused with an organized knowledge matrix to obtain a knowledge-enhanced representation vector. Finally, entity recognition and standard code mapping are performed on the knowledge-enhanced representation vector to obtain the job transfer task. The compliance decision-making layer is used to determine a risk score based on the job transfer task, and then an execution control result is generated based on the risk score; the execution control result includes execution actions; The source job permission set and the target job permission set are determined using the human-computer interaction layer and based on the job transfer task and the execution control result. The permission change set is then determined based on the source job permission set and the target job permission set. Permission execution data is then generated based on the execution control result and the permission change set. If the execution action is a preset action, two-factor authentication is performed based on the permission execution data to obtain a confirmation result. Using the execution audit layer and based on the permission execution data and the confirmation result, the permissions of the research personnel accounts corresponding to the employee identifiers in the reassignment task are adjusted and reassigned, resulting in the reassignment completion result.

[0008] Optionally, the step of using an intelligent parsing layer to perform text standardization and image OCR recognition on the original instruction data to obtain a text sequence includes: The multimodal job transfer instruction intelligent parsing module in the intelligent parsing layer is used to filter noise in the text data of the original instruction data to obtain noise-filtered data, and the length of the noise-filtered data is controlled to obtain length-controlled data. Optical character recognition and conversion are performed on the image data in the original instruction data to obtain the converted result, and a standardized text sequence with a unified format is generated based on the length-controlled data and the converted result.

[0009] Optionally, the step of performing text semantic encoding and fusing the organized knowledge matrix on the text sequence to obtain a knowledge-enhanced representation vector includes: The text sequence is semantically encoded using a large language model to obtain a text vector representation; The text vector representation is subjected to attention-weighted fusion processing based on the organizational knowledge matrix to obtain a knowledge-enhanced representation vector.

[0010] Optionally, the step of performing entity recognition and standard code mapping on the knowledge-enhanced representation vector to obtain the job transfer task includes: The knowledge-enhanced representation vector is subjected to sequence labeling to obtain labeling results, and entity recognition is performed on the labeling results to obtain entity recognition results; the entity recognition results include personnel entity information and job entity information; Based on preset standard coding mapping rules, the entity identification results are mapped to employee identifiers, source job codes, and target job codes respectively. A job transfer task is generated based on the employee identifier, the source job code, the target job code, and the task effective time. The job transfer task is then sent to the compliance decision-making layer.

[0011] Optionally, the step of utilizing the compliance decision-making layer and determining a risk score based on the job transfer task, and then generating an execution control result based on the risk score, includes: The compliance decision-making layer is used to determine the personnel status data set based on the job transfer task, and the personnel status data set is matched one by one with the preset rule set to obtain the matching result; Determine the rule weight corresponding to each rule, and determine whether each rule satisfies the state condition to obtain the judgment result. Then, determine the risk score based on the judgment result, the rule weight and the matching result. The risk score is compared with a preset threshold, and an execution control result corresponding to the job transfer task is generated based on the comparison result. The execution control result is then sent to the human-computer interaction layer. The execution control result includes a risk level and an execution action.

[0012] Optionally, the process involves using the human-computer interaction layer and determining the source and target job permission sets based on the job transfer task and the execution control result, then determining a permission change set based on the source and target job permission sets, and finally generating permission execution data based on the execution control result and the permission change set. If the execution action is a preset action, then a two-factor authentication is performed based on the permission execution data to obtain a confirmation result, including: The source job permission set and the target job permission set are determined using the human-computer interaction layer and based on the job transfer task and the execution control result. Then, set difference and set union operations are performed on the source job permission set and the target job permission set to obtain the permissions to be added and the permissions to be removed. A permission change set is generated based on the permissions to be added and the permissions to be removed. Permission execution data is generated based on the execution control result and the permission change set; the permission execution data includes the permission change set and the execution action. If the execution action is manual confirmation or blocking, then the human-computer interaction layer is used to perform two-factor authentication based on the permission execution data, and then a confirmation result is generated after the authentication is passed, so that the confirmation result is sent to the execution audit layer; the two-factor authentication includes sending a dynamic verification code to a preset security device and performing biometric identification.

[0013] Optionally, the step of using the execution audit layer and adjusting the permissions and reassigning the research personnel account corresponding to the employee identifier in the reassignment task based on the permission execution data and the confirmation result to obtain the reassignment completion result includes: Using the execution audit layer and based on the permission execution data and the confirmation result, the permissions of the research personnel account corresponding to the employee identifier in the job transfer task are adjusted to obtain the permission execution result. Based on the permission execution result, the job status corresponding to the job transfer task is updated to obtain the job transfer completion result. Based on the job transfer completion result, the job transfer task, and the confirmation result, an audit evidence chain is generated to perform closed-loop processing from receiving the job transfer instruction to executing permissions and auditing.

[0014] Secondly, this application provides a job transfer instruction response device based on knowledge enhancement and human-computer interaction, applied to a research position transfer system including an instruction access layer, an intelligent parsing layer, a compliance decision-making layer, a human-computer interaction layer, and an execution audit layer. The device includes: The raw instruction data generation module is used to generate raw instruction data based on the instruction access layer and the transfer instructions from various channels. The text sequence generation module is used to perform text standardization and image OCR recognition on the original instruction data using the intelligent parsing layer to obtain a text sequence. Then, the text sequence is subjected to text semantic encoding and knowledge matrix fusion to obtain a knowledge-enhanced representation vector. Finally, entity recognition and standard code mapping are performed on the knowledge-enhanced representation vector to obtain the job transfer task. The execution control result generation module is used to determine a risk score based on the compliance decision-making layer and the job transfer task, and then generate an execution control result based on the risk score; the execution control result includes execution actions; The permission execution data generation module is used to determine the source position permission set and the target position permission set based on the human-computer interaction layer and the job transfer task and the execution control result, to determine the permission change set based on the source position permission set and the target position permission set, and then generate permission execution data based on the execution control result and the permission change set. If the execution action is a preset action, then perform two-factor verification based on the permission execution data to obtain a confirmation result. The job transfer completion result determination module is used to adjust the permissions and transfer the research personnel account corresponding to the employee identifier in the job transfer task by using the execution audit layer and based on the permission execution data and the confirmation result, so as to obtain the job transfer completion result.

[0015] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor is used to execute the computer program to implement the aforementioned method for responding to job transfer instructions based on knowledge enhancement and human-computer interaction.

[0016] Fourthly, this application provides a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the aforementioned method for responding to job transfer instructions based on knowledge enhancement and human-computer interaction.

[0017] As can be seen from the above, before responding to the job transfer instruction, this application needs to utilize the instruction access layer to generate original instruction data based on job transfer instructions from various channels; utilize the intelligent parsing layer to perform text standardization and image OCR recognition on the original instruction data to obtain a text sequence, then perform text semantic encoding and knowledge matrix fusion on the text sequence to obtain a knowledge-enhanced representation vector, and perform entity recognition and standard code mapping on the knowledge-enhanced representation vector to obtain the job transfer task; utilize the compliance decision layer to determine a risk score based on the job transfer task, and then generate an execution control result based on the risk score; the execution control result includes execution actions; utilize the human-computer interaction layer to determine the source job permission set and the target job permission set based on the job transfer task and the execution control result, determine the permission change set based on the source job permission set and the target job permission set, and then generate permission execution data based on the execution control result and the permission change set. If the execution action is a preset action, then perform two-factor validation based on the permission execution data to obtain a confirmation result; utilize the execution audit layer to adjust the permissions and transfer the research personnel account corresponding to the employee identifier in the job transfer task based on the permission execution data and the confirmation result, and obtain the job transfer completion result.

[0018] Therefore, this application first needs to utilize the instruction access layer to generate raw instruction data based on transfer instructions from various channels; then, the intelligent parsing layer performs text standardization and image OCR recognition on the raw instruction data to obtain a text sequence, followed by text semantic encoding and knowledge matrix fusion to obtain a knowledge-enhanced representation vector, and finally entity recognition and standard code mapping on the knowledge-enhanced representation vector to obtain the transfer task; secondly, the compliance decision layer determines a risk score based on the transfer task, and then generates an execution control result based on the risk score; the execution control result includes execution actions; the human-computer interaction layer determines the source position permission set and the target position permission set based on the transfer task and the execution control result, and determines the permission change set based on the source position permission set and the target position permission set, and then generates permission execution data based on the execution control result and the permission change set. If the execution action is a preset action, two-factor validation is performed based on the permission execution data to obtain a confirmation result; finally, the execution audit layer adjusts the permissions and transfers the research personnel accounts corresponding to the employee identifiers in the transfer task based on the permission execution data and the confirmation result to obtain the transfer completion result. This improves the efficiency of reassigning research positions during the reassignment process, thereby reducing the error rate and safety risks associated with reassignment. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This application discloses a flowchart of a job transfer instruction response method based on knowledge enhancement and human-computer interaction. Figure 2 This is a schematic diagram of the overall architecture of a specific job transfer instruction response based on knowledge enhancement and human-computer interaction disclosed in this application; Figure 3 This is a schematic diagram of a specific HR-InstructionNet model structure disclosed in this application; Figure 4 This is a schematic diagram of a specific compliance rule engine workflow disclosed in this application; Figure 5 This is a schematic diagram of the interface of a specific visual decision-making auxiliary device disclosed in this application; Figure 6 This is a schematic diagram of a job transfer instruction response device based on knowledge enhancement and human-computer interaction disclosed in this application; Figure 7 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Currently, automatically, accurately, and securely converting unstructured and ambiguous personnel reassignment instructions into standardized permission adjustment operations across multiple information systems presents the following challenges: First, the problem of accurately parsing unstructured and unreliable information at the information flow level. Second, the problem of manual driving and operational risk control at the control flow level. Third, the problem of system fragmentation and logical gaps at the data flow level. The essence of current methods is a "breakpoint" process entirely reliant on manual bridging. The lack of digital connection and automated logic between the command system and the execution system leads to low efficiency, high error rates, and significant security risks. Therefore, this application provides a reassignment instruction response method based on knowledge enhancement and human-computer interaction, which can improve the efficiency of reassigning research positions during the reassignment instruction response process.

[0023] See Figure 1 As shown in the figure, this invention discloses a method for responding to job transfer instructions based on knowledge enhancement and human-computer interaction. This method is applied to a research position transfer system that includes an instruction access layer, an intelligent parsing layer, a compliance decision-making layer, a human-computer interaction layer, and an execution audit layer. The method includes: Step S11: Utilize the instruction access layer and generate raw instruction data based on the job transfer instructions from each channel.

[0024] In this embodiment, the present application proposes a five-layer architecture system, specifically involving the fields of natural language processing, rule engine, and visual interaction technology, and the overall system architecture diagram is shown below. Figure 2 As shown, this illustrates the system's five-layer architecture and data flow relationships. Figure 2 The annotations indicate the data object transfer between each layer, including raw commands (Raw_Command), structured tasks (Structured_Task), execution commands (Execution_Order), and audit evidence chains (Audit_Evidence_Chain). External systems, such as HRMS, OA, and internal instant messaging tools, input through the command access layer, and the execution audit layer ultimately adjusts permissions and generates immutable logs, forming a complete "perception-decision-execution-audit" closed loop.

[0025] In this embodiment, the present application embodiment achieves full-process automation from instruction parsing to permission execution through three core modules: a multimodal job transfer instruction intelligent parsing module, a dynamic compliance rule engine module, and a visualization auxiliary decision-making device module.

[0026] The system comprises the following layers: Instruction Access Layer: Unified reception of job transfer instructions from multiple channels; Intelligent Parsing Layer: Parsing unstructured instructions using the HR-InstructionNet model; Compliance Decision Layer: Real-time risk assessment based on a rule engine; Human-Computer Interaction Layer: Providing visual decision support and security confirmation; and Execution Audit Layer: Executing permission adjustments and generating audit evidence chains.

[0027] It's worth noting that the three functional modules mentioned above are the core technology engines vertically embedded in the five-layer architecture. They respectively endow the intelligent parsing layer, compliance decision-making layer, and human-computer interaction layer with intelligence and processing capabilities. The five-layer architecture ensures the clarity and orderliness of data and control flows, while the three modules, within this orderly framework, achieve intelligent breakthroughs in key areas. Together, they form a complete, intelligent, and secure closed-loop system from "perception" to "execution" to "auditing."

[0028] In this embodiment, the intelligent parsing module for multimodal job reassignment instructions serves as the system's perception center, responsible for transforming unstructured, multimodal (text / image) research job reassignment instructions into structured, standardized business tasks. Its core lies in establishing the HR-InstructionNet model, which, through a dual-channel encoding mechanism enhanced with organizational knowledge, outputs structured job reassignment instruction response tasks. Task={Employee_ID,Source_Role_Code,Target_Role_Code,Effective_Time}; Wherein, Employee_ID is the employee identifier, Source_Role_Code is the source role code, Target_Role_Code is the target role code, and Effective_Time is the effective time.

[0029] This solves the problems of low accuracy in identifying entity relationships and difficulty in code mapping in the field of job transfer business using existing technologies.

[0030] The schematic diagram of the HR-InstructionNet model structure is shown below. Figure 3 As shown, where, Figure 3 The model's three-layer core structure is shown, including: 1) Input layer: a multimodal input processor component that receives multi-source research position adjustment instructions such as text or images, and unifies them into standardized text sequences to provide standardized input for subsequent encoding; 2) Encoding Fusion Layer: This is a dual-channel encoder and fusion component. The text semantic channel uses a pre-trained language model (such as RoBERTa-wwm-ext) to extract 768-dimensional character vectors. The organization knowledge channel loads a learnable organization knowledge matrix K, which is then fused through a knowledge attention mechanism to generate a 1536-dimensional organization knowledge enhanced representation vector, thus achieving knowledge attention fusion. 3) Output layer: The original representation and the knowledge-enhanced representation are concatenated, and the sequence is labeled by the CRF decoder (Conditional Random Field Decoder) to identify entities such as personnel, source positions, and target positions and map them to standard system code, directly outputting the structured job transfer task.

[0031] In this way, the above module consists of three sub-components: multimodal input processor, dual-channel encoder and fusion unit, and structured output: (1) Multimodal input processor component: responsible for receiving multi-source scientific research post adjustment instructions, completing modal unification and text standardization, and providing standardized input for subsequent encoding.

[0032] Step S12: Use the intelligent parsing layer to perform text standardization and image OCR recognition on the original instruction data to obtain a text sequence. Then, perform text semantic encoding and knowledge matrix fusion on the text sequence to obtain a knowledge-enhanced representation vector. Finally, perform entity recognition and standard code mapping on the knowledge-enhanced representation vector to obtain the job transfer task.

[0033] In this embodiment, the present application supports direct text input and image OCR (Optical Character Recognition) recognition (based on the Tesseract engine). Through noise filtering (such as removing irrelevant phrases like "after research and decision") and length control (matching the BERT512 character input limit), a unified text sequence is output. S =[ s 1, s 2,..., sn ].

[0034] Specifically, the intelligent parsing layer performs text standardization and image OCR recognition on the original instruction data to obtain a text sequence. This can include: using the multimodal job transfer instruction intelligent parsing module in the intelligent parsing layer to filter noise from the text data in the original instruction data to obtain noise-filtered data, and then performing length control processing on the noise-filtered data to obtain length-controlled data; performing optical character recognition and conversion on the image data in the original instruction data to obtain the converted result, and generating a standardized text sequence in a unified format based on the length-controlled data and the converted result.

[0035] Furthermore, embodiments of this application utilize a dual-channel encoder and fusion component to process the input instruction text sequence. S =[ s 1, s 2,..., sn A dual-channel encoder and fusion processor are constructed. The text semantic channel uses a pre-trained language model (such as RoBERTa-wwm-ext) to extract the context vector representation H=[h1,h2,...,h] for each character. n ],in, For the first i Vector representation of each character, d =768; Organize knowledge channels and build a learnable organizational knowledge matrix. , m The number of standard entities (e.g., departments, positions), k per row. j A standard vector represents an entity; the fusion processor generates an organizational knowledge-enhanced representation by calculating the interaction weights between text and knowledge, thereby achieving knowledge attention fusion.

[0036] The formula for calculating the interaction weight between text and knowledge is as follows: ; in, Aij Let be the attention weight, representing the th i The character is paired with the first j The degree of attention paid to each knowledge entity; h i For text character vectors, the output of the pre-trained model is the first... i Individual vectors; k j Let K be the knowledge entity vector, and let K be the _th_ in the knowledge matrix. j One standard entity vector; d The vector dimension is a fixed value of 768 (a fixed architecture parameter of the RoBERTa-wwm-ext model). m The number of knowledge entities is the total number of standard entities in the organization (e.g., 1000 departments or positions). This is a scaling factor used to prevent the gradient from vanishing due to an excessively large dot product.

[0037] Furthermore, in this embodiment, a knowledge-enhanced text representation is obtained through weighted summation, and the generation formula is shown below: ; in, Let be the knowledge enhancement vector, representing the th digit after integrating and organizing knowledge. i Each word vector represents; αij The attention weights are derived from the calculation results of Equation 1; k j For knowledge entity vectors, the definition is the same as before.

[0038] Specifically, performing text semantic encoding on a text sequence and fusing it with an organized knowledge matrix to obtain a knowledge-enhanced representation vector can include: using a large language model to perform semantic encoding on the text sequence to obtain a text vector representation; and performing attention-weighted fusion processing on the text vector representation based on the organized knowledge matrix to obtain a knowledge-enhanced representation vector.

[0039] In this embodiment, the present application selects to use a structured output component to concatenate the original representation and the knowledge-enhanced representation. The sequence is then fed into a CRF (Conditional Random Field Layer) for sequence labeling to identify entities such as personnel, departments, and positions. Subsequently, the identified entity names are looked up in the metadata corresponding to the knowledge matrix K, mapped to standard system code, and a structured job reassignment task is output. Task={Employee_ID,Source_Role_Code,Target_Role_Code,Effective_Time}.

[0040] Specifically, the process of performing entity recognition and standard code mapping on the knowledge-enhanced representation vector to obtain the job transfer task can include: performing sequence labeling on the knowledge-enhanced representation vector to obtain the labeling results, and performing entity recognition on the labeling results to obtain the entity recognition results; the entity recognition results include personnel entity information and job entity information; mapping the entity recognition results to employee identifiers, source job codes, and target job codes respectively based on preset standard coding mapping rules, and generating the job transfer task based on the employee identifier, source job code, target job code, and task effective time, and then issuing the job transfer task to the compliance decision-making level.

[0041] Step S13: Utilize the compliance decision-making layer and determine the risk score based on the job transfer task, and then generate the execution control result based on the risk score; the execution control result includes execution actions.

[0042] In this embodiment, the present application includes a dynamic compliance rule engine module, and the workflow diagram of the compliance rule engine is shown below. Figure 4 As shown, its core lies in upgrading static rules into dynamic, context-aware decision flows to address issues such as the lack of real-time compliance checks and risk quantification in traditional permission systems. Specifically, the aforementioned module receives structured tasks and outputs a risk level (Risk_Level) and an execution suggestion (Action) (automatic execution, manual confirmation required, or blocking).

[0043] It is worth mentioning that the above module is the decision center of the system. Through formal rule modeling and real-time status verification, it realizes the advance assessment and automatic blocking of job transfer risks.

[0044] The rule formalization engine is used to define compliance rules as quintuples: ; Where ID is the unique identifier of the rule; Trigger is the rule triggering condition, such as Target_Role_Codein['ROLE_PM','ROLE_SEC']; Condition is the set of state conditions to be satisfied, represented as a logical expression, whose atomic propositions are used to query the real-time status of personnel, such as: isTrainingCompleted(Employee_ID,'Confidential Training') == True; Furthermore, Weight represents the rule weight; Action represents the execution action, either BLOCK / MANUAL / AUTO.

[0045] Subsequently, real-time status queries and risk scoring are performed: the engine parses the Condition, generates parallel query tasks, and sends requests to the job training system, project management system, and funding management system to obtain status values. s 1, s 2,..., sp Whether a rule is satisfied is determined by a logical function. F ( s 1, s 2,..., sp Determine the risk score. R : ; in, R A risk score is assigned to quantify the overall risk of this job transfer; q The number of rules triggered indicates the number of rules that were matched. wi Let be the weight of the i-th rule, with a value of {0,5}, determined by the analytic hierarchy process. This is an indicator function; it returns 1 if the condition is not met, and 0 otherwise. Fi The logical function for the i-th rule takes the values ​​{True, False}, and represents the logical combination of multi-state query results; s 1... sp The status query results are obtained from the job training system, project management system, and financial management system, including personnel training status, security classification, and financial status.

[0046] Subsequently, decision thresholds are determined: risk decision thresholds are set. θ Based on the risk score obtained above R， Determine the action to be performed: ; in, θ high The high-risk threshold is set to 1.0. θ low The low-risk threshold is set to 0.5; Action is the action to be performed, with values ​​of BLOCK / MANUAL / AUTO.

[0047] Specifically, the risk score is determined using the compliance decision-making layer based on the job transfer task, and then execution control results are generated based on the risk score. This process may include: determining a set of personnel status data using the compliance decision-making layer based on the job transfer task, and matching the personnel status data set with a set of preset rules one by one to obtain matching results; determining the rule weight corresponding to each rule, and judging whether each rule meets the status conditions to obtain judgment results, and then determining the risk score based on the judgment results, rule weights, and matching results; comparing the risk score with a preset threshold, and generating execution control results corresponding to the job transfer task based on the comparison results, so as to send the execution control results to the human-computer interaction layer; the execution control results include risk level and execution actions.

[0048] Step S14: Utilize the human-computer interaction layer and determine the source job permission set and target job permission set based on the job transfer task and the execution control result. Then, determine the permission change set based on the source job permission set and the target job permission set. Finally, generate permission execution data based on the execution control result and the permission change set. If the execution action is a preset action, perform two-factor authentication based on the permission execution data to obtain a confirmation result.

[0049] In this embodiment, the present application selects to introduce a visualization-assisted decision-making device module, wherein the interface diagram of the visualization-assisted decision-making device is as follows. Figure 5 As shown, Figure 5 The presentation showcases a visual decision-making interface layout. Its core lies in transforming abstract permission changes into a perceptible graph structure and designing a fault-proofing interactive process. It provides a dedicated interface for high-risk or batch tasks. The central element is a permission impact graph, centered on the user node. Solid green lines indicate newly added permissions, while red / gray dashed lines indicate removed permissions (high-risk items are highlighted in red). The top of the interface displays task information and risk warning boxes, while the bottom is an operation confirmation area. It supports checkbox confirmation and two-factor authentication via OTP (One-Time Password) verification codes. Upon confirmation, vouchers are generated and linked to audit logs to ensure the traceability of high-risk operations.

[0050] It is worth mentioning that the above module is the system's security operation console. Through the permission impact graph and two-factor confirmation protocol, it solves the problems of unintuitive change impact, lack of warnings for high-risk operations, and insecure confirmation process in traditional permission change management.

[0051] Subsequently, this embodiment of the application can perform permission change set calculation: based on the task list to be adjusted, the permission set of the relevant roles is extracted from the permission configuration library. Psrc and Ptgt .

[0052] In this embodiment, the application requires calculating the permission change set: ; Where, Δ P This is the set of permission changes, representing all permission change operations that need to be performed. Psrc The source job permission set represents all the permissions the employee had before the job transfer; Ptgt The target job permission set represents all the permissions the employee should have after the job transfer; The ∪ operator is used for set difference operations to calculate permissions that are "existing but no longer needed" or "needed but not yet available"; the ∪ operator is used for set union operations to merge added and removed permissions.

[0053] Subsequently, the permission-influenced atlas is generated: using the graph G =( V , E Visualization, nodes V Representatives and system resources, on the side E This represents the granting or removal of permissions, and uses color and thickness to indicate the type and risk level of the change to the role: ; in ，G This is a permission-based impact graph, representing the data structure for visualization. V This is a set of nodes, representing user nodes plus system resource nodes; E Let be a set of edges, representing permission relationships (granting / removing).

[0054] The definition of an edge is as follows: ; Furthermore, this application embodiment requires two-factor authentication for security verification: for high-risk operations, the device requires "two-factor authentication." The administrator first selects the tasks to be processed on the visual interface, and then must enter a dynamic verification code (sent to their reserved security device) or perform biometric identification before finally submitting. All operation steps, dwell time, and confirmation actions are recorded as audit events (Eventaudit).

[0055] Specifically, the source and target job permission sets are determined using the human-computer interaction layer based on the job transfer task and execution control results. A permission change set is then determined based on these sets. Permission execution data is generated based on the execution control results and the permission change set. If the execution action is a preset action, two-factor authentication is performed on the permission execution data to obtain a confirmation result. This process may include: determining the source and target job permission sets using the human-computer interaction layer based on the job transfer task and execution control results; performing set difference and set union operations on the source and target job permission sets to obtain permissions to be added and permissions to be removed; generating a permission change set based on these permissions; and generating permission execution data based on the execution control results and the permission change set. The permission execution data includes the permission change set and the execution action. If the execution action requires manual confirmation or blocking, two-factor authentication is performed using the human-computer interaction layer based on the permission execution data. After successful authentication, a confirmation result is generated and sent to the execution audit layer. Two-factor authentication includes sending dynamic verification codes to preset security devices and performing biometric identification.

[0056] Step S15: Using the execution audit layer and based on the permission execution data and the confirmation result, adjust the permissions and transfer the research personnel account corresponding to the employee identifier in the transfer task to obtain the transfer completion result.

[0057] In this embodiment, the present application requires the use of an execution audit layer to adjust the permissions and transfer the research personnel account corresponding to the employee identifier in the transfer task based on the permission execution data and confirmation results. Specifically, the execution audit layer is used to adjust the permissions of the research personnel account corresponding to the employee identifier in the transfer task based on the permission execution data and confirmation results, obtaining the permission execution result. The job status corresponding to the transfer task is updated based on the permission execution result, resulting in the transfer completion result. An audit evidence chain is generated based on the transfer completion result, the transfer task, and the confirmation result, so as to perform closed-loop processing from the receipt of the transfer instruction to the permission execution and audit based on the audit evidence chain.

[0058] In one specific implementation, in a scenario where a technical position is transferred to a planning position, after the human resources department approves the transfer through the internal system, a research position adjustment order is generated: "Effective immediately, Zhang San (ID: EMP2025) will be transferred from the application development group to the planning group project management position."

[0059] Then, proceed to step 1: instruction access and parsing: (1) The instruction is captured by the instruction access layer in text form and sent to the system.

[0060] (2) The HR-InstructionNet of the intelligent parsing layer is started to perform instruction parsing.

[0061] Then, the text encoder transforms the instructions into a vector sequence H, where the organizational knowledge matrix K contains standard vectors such as Application Development Group (DEPT_DEV), Planning and Demonstration Group (DEPT_PLAN), Software Development Role (ROLE_DEV), and Project Management Role (ROLE_PM). A knowledge attention mechanism is used to calculate the high weights of "Planning and Demonstration Group" and "Planning Group" in the instructions, as well as the high weights of "Software Development Role" and "Development Role" in the instructions, thereby enhancing their vector representation. Furthermore, the CRF layer is used to identify entities: [Personnel: Zhang San], [Source Position: Application Development Team, Software Development Position], [Target Position: Planning and Demonstration Team, Project Management Position].

[0062] Structured output: Through knowledge matrix mapping, output structured tasks: Task={Employee_ID:'EMP2025',Source_Role_Code:'DEPT_DEV.ROLE_DEV',Target_Role_Code:'DEPT_PLAN.ROLE_PM',Effective_Time:'Current Time'}

[0063] Next, proceed to step 2: compliance risk assessment: (1) The compliance decision-making level receives the Task. The rule engine matches the triggering rule: Trigger:Target_Role_Code=='DEPT_PLAN.ROLE_PM'.

[0064] (2) Rule association condition: Condition: isTrainingCompleted('EMP2025','Confidential Training') AND clearanceLevel('EMP2025')>='Secret'.

[0065] (3) The compliance engine concurrently initiates two queries: querying the training system for the status of confidentiality and strategic planning training. s 1. Query the security classification from the project management system. s 2.

[0066] (4) Assuming the query returns s 1 = True (Completed) s 2 = 'Internal' (conflict), condition not met.

[0067] (5) Calculate the risk score Due to unmet conditions and the sensitivity of the target position, the risk level is assessed as "high".

[0068] Furthermore, proceed to step 3: Human-machine collaboration safety verification: (1) Due to the “high” risk level, the Task is routed to the visualization device in the human-computer interaction layer.

[0069] (2) Perform calculate: New additions: Project Management System, Strategic Document Library (2 items); Removed: Code repository commits, test environment deployments (2 items); (3) The device generates an access permission impact graph, which will add access permissions for the project management system and strategic document repository, and remove permissions for code repository submission and test environment deployment. The code repository submission permission is highlighted in the graph and marked as high risk in red.

[0070] (4) The system administrator views the graph in the interface and confirms the impact of the change; selects the task and prepares to execute it; the system sends a dynamic verification code 123456 to the administrator through the internal instant messaging tool, which is valid for 5 minutes; the administrator enters the verification code 123456 in the device to complete the two-factor confirmation.

[0071] (5) Confirm credentials: Generate a token, the device records the event Event_{audit}={Task_ID,Admin_ID,Confirm_Time,Confirm_Method:'OTP'}, and sends the final execution instruction to the execution audit layer.

[0072] Finally, proceed to step 4: Execution and Audit: (1) The execution audit layer calls the APIs of each target system (project management system, document management system, code repository, etc.) according to... Perform atomic operations to automatically add or remove four permissions.

[0073] (2) Execution record: All API call results (success / failure) are recorded.

[0074] (3) Generate and store complete audit logs: associated instructions, parsed tasks, risk assessment results, manually confirmed events and execution results of each system, forming a closed loop.

[0075] In another specific implementation, in an abnormal scenario, such as: Wang Wu (ID: EMP2024) is transferred to a classified position but has not completed the confidentiality training.

[0076] The processing flow is as follows: (1) The compliance engine detected that the training status was False.

[0077] (2) Risk score (Exceeding the blocking threshold).

[0078] (3) The system automatically blocks the process and does not proceed to the manual confirmation stage.

[0079] (4) Interface prompt: "Confidentiality training not completed, process blocked".

[0080] (5) Audit record: execution_result: BLOCKED, block_reason: Training not completed.

[0081] As can be seen from the above, the embodiments of this application first need to utilize the instruction access layer to generate original instruction data based on the job transfer instructions from various channels; utilize the intelligent parsing layer to perform text standardization and image OCR recognition on the original instruction data to obtain a text sequence, and then perform text semantic encoding and knowledge matrix fusion on the text sequence to obtain a knowledge-enhanced representation vector, and perform entity recognition and standard code mapping on the knowledge-enhanced representation vector to obtain the job transfer task; secondly, utilize the compliance decision layer to determine the risk score based on the job transfer task, and then generate the execution control result based on the risk score; then, the execution control result includes the execution action; utilize the human-computer interaction layer to determine the source job permission set and the target job permission set based on the job transfer task and the execution control result, and determine the permission change set based on the source job permission set and the target job permission set, and then generate permission execution data based on the execution control result and the permission change set. If the execution action is a preset action, then perform two-factor validation based on the permission execution data to obtain the confirmation result; finally, utilize the execution audit layer to adjust the permissions and transfer the research personnel account corresponding to the employee identifier in the job transfer task based on the permission execution data and the confirmation result to obtain the job transfer completion result. This improves the efficiency of reassigning research positions during the reassignment process, thereby reducing the error rate and safety risks associated with reassignment.

[0082] Accordingly, see Figure 6 As shown, this application also provides a job transfer instruction response device based on knowledge enhancement and human-computer interaction, including: The raw instruction data generation module 11 is used to generate raw instruction data using the instruction access layer and based on the job transfer instructions from each channel; The text sequence generation module 12 is used to perform text standardization and image OCR recognition on the original instruction data using the intelligent parsing layer to obtain a text sequence. Then, the text sequence is subjected to text semantic encoding and knowledge matrix fusion to obtain a knowledge-enhanced representation vector. Finally, the knowledge-enhanced representation vector is subjected to entity recognition and standard code mapping to obtain the job transfer task. The execution control result generation module 13 is used to determine a risk score based on the compliance decision-making layer and the job transfer task, and then generate an execution control result based on the risk score; the execution control result includes execution actions; The permission execution data generation module 14 is used to determine the source position permission set and the target position permission set by utilizing the human-computer interaction layer and based on the job transfer task and the execution control result, to determine the permission change set based on the source position permission set and the target position permission set, and then to generate permission execution data based on the execution control result and the permission change set. If the execution action is a preset action, then a two-factor verification is performed based on the permission execution data to obtain a confirmation result. The job transfer completion result determination module 15 is used to adjust the permissions and transfer the research personnel account corresponding to the employee identifier in the job transfer task by using the execution audit layer and based on the permission execution data and the confirmation result, so as to obtain the job transfer completion result.

[0083] In some specific embodiments, the text sequence generation module 12 may specifically include: The text data noise filtering unit is used to filter the text data in the original instruction data using the multimodal job transfer instruction intelligent parsing module in the intelligent parsing layer to obtain noise-filtered data, and to perform length control processing on the noise-filtered data to obtain length-controlled data. The conversion result generation unit is used to perform optical character recognition and conversion on the image data in the original instruction data to obtain the conversion result, and generate a standardized text sequence in a unified format based on the length-controlled data and the conversion result.

[0084] In some specific embodiments, the text sequence generation module 12 may specifically include: The text vector representation generation unit is used to semantically encode the text sequence using a large language model to obtain a text vector representation. The knowledge-enhanced representation vector generation unit is used to perform attention-weighted fusion processing on the text vector representation based on the organized knowledge matrix to obtain the knowledge-enhanced representation vector.

[0085] In some specific embodiments, the text sequence generation module 12 may specifically include: The entity recognition result determination unit is used to perform sequence labeling on the knowledge-enhanced representation vector to obtain labeling results, and to perform entity recognition on the labeling results to obtain entity recognition results; the entity recognition results include personnel entity information and job entity information; The job transfer task generation unit is used to map the entity identification results into employee identifiers, source job codes, and target job codes respectively based on preset standard coding mapping rules, and to generate job transfer tasks based on the employee identifiers, source job codes, target job codes, and task effective time, and then issue the job transfer tasks to the compliance decision-making layer.

[0086] In some specific embodiments, the execution control result generation module 13 may specifically include: The matching result generation unit is used to determine the personnel status data set based on the compliance decision-making layer and the job transfer task, and to perform a matching process between the personnel status data set and a preset rule set to obtain the matching result. The judgment result determination unit is used to determine the rule weight corresponding to each rule, and to determine whether each rule satisfies the state condition, thereby obtaining the judgment result. Then, based on the judgment result, the rule weight, and the matching result, a risk score is determined. An execution control result generation unit is used to compare the risk score with a preset threshold, and generate an execution control result corresponding to the job transfer task based on the comparison result, so as to send the execution control result to the human-computer interaction layer; the execution control result includes risk level and execution action.

[0087] In some specific embodiments, the permission execution data generation module 14 may specifically include: The job permission set determination unit is used to determine the source job permission set and the target job permission set by utilizing the human-computer interaction layer and based on the job transfer task and the execution control result. Then, it performs set difference operation and set union operation on the source job permission set and the target job permission set to obtain the permissions to be added and the permissions to be removed. Finally, it generates a permission change set based on the permissions to be added and the permissions to be removed. A permission execution data generation unit is used to generate permission execution data based on the execution control result and the permission change set; the permission execution data includes the permission change set and the execution action. The confirmation result generation unit is used to perform two-factor authentication based on the permission execution data using the human-computer interaction layer if the execution action is manual confirmation or blocking, and then generate a confirmation result after the authentication is successful, so as to send the confirmation result to the execution audit layer; the two-factor authentication includes sending a dynamic verification code to a preset security device and performing biometric identification.

[0088] In some specific embodiments, the job transfer completion result determination module 15 may specifically include: The permission execution result generation unit is used to adjust the permissions of the research personnel account corresponding to the employee identifier in the job transfer task by using the execution audit layer and based on the permission execution data and the confirmation result, to obtain the permission execution result, and to update the job status corresponding to the job transfer task based on the permission execution result, to obtain the job transfer completion result; An audit evidence chain generation unit is used to generate an audit evidence chain based on the job transfer completion result, the job transfer task, and the confirmation result, so as to perform closed-loop processing from receiving the job transfer instruction to the execution of permissions and auditing based on the audit evidence chain.

[0089] Furthermore, embodiments of this application also disclose an electronic device, Figure 7 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the knowledge-enhanced and human-computer interaction-based job transfer instruction response method disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be a computer.

[0090] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0091] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0092] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the knowledge-enhanced and human-computer interaction-based job transfer instruction response method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0093] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned disclosed method for responding to job transfer instructions based on knowledge enhancement and human-computer interaction. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0094] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0095] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0096] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0097] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0098] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for responding to job transfer instructions based on knowledge enhancement and human-computer interaction, characterized in that, The method, applied to a research post reassignment system comprising an instruction access layer, an intelligent parsing layer, a compliance decision-making layer, a human-computer interaction layer, and an execution audit layer, includes: The original instruction data is generated using the instruction access layer and based on the job transfer instructions from each channel; The original instruction data is normalized and image OCR is performed using an intelligent parsing layer to obtain a text sequence. Then, the text sequence is semantically encoded and fused with an organized knowledge matrix to obtain a knowledge-enhanced representation vector. Finally, entity recognition and standard code mapping are performed on the knowledge-enhanced representation vector to obtain the job transfer task. The compliance decision-making layer is used to determine a risk score based on the job transfer task, and then an execution control result is generated based on the risk score; the execution control result includes execution actions; The source job permission set and the target job permission set are determined using the human-computer interaction layer and based on the job transfer task and the execution control result. The permission change set is then determined based on the source job permission set and the target job permission set. Permission execution data is then generated based on the execution control result and the permission change set. If the execution action is a preset action, two-factor authentication is performed based on the permission execution data to obtain a confirmation result. Using the execution audit layer and based on the permission execution data and the confirmation result, the permissions of the research personnel accounts corresponding to the employee identifiers in the reassignment task are adjusted and reassigned, resulting in the reassignment completion result.

2. The job transfer instruction response method based on knowledge enhancement and human-computer interaction according to claim 1, characterized in that, The method utilizes an intelligent parsing layer to perform text standardization and image OCR recognition on the original instruction data to obtain a text sequence, including: The multimodal job transfer instruction intelligent parsing module in the intelligent parsing layer is used to filter noise in the text data of the original instruction data to obtain noise-filtered data, and the length of the noise-filtered data is controlled to obtain length-controlled data. Optical character recognition and conversion are performed on the image data in the original instruction data to obtain the converted result, and a standardized text sequence with a unified format is generated based on the length-controlled data and the converted result.

3. The job transfer instruction response method based on knowledge enhancement and human-computer interaction according to claim 1, characterized in that, The step of performing text semantic encoding and knowledge matrix fusion on the text sequence to obtain a knowledge-enhanced representation vector includes: The text sequence is semantically encoded using a large language model to obtain a text vector representation; The text vector representation is subjected to attention-weighted fusion processing based on the organizational knowledge matrix to obtain a knowledge-enhanced representation vector.

4. The job transfer instruction response method based on knowledge enhancement and human-computer interaction according to claim 1, characterized in that, The process of performing entity recognition and standard code mapping on the knowledge-enhanced representation vector to obtain the job reassignment task includes: The knowledge-enhanced representation vector is subjected to sequence labeling to obtain labeling results, and entity recognition is performed on the labeling results to obtain entity recognition results; the entity recognition results include personnel entity information and job entity information; Based on preset standard coding mapping rules, the entity identification results are mapped to employee identifiers, source job codes, and target job codes respectively. A job transfer task is generated based on the employee identifier, the source job code, the target job code, and the task effective time. The job transfer task is then sent to the compliance decision-making layer.

5. The job transfer instruction response method based on knowledge enhancement and human-computer interaction according to claim 1, characterized in that, The process of utilizing the compliance decision-making layer and determining a risk score based on the job transfer task, and then generating execution control results based on the risk score, includes: The compliance decision-making layer is used to determine the personnel status data set based on the job transfer task, and the personnel status data set is matched one by one with the preset rule set to obtain the matching result; Determine the rule weight corresponding to each rule, and determine whether each rule satisfies the state condition to obtain the judgment result. Then, determine the risk score based on the judgment result, the rule weight and the matching result. The risk score is compared with a preset threshold, and an execution control result corresponding to the job transfer task is generated based on the comparison result. The execution control result is then sent to the human-computer interaction layer. The execution control result includes a risk level and an execution action.

6. The job transfer instruction response method based on knowledge enhancement and human-computer interaction according to claim 1, characterized in that, The process involves using the human-computer interaction layer and determining the source and target job permission sets based on the job transfer task and the execution control result. Then, a permission change set is determined based on the source and target job permission sets. Finally, permission execution data is generated based on the execution control result and the permission change set. If the execution action is a preset action, a two-factor authentication is performed based on the permission execution data to obtain a confirmation result, including: The source job permission set and the target job permission set are determined using the human-computer interaction layer and based on the job transfer task and the execution control result. Then, set difference and set union operations are performed on the source job permission set and the target job permission set to obtain the permissions to be added and the permissions to be removed. A permission change set is generated based on the permissions to be added and the permissions to be removed. Permission execution data is generated based on the execution control result and the permission change set; the permission execution data includes the permission change set and the execution action. If the execution action is manual confirmation or blocking, then the human-computer interaction layer is used to perform two-factor authentication based on the permission execution data, and then a confirmation result is generated after the authentication is passed, so that the confirmation result is sent to the execution audit layer; the two-factor authentication includes sending a dynamic verification code to a preset security device and performing biometric identification.

7. The method for responding to job transfer instructions based on knowledge enhancement and human-computer interaction according to any one of claims 1 to 6, characterized in that, The process of adjusting permissions and transferring positions for the research personnel accounts corresponding to the employee identifiers in the transfer task using the execution audit layer and based on the permission execution data and the confirmation result, to obtain the transfer completion result, includes: Using the execution audit layer and based on the permission execution data and the confirmation result, the permissions of the research personnel account corresponding to the employee identifier in the job transfer task are adjusted to obtain the permission execution result. Based on the permission execution result, the job status corresponding to the job transfer task is updated to obtain the job transfer completion result. Based on the job transfer completion result, the job transfer task, and the confirmation result, an audit evidence chain is generated to perform closed-loop processing from receiving the job transfer instruction to executing permissions and auditing.

8. A job transfer instruction response device based on knowledge enhancement and human-computer interaction, characterized in that, A research post reassignment system, applicable to a research post system comprising an instruction access layer, an intelligent parsing layer, a compliance decision-making layer, a human-computer interaction layer, and an execution audit layer, wherein the device includes: The raw instruction data generation module is used to generate raw instruction data based on the instruction access layer and the transfer instructions from various channels. The text sequence generation module is used to perform text standardization and image OCR recognition on the original instruction data using the intelligent parsing layer to obtain a text sequence. Then, the text sequence is subjected to text semantic encoding and knowledge matrix fusion to obtain a knowledge-enhanced representation vector. Finally, entity recognition and standard code mapping are performed on the knowledge-enhanced representation vector to obtain the job transfer task. The execution control result generation module is used to determine a risk score based on the compliance decision-making layer and the job transfer task, and then generate an execution control result based on the risk score; the execution control result includes execution actions; The permission execution data generation module is used to determine the source position permission set and the target position permission set based on the human-computer interaction layer and the job transfer task and the execution control result, to determine the permission change set based on the source position permission set and the target position permission set, and then generate permission execution data based on the execution control result and the permission change set. If the execution action is a preset action, then two-factor verification is performed based on the permission execution data to obtain a confirmation result. The job transfer completion result determination module is used to adjust the permissions and transfer the research personnel account corresponding to the employee identifier in the job transfer task by using the execution audit layer and based on the permission execution data and the confirmation result, so as to obtain the job transfer completion result.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor is configured to execute the computer program to implement the job transfer instruction response method based on knowledge enhancement and human-computer interaction as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program, wherein the computer program, when executed by a processor, implements the job transfer instruction response method based on knowledge enhancement and human-computer interaction as described in any one of claims 1 to 7.

Images