Private cloud collaborative management platform

Abstract

The invention discloses a privatized cloud collaborative management platform, and relates to the technical field of data management, the privatized cloud collaborative management platform comprises an edge-cloud fusion architecture, an AI dynamic security module, a self-adaptive collaborative engine and a zero code customization center, the edge-cloud fusion architecture forms a distributed management center of a collaborative management network, a quantum key distribution link based on a BB84 protocol is used for connecting an edge computing node and a central cloud end, the transmission rate is 1Gbps, the key updating period is 1 hour, and a distributed hash table based on a consistent hash algorithm is adopted to realize dynamic mapping and load balancing of data fragments. The method has the advantages that the distributed hash table based on the consistent hash algorithm is adopted, dynamic fragmentation and load balancing of data among 2-256 nodes are achieved, the method has the capabilities of automatic fault tolerance and copy migration, the edge nodes support automatic switching to local operation when the network is interrupted, and service continuity is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of data management technology, specifically a private cloud-based collaborative management platform. Background Technology

[0002] In recent years, with the deepening of enterprise digital transformation, cloud-based collaborative management platforms have been widely used in the operation of multi-branch and cross-regional organizations. However, existing public cloud-based platforms usually host enterprise data on third-party servers, which weakens data control and poses a risk of sensitive information leakage. Traditional encryption and permission mechanisms are mostly static settings, which are difficult to cope with increasingly complex network attacks. In addition, collaboration between different business systems relies on customized interface development, resulting in poor data format compatibility and significant synchronization delays. In cross-time zone collaboration scenarios, version conflicts and operational inconsistencies are prone to occur, which seriously affect collaboration efficiency. Although some enterprises adopt private deployment to strengthen local data control, traditional private platforms still have significant shortcomings in dynamic security protection, intelligent collaboration and flexible configuration. Security policies cannot be adjusted in real time according to the threat situation, cross-system data interoperability still relies heavily on manual processing, business customization requires professional development support, and the implementation cycle is long and costly, making it difficult to meet the needs of SMEs to respond quickly to business changes. Therefore, there is an urgent need for a new type of private cloud collaborative management platform that supports distributed management, has AI-driven security capabilities and automated collaboration functions. Summary of the Invention

[0003] The purpose of this invention is to provide a private cloud-based collaborative management platform.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a private cloud collaborative management platform, comprising an edge-cloud fusion architecture, an AI dynamic security module, an adaptive collaborative engine, and a zero-code customization center; The edge-cloud converged architecture constitutes the distributed management center of the collaborative management network. It uses a quantum key distribution link based on the BB84 protocol to connect the edge computing nodes and the central cloud, with a transmission rate of 1Gbps and a key update cycle of 1 hour. A distributed hash table based on the consistent hashing algorithm is used to realize dynamic mapping and load balancing of data shards, supporting elastic expansion of 256 nodes and automatic fault tolerance. The AI ​​dynamic security module includes a CNN-LSTM hybrid security situation awareness model, a risk-adaptive permission mechanism, and an intelligent sandbox isolation unit. The adaptive collaboration engine includes a cross-system semantic mapping unit and a spatiotemporal collaborative scheduling unit, which are used to realize semantic collaboration and asynchronous collaborative management among multiple systems. The zero-code customization center includes an industry-specific business template library, an atomic functional component library, and a user intent recognition unit, supporting the generation of business processes and data forms through natural language commands. As a further aspect of the present invention: the dynamic risk prevention and control index in the AI ​​dynamic security module is calculated by the following formula: ; in, This is a dynamic risk control index. , as well as These are the weighting coefficients. Delay in identifying abnormal operations For the operation response threshold, This represents the actual number of risk scenarios that trigger permission adjustments. The total number of risk scenarios. For the sandbox false positive rate, For the incidence of safety incidents.

[0005] As a further aspect of the present invention: the zero-code customization center parses user instructions using natural language processing technology, automatically generates corresponding business process logic and data collection forms, and supports typical business scenario templates for the manufacturing and financial industries.

[0006] As a further aspect of the present invention: the cross-system semantic mapping unit realizes semantic association and format conversion between heterogeneous systems based on a knowledge graph, and its comprehensive optimization score calculation formula is as follows: ; in, To optimize the score for cross-system semantic mapping, The total number of data types to be converted. For the first The semantic association matching degree of the class data, with a value range of [0,1]. For the first The data type format is converted to power, with a value range of [0,1]. Optimize coefficients for federated learning.

[0007] As a further aspect of the present invention: the spatiotemporal collaborative scheduling unit dynamically selects a real-time transmission protocol or an asynchronous collaboration strategy based on the user's time zone distribution and service priority, and provides an operation impact pre-simulation function to reduce version conflicts.

[0008] As a further aspect of the present invention: the edge-cloud converged architecture has the ability to switch autonomously in case of network anomalies, automatically activates the edge autonomous mode when the network is interrupted, and performs data consistency verification through incremental synchronization algorithm after recovery.

[0009] As a further aspect of the present invention: the cross-system semantic mapping unit continuously optimizes the conversion accuracy through federated learning, supporting intelligent conversion of CAD drawings, ERP bills of materials, and MES production data in industrial formats.

[0010] As a further aspect of the present invention: the zero-code customization center provides three levels of atomic functional components, including atomic level, composition level and application level. The component composition follows the principle of no circular dependencies and has a depth of 5 layers. It supports the generation of business processes through natural language instruction parsing and provides semantic error correction suggestions when parsing fails.

[0011] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows: 1. This invention uses a distributed hash table based on the consistent hashing algorithm to achieve dynamic data sharding and load balancing among 2-256 nodes. It has automatic fault tolerance and replica migration capabilities. Edge nodes can autonomously switch to local operation when the network is interrupted to ensure business continuity. After recovery, it can quickly complete data consistency verification through incremental synchronization, effectively supporting the stable and efficient operation of large-scale cross-regional collaborative networks. 2. This invention achieves intelligent scheduling and precise collaboration of cross-system services through an adaptive collaborative engine. Based on knowledge graph and federated learning technology, the cross-system semantic mapping unit can accurately realize semantic association and format conversion between heterogeneous data. The spatiotemporal collaborative scheduling unit dynamically selects transmission strategies according to user time zone and service priority, which significantly reduces version confusion and operational conflicts in cross-regional collaboration and improves collaboration efficiency and order. 3. This invention deeply integrates zero-code customization with AI-driven operation and maintenance to achieve efficient configuration and intelligent control of distributed management networks. It automatically generates business processes and forms through natural language command parsing, allowing business personnel to directly participate in customization and greatly shortening deployment time. The AI ​​dynamic security module analyzes user behavior and network status in real time, and dynamically adjusts permission policies through quantitative risk index to achieve proactive early warning and protection, significantly enhancing overall network security and reducing operation and maintenance complexity. Attached Figure Description

[0012] Figure 1 This is a flowchart illustrating the zero-code customization process in an embodiment of the present invention; Figure 2 This is an adaptive collaboration flowchart in an embodiment of the present invention; Figure 3 This is a flowchart of the edge-cloud process in an embodiment of the present invention. Detailed Implementation

[0013] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0014] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0015] Please see the appendix Figure 1 - Appendix Figure 3 This invention discloses a private cloud-based collaborative management platform, comprising an edge-cloud fusion architecture, an AI dynamic security module, an adaptive collaborative engine, and a zero-code customization center, characterized in that: The edge-cloud converged architecture forms the distributed management center of the collaborative management network. It uses a quantum key distribution link based on the BB84 protocol to connect the edge computing nodes and the central cloud, with a transmission rate of 1Gbps and a key update cycle of 1 hour. A distributed hash table based on the consistent hashing algorithm is used to realize dynamic mapping and load balancing of data shards, supporting elastic expansion of 256 nodes and automatic fault tolerance. The AI ​​dynamic security module includes a CNN-LSTM hybrid security situation awareness model, a risk-adaptive permission mechanism, and an intelligent sandbox isolation unit. The adaptive collaboration engine includes a cross-system semantic mapping unit and a spatiotemporal collaborative scheduling unit, which are used to realize semantic collaboration and asynchronous collaborative management among multiple systems; The zero-code customization center includes an industry-specific business template library, an atomic functional component library, and a user intent recognition unit, supporting the generation of business processes and data forms through natural language commands.

[0016] In one embodiment of the present invention: the formula for the dynamic risk prevention and control index in the AI ​​dynamic security module is: ; in, This is a dynamic risk control index. These are the weighting coefficients. Delay in identifying abnormal operations For the operation response threshold, This represents the actual number of risk scenarios that trigger permission adjustments. The total number of risk scenarios. For the sandbox false positive rate, For the incidence of safety incidents.

[0017] In one embodiment of the present invention: the quantum key distribution link is constructed using the BB84 protocol, supports a transmission rate of 1Gbps or higher, has an automatic key update mechanism with an update cycle of ≤1 hour, generates a true random key through a quantum random number generator, and its resistance to quantum computing attacks meets the Level 2 or higher security standards of the State Cryptography Administration. The distributed hash table uses a consistent hashing algorithm to implement data sharding mapping, supports dynamic expansion of 2-256 physical nodes, each data shard corresponds to 3 redundant replicas, and automatically triggers replica migration when a node fails. The data mapping update time is ≤100ms. Federated learning optimization coefficient It is a correction parameter used in the cross-system semantic mapping unit to dynamically improve the accuracy of heterogeneous data conversion. Its essence is to calculate it based on the weighted calculation of the model contribution of each participating business system. It is used to further optimize the accuracy of cross-system data mapping on the basis of semantic association matching and format conversion. It is especially suitable for multi-node distributed collaboration scenarios. Its value range is [0, 0.5]. It is calculated based on the weighted calculation of the model contribution of the participants. Contribution = local data volume × model accuracy. It is updated once a day at 2:00 by the federated average algorithm. The initial value is 0.1.

[0018] In one embodiment of the present invention: the security situation awareness model adopts a CNN-LSTM hybrid network structure, and the training data includes user operation logs, network traffic characteristics and known threat samples from the past 90 days. It supports online incremental learning and completes model iteration every 48 hours based on newly collected security event data. Natural language commands support bilingual input in Chinese and English, covering eight core command syntaxes including process creation, node configuration, and permission settings. Based on testing with 10,000 industry command samples, the parsing accuracy is ≥95%. When parsing fails, semantic error correction suggestions are returned, and manual intervention to correct commands is supported. Example 1

[0019] New energy vehicle enterprise supply chain collaborative management scenario: Basic configuration and architecture deployment: Edge-to-cloud node deployment: Edge computing nodes are deployed at the automaker's headquarters, three regional production bases (North China, East China, and South China), and two core component suppliers. They are connected via a quantum key distribution link based on the BB84 protocol, maintaining a stable transmission rate of 1Gbps. Key updates are automatically completed at 1:00 AM daily. A distributed hash table is constructed using a consistent hashing algorithm, with 32 nodes configured: 8 at headquarters, 6 at each production base, and 2 at each supplier. Each supply chain data shard corresponds to 3 redundant replicas, and the data mapping update time is ≤100ms in the event of a node failure. Module initialization settings: The AI ​​dynamic security module imports supply chain-related operation logs, network traffic characteristics, and known threat samples from the past 90 days, completes the training of the CNN-LSTM security situation awareness model, and sets the operation response threshold. =2.5 seconds, the adaptive collaborative engine initializes the federated learning optimization coefficients by associating the ERP system, MES system, and SRM system through the knowledge graph. =0.18, the zero-code customization center loads a template library dedicated to the manufacturing supply chain and enables bilingual natural language parsing in Chinese and English; Specific operating methods and content: Cross-system collaborative operation: Operation trigger: The MES system at the East China production base shows that the inventory of a certain batch of motor parts is below the safety threshold, requiring the triggering of an emergency procurement process; Semantic Mapping and Data Transformation: The MES system automatically sends inventory warning data to the adaptive collaborative engine. The cross-system semantic mapping unit, based on a knowledge graph, semantically associates MES production inventory data with ERP procurement demand data. The semantic association matching degree of the first type of data... Format converted to power Simultaneously, it links to SRM system supplier qualification data, and the semantic correlation matching degree of the second type of data (SRM qualification data). Format converted to power Combined with federated learning optimization coefficient =0.18, substituting into the cross-system semantic mapping comprehensive optimization score formula, we get... The actual conversion rate reached 97.5%, realizing the automatic conversion of inventory data into procurement demand data without the need for manual secondary entry; Spatiotemporal Cooperative Scheduling: The Spatiotemporal Cooperative Scheduling Unit is the core functional module of the Adaptive Cooperative Engine. It aims to solve the collaboration conflict problem caused by "time difference" and "spatial distribution" in multi-system and cross-regional collaboration. It achieves efficient and conflict-free cross-node business collaboration through dynamic strategy adjustment. Specifically, it can be understood from the "spatiotemporal dual dimensions". The Adaptive Cooperative Engine analyzes the user's time zone and business priority, selects the real-time transmission protocol, and synchronizes the converted procurement demand data to the ERP system and the supplier SRM system. It also triggers the operation impact simulation function to simulate the impact of the purchase order on the production plan and inventory. After confirming that there is no version conflict, it generates the purchase order number "CG20240508001". Zero-code business process customization operation command input: Supply chain management specialists input Chinese commands through the zero-code customization center, such as "Create an emergency procurement approval process for motor parts, including four nodes: procurement specialist submission, procurement manager review, finance confirmation of payment, and supplier order acceptance. Each node has a 2-hour approval time limit, with automatic reminders for exceeding the time limit." Command parsing and process generation: The zero-code customization center's user intent recognition unit, combined with natural language processing technology, achieves a 96% accuracy rate in parsing commands and generates process logic within 12 seconds: Procurement specialists submit purchase orders in the ERP system → The system automatically pushes a review notification to the procurement manager → After the procurement manager approves the order, it is pushed to finance → After finance confirms the payment method, it is pushed to the supplier → After the supplier accepts the order, the process loop is closed. Each node has a 2-hour countdown, and an SMS + system message reminder is automatically sent when there are 30 minutes left. Component combination and verification: During the process generation process, the system calls atomic-level components, composite-level components, and application-level components. The component combination depth is 4 layers, and there are no circular dependencies. The process is activated immediately after automatic verification, and dynamic security protection and abnormal operation monitoring are implemented. A supplier's personnel attempted to log into the SRM system outside of working hours (22:00) to download the lowest purchase prices for all motor components over the past year (sensitive data). The AI ​​dynamic security module's security situation awareness model monitored the operation in real time and identified the anomaly within 1.8 seconds. The anomaly detection was delayed. =1.8 seconds, triggering a risk warning; Risk prevention and control index calculation and response: taking weighting coefficients. , , Total number of current risk scenarios The actual number of risk scenarios that trigger permission adjustments Sandbox misjudgment rate security incident incidence Substitute into the dynamic risk prevention and control index formula The system determined the risk level to be "medium-high" and immediately froze the sensitive data download permissions of the supplier's personnel. They were required to pass two-factor authentication of "account password + hardware key" before they could unlock the data. At the same time, a security log was generated and uploaded to the blockchain for evidence storage to ensure that the operation was traceable. Edge-cloud converged architecture applications: Edge autonomy switch: Due to a regional network failure, the South China production base lost connection with the central cloud. The edge node automatically detected the network anomaly and switched to edge autonomy mode within 10 seconds. Locally stored production progress data and quality inspection report data were normal and available. The production workshop could continue to enter the daily motor assembly progress to ensure uninterrupted production. Data synchronization recovery: After 2.5 hours, the network is restored. The edge nodes use an incremental synchronization algorithm to transmit only the assembly progress data of the 200 motors added during the interruption to the central cloud. The incremental data volume is 80MB. Data consistency verification is completed within 5 minutes to ensure data synchronization between the central cloud and the edge nodes, avoiding network bandwidth occupation and data redundancy caused by full synchronization. Example 2

[0020] Collaborative operation scenarios for chain retail enterprises: Basic configuration and architecture deployment: Edge-to-cloud node deployment: Edge computing nodes are deployed at the retail enterprise headquarters (central cloud) and 10 core stores (5 in first-tier cities and 5 in second-tier cities). They are connected via a quantum key distribution link based on the BB84 protocol, with a transmission rate of 1Gbps and a key update cycle of 1 hour. A distributed hash table is constructed using a consistent hashing algorithm, with 20 nodes configured: 10 at the headquarters and 1 at each store. In practice, the nodes are allocated in integers as 2 for each store in first-tier cities and 1 for each store in second-tier cities. Each operational data shard corresponds to 3 redundant replicas. In the event of a node failure, the data mapping update time is ≤100ms. Module initialization settings: The AI ​​dynamic security module imports nearly 90 days of store operation logs, network traffic characteristics, and known threat samples, trains the CNN-LSTM security situation awareness model, and sets operation response thresholds. Within seconds, the adaptive collaborative engine connects the POS system, inventory management system, and membership management system through a knowledge graph, and initializes the federated learning optimization coefficients. The zero-code customization center loads a template library specifically for the retail industry and supports Chinese natural language input; Specific operation methods and content: Cross-system collaborative operation. Operation trigger: The POS system of a store in a first-tier city shows that the "bottled beverage" category is out of stock. The system automatically triggers a stockout warning. Semantic mapping and data transformation: The POS system sends the stockout data to the adaptive collaborative engine. The cross-system semantic mapping unit, based on a knowledge graph, associates the "replenishment demand" data format of the inventory management system with the "member preference" data of the membership management system. The semantic association matching degree of the first type of data. =0.96, format converted to power The semantic relevance matching degree of the second type of data (member preference data) Format converted to power Combined with federated learning optimization coefficient =0.15, substitute into the cross-system semantic mapping comprehensive optimization score formula The actual conversion rate reached 96.8%, and a "bottled beverage replenishment request form" was automatically generated and synchronized to the inventory management system. Spatiotemporal collaborative scheduling: The adaptive collaborative engine analyzes time zones and business priorities, selects asynchronous collaboration strategies, temporarily stores replenishment request orders in the cloud, and pushes reminders when the headquarters inventory administrator logs into the system during the next working period (9:00 am). At the same time, it provides an operation impact simulation function to simulate the impact of replenishing 150 bottles on store inventory and procurement costs. After confirming that there are no conflicts, it generates a replenishment order number "BH20240509001". Zero-code business process customization operation command input: The retail headquarters operations specialist inputs the Chinese command "Create monthly inventory process for stores, including three nodes: initial inventory by store staff, review by store manager, and random inspection by headquarters. The initial inventory time limit is 1 day, the review time limit is half a day, the random inspection ratio is 10%, and an inventory difference report will be automatically generated after the inventory is completed" through the zero-code customization center. Command parsing and process generation: The user intent recognition unit of the zero-code customization center parses the command, calls the "Store Inventory" basic template in the industry-specific business template library, and combines the "Time Limit Setting", "Ratio Configuration", and "Report Generation" components in the atomic functional component library to generate the process logic within 10 seconds: On the last day of each month, store staff enter the initial inventory data in the inventory management system → the system automatically... The system automatically pushes a review notification to the store manager, who must complete the review by 12 PM the following day. Headquarters randomly selects 10% of product categories for spot checks. After all nodes are completed, the system automatically compares the initial review, review, and spot check data, generating an inventory discrepancy report. The analysis process is error-free with 100% accuracy. The process is activated immediately after generation. Component combination and validation: The component combination depth is 3 levels with no circular dependencies. The system automatically validates and passes checks, supporting up to 10 stores using the inventory process simultaneously. Dynamic security protection and abnormal operation monitoring: In a second-tier city store, an employee attempted to modify a member's purchase record. The AI ​​dynamic security module monitored the operation in real time, identifying the anomaly within 2.2 seconds and triggering a risk warning. Risk control index calculation and response: Weighted coefficients are used. , , Total number of current risk scenarios The actual number of risk scenarios that trigger permission adjustments Sandbox misjudgment rate security incident incidence Substitute into the dynamic risk prevention and control index formula The system determined the risk level to be "medium" and immediately locked the employee's access to modify member data, requiring the store manager's approval before unlocking it. At the same time, the abnormal operation record was uploaded to the blockchain for evidence storage to prevent tampering. Edge-cloud converged architecture applications: Edge autonomy switching: When a store in a first-tier city experiences a network outage due to a typhoon, the edge node automatically detects the network status and switches to edge autonomy mode within 15 seconds. The local POS system continues to process payments normally, and store staff can check local inventory as usual, ensuring the normal operation of the store. Data synchronization and recovery: After 4 hours, the network is restored. The edge nodes use an incremental synchronization algorithm to transmit only the 80 transaction data and inventory change data during the interruption to the central cloud. Data consistency verification is completed within 3 minutes to ensure that the headquarters can keep abreast of store sales and inventory in real time, providing data support for subsequent replenishment and promotional strategy adjustments.

[0021] As can be seen from the above two sets of embodiments, this platform can effectively support the stable and efficient operation of large-scale cross-regional collaborative networks.

[0022] While the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Any variations and modifications can be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention, without departing from the scope of the invention, fall within the protection scope defined by the claims of the present invention.

Claims

1. A private cloud-based collaborative management platform, comprising an edge-cloud fusion architecture, an AI dynamic security module, an adaptive collaborative engine, and a zero-code customization center, characterized in that: The edge-cloud converged architecture constitutes the distributed management center of the collaborative management network. It uses a quantum key distribution link based on the BB84 protocol to connect the edge computing nodes and the central cloud, with a transmission rate of 1Gbps and a key update cycle of 1 hour. A distributed hash table based on the consistent hashing algorithm is used to realize dynamic mapping and load balancing of data shards, supporting elastic expansion of 256 nodes and automatic fault tolerance. The AI ​​dynamic security module includes a CNN-LSTM hybrid security situation awareness model, a risk-adaptive permission mechanism, and an intelligent sandbox isolation unit. The adaptive collaboration engine includes a cross-system semantic mapping unit and a spatiotemporal collaborative scheduling unit, which are used to realize semantic collaboration and asynchronous collaborative management among multiple systems. The zero-code customization center includes an industry-specific business template library, an atomic functional component library, and a user intent recognition unit, which supports the generation of business processes and data forms through natural language commands.

2. The private cloud collaborative management platform according to claim 1, characterized in that: The dynamic risk prevention and control index in the AI ​​dynamic security module is calculated using the following formula: ； in, This is a dynamic risk control index. , as well as These are the weighting coefficients. Delay in identifying abnormal operations For the operation response threshold, This represents the actual number of risk scenarios that trigger permission adjustments. The total number of risk scenarios. For the sandbox false positive rate, For the incidence of safety incidents.

3. The private cloud collaborative management platform according to claim 2, characterized in that: The zero-code customization center uses natural language processing technology to parse user commands and automatically generate corresponding business process logic and data collection forms, supporting typical business scenario templates in the manufacturing and financial industries.

4. The private cloud collaborative management platform according to claim 3, characterized in that: The cross-system semantic mapping unit realizes semantic association and format conversion between heterogeneous systems based on knowledge graphs, and its comprehensive optimization score calculation formula is as follows: ； in, To optimize the score for cross-system semantic mapping, The total number of data types to be converted. For the first The semantic association matching degree of the class data, with a value range of [0,1]. For the first The data type format is converted to power, with a value range of [0,1]. Optimize coefficients for federated learning.

5. A private cloud-based collaborative management platform according to claim 4, characterized in that: The spatiotemporal collaborative scheduling unit dynamically selects real-time transmission protocols or asynchronous collaboration strategies based on user time zone distribution and service priorities, and provides an operation impact pre-simulation function to reduce version conflicts.

6. The private cloud collaborative management platform according to claim 5, characterized in that: The edge-cloud converged architecture has the ability to switch autonomously in case of network failure. It automatically activates the edge autonomous mode when the network is interrupted, and performs data consistency verification through incremental synchronization algorithm after recovery.

7. A private cloud-based collaborative management platform according to claim 6, characterized in that: The cross-system semantic mapping unit continuously optimizes the conversion accuracy through federated learning, supporting intelligent conversion of CAD drawings, ERP bills of materials, and MES production data in industrial formats.

8. A private cloud-based collaborative management platform according to claim 7, characterized in that: The zero-code customization center provides three levels of atomic functional components, including atomic level, composition level and application level. The component composition follows the principle of no circular dependencies and has a depth of 5 layers. It supports the generation of business processes through natural language instruction parsing and provides semantic error correction suggestions when parsing fails.

Images