Information processing system

Abstract

The application provides an information processing system. An information processing system, characterized by comprising: a processor; wherein the processor is configured to: save a network introduction condition into a database; based on the saved data, generate a prompt text for indicating a required item of a network composition graph by using a generative artificial intelligence model; input the prompt text into the generative artificial intelligence model to generate an optimal network composition graph.

Description

Technical Field

[0001] The technology disclosed herein relates to an information processing system. Background Technology

[0002] Japanese Patent Application Publication No. 2022-180282 discloses a method for controlling a role-based chatbot executed by at least one processor. The method includes the following steps: receiving a user's speech; adding the user's speech to a prompt word, the prompt word containing instruction statements associated with an explanation of the chatbot's role; encoding the prompt word; and inputting the encoded prompt word into a language model to generate a chatbot response to the user's speech.

[0003] When enterprises or organizations build and expand their networks, on-site personnel often lack systematic network design knowledge and experience, making it difficult to accurately design the network structure and generate network configuration diagrams based on the actual network deployment status. In existing technologies, engineers with specialized network knowledge typically need to manually analyze the network deployment status, design network address schemes and routing control strategies, and manually draw the network configuration diagram accordingly. This not only consumes a significant amount of manpower and time but is also prone to design flaws, network redundancy, or potential failure risks due to human error or misunderstanding. Furthermore, existing systems are mostly static configuration tools, unable to dynamically optimize the network configuration diagram based on constantly changing network deployment status, nor can they fully utilize generative artificial intelligence models to intelligently assist the design process. Therefore, how to provide a system that can automatically acquire and utilize network deployment status data, automatically generate the optimal network configuration diagram using generative artificial intelligence models, and simultaneously provide users with relevant information such as network address design and routing control, is the technical challenge this invention aims to address. Summary of the Invention

[0004] To address the aforementioned technical challenges, this invention proposes an information processing system. This system includes a processor configured to store network import status in a database. By structurally storing and uniformly managing the network import status, a reliable data foundation is provided for subsequent intelligent analysis. The processor is further configured to generate prompt text based on the stored data using a generative artificial intelligence model. This prompt text, containing information such as network size, device type, topology hierarchy, security requirements, and bandwidth requirements, guides the generative artificial intelligence model to output a compliant network configuration scheme according to predetermined design rules and best practices. The processor is also configured to input the prompt text into the generative artificial intelligence model to generate an optimal network configuration diagram, thereby achieving automatic generation and optimization of the network configuration diagram without requiring the user to possess professional network design capabilities.

[0005] Furthermore, to provide users with more comprehensive network design assistance, the processor in this invention is also configured to provide information related to network address design and routing control. This allows users to obtain corresponding address planning suggestions and routing strategy explanations while viewing the network structure diagram, reducing the risk of configuration errors and improving the consistency and rationality of the overall network design. Further, the processor is configured to dynamically adjust the prompt text based on the network's deployment status. When the network deployment status changes (e.g., adding or removing devices, changing bandwidth requirements, or changing security policies), the processor can automatically update the prompt text and re-invoke the generative artificial intelligence model to generate a new network structure diagram. This enables continuous optimization and iterative updates to the network design, effectively solving the problem in existing technologies where network design is difficult to flexibly adjust according to actual deployment conditions.

[0006] A "system" refers to an overall device or platform composed of multiple interrelated hardware components and software modules, used to perform functions such as storing, processing, analyzing network import status, and generating network structure diagrams.

[0007] "Processor" refers to a hardware device or its virtualized instance that is capable of executing program instructions to perform data processing and logical operations, including but not limited to a central processing unit (CPU), a graphics processing unit (GPU), an application-specific integrated circuit (ASIC), or a virtual computing unit in a cloud computing environment.

[0008] "Network import status" refers to various status and configuration information related to the construction, deployment, or expansion of the target network, including but not limited to the type, quantity, model, connection relationship, bandwidth requirements, service requirements, security policies, and deployment stage of network devices.

[0009] A "database" refers to a storage system used to store and manage network import status data and related information in a structured or semi-structured form, including relational databases, non-relational databases, or distributed data storage systems.

[0010] "Generative artificial intelligence models" refer to artificial intelligence models trained through machine learning or deep learning methods that can automatically generate corresponding output content based on input prompt text, including but not limited to large language models, text generation models, or specialized generative models that can generate structured network design results.

[0011] "Prompt text" refers to textual instructions or descriptive information that is constructed by the processor and input into the generative artificial intelligence model, instructing the generative artificial intelligence model to generate network structure diagrams or related design content according to predetermined intentions and constraints.

[0012] A "network structure diagram" is a graphical or logical representation of a network structure and its connections. It includes information about each device node, link connection, hierarchical structure, and related attributes in the network, and can be used to guide the deployment, operation, and management of the network.

[0013] "Optimal network configuration" refers to a network configuration that, under given network implementation conditions and design constraints, has preferred performance or a compromise solution relative to preset evaluation criteria (such as reliability, scalability, cost, performance, security, etc.).

[0014] "Network address design" refers to the design process of planning and allocating IP addresses or other network address resources for various devices, subnets and business scenarios in a network, including network segmentation, address pool planning, subnet mask setting and address allocation strategies.

[0015] "Routing control" refers to the process of configuring and managing routing protocols, routing policies, and forwarding paths to enable data forwarding between different subnets or networks within a network. This includes static route configuration, dynamic routing protocol settings, and routing policy optimization. Attached Figure Description

[0016] Figure 1 This is a conceptual diagram illustrating an example of the configuration of the data processing system according to the first embodiment.

[0017] Figure 2 This is a conceptual diagram illustrating an example of the main functions of the data processing apparatus and smart device according to the first embodiment.

[0018] Figure 3 This is a conceptual diagram illustrating an example of the configuration of the data processing system according to the second embodiment.

[0019] Figure 4 This is a conceptual diagram illustrating an example of the main functions of the data processing device and smart glasses according to the second embodiment.

[0020] Figure 5 This is a conceptual diagram illustrating an example of the configuration of the data processing system according to the third embodiment.

[0021] Figure 6 This is a conceptual diagram illustrating an example of the main functions of the data processing apparatus and head-mounted terminal according to the third embodiment.

[0022] Figure 7 This is a conceptual diagram illustrating an example of the configuration of the data processing system according to the fourth embodiment.

[0023] Figure 8 This is a conceptual diagram illustrating an example of the main functions of the data processing device and robot according to the fourth embodiment.

[0024] Figure 9 This represents an emotion map that maps multiple emotions.

[0025] Figure 10 This represents an emotion map that maps multiple emotions.

[0026] Figure 11 This is a sequence diagram illustrating the processing flow of the data processing system of the first embodiment.

[0027] Figure 12 This is a sequence diagram illustrating the processing flow of the data processing system in Application Example 1.

[0028] Figure 13 This is a sequence diagram illustrating the processing flow of the data processing system of the second embodiment.

[0029] Figure 14 This is a sequence diagram illustrating the processing flow of the data processing system in Application Example 2. Detailed Implementation

[0030] Hereinafter, an example of an implementation of the system according to the present disclosure will be described with reference to the accompanying drawings.

[0031] First, let me explain the terminology used in the following instructions.

[0032] In the following embodiments, the processor (hereinafter referred to as "processor") with reference numerals may be a single computing device or a combination of multiple computing devices. Furthermore, the processor may be a single computing device or a combination of multiple computing devices. Examples of computing devices include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), APU (Accelerated Processing Unit), etc.

[0033] In the following embodiments, RAM (Random Access Memory), as indicated in the figures, is a memory that temporarily stores information and is used as working memory by the processor.

[0034] In the following embodiments, the memory, as indicated by the reference numerals, is one or more non-volatile storage devices that store various programs and parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), disks (e.g., hard disks), or magnetic tapes.

[0035] In the following embodiments, the communication I / F (Interface) with reference numerals is an interface that includes a communication processor and an antenna, etc. The communication I / F is responsible for communication between multiple computers. As an example of a communication specification applicable to the communication I / F, wireless communication specifications such as 5G (5th Generation Mobile Communication System), Wi-Fi (wireless fidelity) (registered trademark), or Bluetooth (registered trademark) can be listed.

[0036] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B". That is, "A and / or B" means that it can be only A, only B, or a combination of A and B. Furthermore, in this specification, when "and / or" connects to express more than three items, the same interpretation as "A and / or B" applies.

[0037] First Implementation Method Figure 1 An example of the configuration of the data processing system 10 according to the first embodiment is shown.

[0038] like Figure 1 As shown, the data processing system 10 includes a data processing device 12 and an intelligent device 14. A server can be cited as an example of the data processing device 12.

[0039] The data processing apparatus 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" as understood in this disclosure. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, the database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0040] The smart device 14 includes a computer 36, a receiving device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. In addition, the receiving device 38, output device 40, camera 42, and communication I / F 44 are also connected to the bus 52.

[0041] The receiving device 38 includes a touchscreen 38A and a microphone 38B, and receives user input. The touchscreen 38A receives user input via touch by detecting contact with an indicator (e.g., a pen or finger). The microphone 38B receives user input via sound by detecting the user's voice. The control unit 46A in the processor 46 sends data representing the user input received by the touchscreen 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data representing the user input.

[0042] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting data in a form perceptible to the user 20 (e.g., sound and / or text). The display 40A displays visual information such as text and images according to instructions from the processor 46. The speaker 40B outputs sound according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0043] Communication I / F44 is connected to network 54. Communication I / F44 and 26 are responsible for sending and receiving various information between processor 46 and processor 28 via network 54.

[0044] Figure 2 The diagram shows an example of the main functions of the data processing device 12 and the smart device 14.

[0045] like Figure 2 As shown, in the data processing apparatus 12, specific processing is performed by the processor 28. A specific processing program 56 is stored in the memory 32. The specific processing program 56 is an example of a "program" as understood in this disclosure. The processor 28 reads the specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. Specific processing is implemented by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0046] The memory 32 stores a data generation model 58 and an emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290. The specific processing unit 290 can use the emotion-specific model 59 to infer the user's emotions and perform specific processing based on the user's emotions. In the emotion inference function (emotion-specific function) using the emotion-specific model 59, various inferences and predictions related to the user's emotions are performed, including inferences and predictions of the user's emotions, but this is not limited to this example. Furthermore, emotion inference and prediction may also include, for example, emotion analysis (parsing).

[0047] In the smart device 14, the processor 46 performs the acceptance output processing. The memory 50 stores the acceptance output program 60. The acceptance output program 60 is used in conjunction with the data processing system 10 and the specific processing program 56. The processor 46 reads the acceptance output program 60 from the memory 50 and executes the read acceptance output program 60 on the RAM 48. The acceptance output processing is implemented by the processor 46 acting as the control unit 46A according to the acceptance output program 60 executed on the RAM 48. Furthermore, the smart device 14 has the same data generation model and emotion-specific model as the data generation model 58 and the emotion-specific model 59, and these models can also be used to perform the same processing as the specific processing unit 290. The acceptance output processing is implemented by the processor 46 acting as the control unit 46A according to the acceptance output program 60 executed on the RAM 48.

[0048] Alternatively, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device (e.g., a generation server) may have the data generation model 58. In this case, the data processing device 12 obtains the processing results (prediction results, etc.) using the data generation model 58 by communicating with the server device that has the data generation model 58. Furthermore, the data processing device 12 may be a server device or a user-held terminal device (e.g., a mobile phone, robot, home appliance, etc.). Next, an example of the processing of the data processing system 10 of the first embodiment will be described.

[0049] Example 1 The flow of a specific process in Example 1 will be described. Each part of the system described below is implemented by the data processing device 12 and the smart device 14. Furthermore, the data processing device 12 is referred to as the "server," and the smart device 14 is referred to as the "terminal."

[0050] As communication networks become increasingly complex in scale and structure, network topology design, resource deployment planning, and configuration scheme formulation rely more and more on the experience and knowledge of professionals. Currently, network engineers typically analyze communication resource information manually based on tabular data, static documents, or simple graphical tools, and then draw network structure diagrams or configuration schemes accordingly. This approach has the following problems: (1) The computer's processing of user input of network import status and business requirements is mostly static rule matching, lacking the ability to comprehensively reason about complex constraints and historical data, which makes it difficult for the generated network composition information to reflect the actual needs in a timely and accurate manner. (2) In existing systems, servers mostly execute fixed template rendering or preset scripts, failing to effectively combine user input and stored data with generative artificial intelligence models. They are unable to dynamically construct high-quality prompt statements based on communication resource information and usage scenarios, thus limiting the performance of generative artificial intelligence models in network design tasks. (3) In terms of output results, the server often provides network design results in the form of text or static images. It lacks the ability to automatically extract nodes and connections from the model output and construct structured visualization data, which makes it difficult for the front-end visualization components to perform efficient rendering and interaction, and increases the cost of manual processing. (4) For multi-round interaction and iterative optimization, existing systems generally cannot manage the user's additional or corrected inputs and the existing network import status in a unified manner. The server lacks a computer processing mechanism to dynamically update prompt statements and internal data, making it difficult to form a closed-loop automated optimization process.

[0051] Therefore, there is a need for a computer implementation scheme that can improve data structuring processing, automatic generation of prompts, generative artificial intelligence model invocation, and visualization of data construction processes on the server side. This would enable the server to more efficiently and intelligently link user input with model inference results in a general hardware and software environment, thereby improving the computer processing power and overall system performance in the process of network graph generation and network design support.

[0052] The specific processing performed by the specific processing unit 290 of the data processing apparatus 12 in Embodiment 1 is achieved by the following means.

[0053] In this invention, the server includes means for obtaining and structurally recording information related to communication resources and processing conditions from a user via a communication terminal; means for automatically generating prompt statements expressing instructions for a generative artificial intelligence model based on the structured recorded information; means for inputting the prompt statements into the generative artificial intelligence model and obtaining structural information representing the connection and arrangement relationships of the communication resources; means for extracting elements representing communication devices and their connection relationships from the structural information and converting them into structured data for visualization; and means for updating the recorded information and the prompt statements based on additional or corrected input from the user and repeating the above processing. Thus, on the server side, in a computer-executable manner, the network import status input by the user can be closely integrated with the inference process of the generative artificial intelligence model, enabling dynamic generation and adjustment of prompt statements, improving the consistency between model output results and actual needs, and automatically constructing structured topology data suitable for front-end graphics rendering. This improves the computer technical performance of the network design support system in terms of data processing efficiency, result accuracy, and multi-round interactive optimization.

[0054] A "system" refers to an entire system consisting of one or more information processing devices, storage devices, and communication devices, used to perform collaborative processing of acquiring, processing, and outputting information.

[0055] A "server" refers to an information processing device that provides computing and data management functions in a network. It can send and receive data with a communication terminal through a communication network and execute programs to achieve the various processes described in this invention.

[0056] "Communication terminal" refers to an electronic device that allows users to input, browse, and interact with information, including but not limited to computing devices, mobile terminals, or other devices with network communication functions.

[0057] "User" refers to the entity that inputs information related to communication resources and processing conditions into the system through a communication terminal and receives the system's output results. It can be an individual or an organization.

[0058] "Communication resources" refers to the general term for all kinds of resources used to build and operate communication networks, including communication devices, links, address space, and related configuration parameters.

[0059] "Communication device" refers to a hardware unit in a communication network used to realize functions such as data transmission and reception, exchange, routing, forwarding or security control, including but not limited to routing devices, switching devices, protection devices and access devices.

[0060] "Processing conditions" refers to the collective constraints, requirements, or expected goals related to the design, deployment, or operation of communication resources, such as the number of users, service type, security level, redundancy requirements, and performance requirements.

[0061] "Information storage device" refers to a storage medium and its control components used to record, maintain and retrieve information obtained through server processing. It can be a database system or other device with persistent storage function.

[0062] "Structured information" refers to information organized and represented in a predetermined data structure or format, enabling computers to process and retrieve the information in a programmed manner, such as information recorded using key-value pairs, record sets, or hierarchical data structures.

[0063] "Generative artificial intelligence models" refer to artificial intelligence models that are trained using data-driven methods and can automatically generate text, structured data, or other forms of output based on input data, including but not limited to natural language generation models based on deep learning.

[0064] "Prompt statements" refer to text data used to provide task instructions, constraints, and background information to generative artificial intelligence models. They describe the processing content and output format expected to be performed by the model through natural language or other forms of expression.

[0065] "Instructions" refers to the set of task descriptions, constraints, contextual information, and output requirements conveyed to the generative artificial intelligence model through prompt statements to guide the model in generating results.

[0066] "Construction information" refers to structured information used to represent the connection and arrangement relationships between communication resources, including at least data representing communication devices and the connection relationships between communication devices.

[0067] "Connectivity" refers to the general term for the connection methods and topological relationships between various communication devices in a communication network at the physical or logical layer, such as link connections, forwarding paths, or tunnel relationships.

[0068] "Layout relationship" refers to the distribution and relative position of communication resources in the network structure or hierarchy, including the configuration relationship of each communication device in different network layers, regions or functional partitions.

[0069] "Structured data" refers to data in a format that is represented by a predetermined data model and is suitable for program processing and visualization. It contains fields for representing nodes and connections, such as sets of node information and sets of connection information.

[0070] "Node information" refers to the attribute information used in structured data to represent various communication devices or network entities, including identifiers, types, functional roles, and attributes related to visualization.

[0071] "Connection information" refers to the attribute information used in structured data to represent the connection relationship between nodes, including the connection start point, connection end point, and connection type or direction.

[0072] "Visualization" refers to the process of presenting constituent information or structured data in graphical, chart, or other intuitive forms so that users can understand and analyze the network structure on communication terminals.

[0073] "Additional input" refers to new requirements, constraints, or supplementary information that users input based on the results obtained after the system has generated the initial structural information.

[0074] "Corrected input" refers to input from users that modifies, corrects, or optimizes previously entered information or generated constituent information, used to update the information recorded in the system and the results of subsequent processing.

[0075] In one embodiment of the present invention, the server, as an information processing device, is equipped with general-purpose hardware resources and an operating system environment. The server may employ a multi-core central processing unit, a graphics processing unit, internal memory, and external memory. The server may run a general-purpose operating system, such as a UNIX-like operating system or other server operating systems. The server may deploy application service middleware, such as a web application server, a script execution environment, and a database management system. The server can also interact with terminals and external computing resources via a network interface.

[0076] In an embodiment of the invention, the server runs programs to implement the various functions of the invention. The server loads various functional modules within the process space through an application framework, including a request receiving module, a data structuring module, an information storage module, a prompt statement generation module, a generative artificial intelligence model invocation module, a constituent information parsing module, and a visualization data generation module. The server transfers data structures between these modules via an internal bus or in-process function calls, thereby achieving automated processing of communication resource information.

[0077] In an embodiment of the present invention, the terminal serves as an information presentation and input device, equipped with a display unit, an input unit, and a communication unit. The terminal can run a general-purpose operating system, such as a mobile terminal operating system or a client operating system. The terminal can execute a browser program or a client application to send user input to the server and receive structured data and text descriptions output by the server. The terminal uses a graphics rendering engine to present a network diagram and related text descriptions on the display unit.

[0078] In this embodiment of the invention, the user, as the main user, inputs communication resource information and processing conditions through a graphical interface on a terminal. The user can input information such as the type and quantity of communication devices, deployment scenarios, performance requirements, security requirements, and redundancy requirements. After viewing the network structure diagram, the user can also input additional requirements or corrections so that the server can iteratively process the data based on existing data.

[0079] In an embodiment of this invention, the server uses a relational database management system as an information storage device. The server can employ relational database software to store data such as network import status, user configuration requests, generative artificial intelligence model call records, and topology information in the form of tables on external storage. The server, through a database access interface, converts user-inputted communication resource information and processing conditions into structured records and writes them into pre-designed data tables. The server uses primary key identifiers to associate the network import status of different rounds with the corresponding generation results, thereby achieving unified management of historical and incremental information in subsequent processing.

[0080] In this embodiment of the invention, the server converts the communication resource information and processing conditions input by the user into structured information. The server parses the text data and option data sent by the terminal, generates a unified type identifier for each type of communication device, and performs data type conversion and boundary checks on fields such as quantity, performance, and security. The server constructs a record object in memory containing multiple fields, such as a scene category field, a device type list field, a capacity requirement field, and a redundancy strategy field. The server maps this record object to a database table structure and performs a write operation, thereby saving the network import status in a unified format.

[0081] In an embodiment of the invention, the server generates a prompt statement based on recorded structured information. The server uses templated string processing logic in memory to convert structured fields into natural language fragments. For example, the server iterates through fields representing device type and quantity, mapping types such as "routing device," "switching device," and "protection device" to natural language descriptions, and generates quantity descriptions such as "1 unit" and "2 units" based on the quantity field. The server can also add constraint descriptions such as "remote access required" and "redundant links required" based on processing condition fields. The server then concatenates these natural language fragments into a complete prompt statement according to a pre-defined template order.

[0082] In an embodiment of the present invention, the prompt statement generated by the server may have the following example form: "You are a senior network architect."

[0083] Based on the following network import status, please generate the optimal network topology scheme for a new office.

[0084] - Scene: New office; - Planned equipment to be introduced: 1 enterprise-grade router, 2 Layer 2 switches, and 1 firewall; - Number of employees: Approximately 50; - Requirement: Support VPN remote access and provide basic secure access control.

[0085] please: 1) Describe the recommended network layer structure (core layer / aggregation layer / access layer); 2) Explain the role of each piece of equipment; 3) Describe in detail the connection relationships between the devices; 4) If possible, please output a topology description including nodes and connections for subsequent visualization. In another embodiment of the present invention, the prompt statement generated by the server during iterative optimization may have the following example form: "Based on the network topology previously generated for the new office (including 1 router, 2 switches, and 1 firewall), please perform the following incremental optimizations:" - Add an internet backup line; - Add a redundant firewall to form a high-availability cluster with the existing firewalls; - Explain the failover strategy for primary / backup lines and dual firewalls.

[0086] Please output the updated network topology description and provide an explanation of the new topology structure, making the newly added devices and links clearly visible. In an embodiment of this invention, the server invokes a generative artificial intelligence model to generate network structure information. The server can deploy the generative artificial intelligence model within the same computing environment or invoke external inference services over the network. The server provides the generative artificial intelligence model with prompts and optional control parameters, including maximum output length, generation temperature, sampling strategy, etc. The server receives text and structured information from the model's output. By parsing the text returned by the model, the server extracts information about communication devices, connections, and hierarchical structures.

[0087] In an embodiment of this invention, the server employs a neural network architecture based on a multi-layer self-attention mechanism as a generative artificial intelligence model. This model can use an encoder-decoder structure or a unidirectional autoregressive structure. Each layer of the model includes a multi-head self-attention sublayer and a feedforward network sublayer, and intermediate activation values ​​are stabilized through residual connections and normalization operations. The model uses embedding vectors to represent the input prompts, assigning a high-dimensional vector to each tag and distinguishing different positions through positional encoding. During inference, the model assigns higher attention weights to key terms in the prompts, such as "router," "switch," "firewall," "VPN," "redundancy," and "backup line," thereby paying more attention to these technical elements when generating network composition information.

[0088] In an embodiment of this invention, the server uses a pre-trained generative artificial intelligence model. During the training phase, the model uses large-scale network design documents, configuration examples, and topology descriptions as training data. The server employs a cross-entropy loss function as an error metric during training and updates the model weights using a gradient descent-type optimization algorithm. The server performs data augmentation on the training data, such as perturbing the order of network devices and performing synonym substitutions on some descriptions, to improve the model's robustness to different representations. During the training phase, the server improves training efficiency through batch processing and parallel computation, and during the inference phase, the weights are fixed to ensure the stability of the inference results.

[0089] In an embodiment of the present invention, the server extracts a data structure suitable for visualization from the output of the generative artificial intelligence model. The server can locate structured fragments in the model output based on keyword patterns or marker symbols, such as statements containing expressions like "node," "connection," and "from…to…." The server can parse these fragments according to predetermined syntax rules, mapping information such as "device name," "device type," and "connection endpoint" into internal node information and connection information. The server constructs a data structure in memory containing a set of nodes and a set of connections. The server records each node as a record containing an identifier, a type label, and a display label, and records each connection as a record containing a source node, a target node, and optional attributes.

[0090] In an embodiment of the invention, the server converts internal node information and connection information into visual data that can be recognized by the front end. The server can organize the set of node information and connection information into a tree-like or graph-like data structure, and assign layout attributes to each node, such as hierarchy level or coordinate estimation. The server can use a simple hierarchical layout algorithm to place routing devices, protection devices, and switching devices at different levels according to the roles of the communication devices, thereby making it easier to generate a clear network structure diagram on the client side. The server encapsulates the above data as structured data in the response and sends it to the terminal through the communication interface.

[0091] In an embodiment of the present invention, after receiving structured data returned by the server, the terminal performs graphics rendering locally. The terminal can use a graphics library to map node information to graphic elements and connection information to line elements, and lay them out according to the hierarchical or coordinate information provided by the server. The terminal presents the network structure diagram on the display unit and displays text descriptions provided by the server around the graphics, thereby enabling the user to intuitively understand the network design scheme.

[0092] In an embodiment of this invention, after viewing the network structure diagram, the user can input additional requests or modifications via a terminal. The user can input natural language descriptions such as "add redundancy," "improve security level," "expand user count," or "add branch nodes." Upon receiving this input, the server correlates it with the existing network import status and updates the record in the database. Based on the updated record, the server regenerates a prompt statement, explicitly indicating that this request is an incremental modification based on the existing network structure. By repeatedly calling the generative artificial intelligence model and regenerating the visualization data, the server enables the user to obtain a progressively optimized network structure diagram through multiple rounds of interaction.

[0093] In this embodiment of the invention, the server achieves centralized management and efficient processing of communication resource information within the server through a structured record and dynamic prompt statement generation mechanism. The server no longer simply outputs static images using fixed templates, but instead, after receiving user input, it utilizes a complete data flow—data structuring, prompt statement generation, model invocation, result parsing, and visualization data generation—to enable each network design request to perform comprehensive reasoning based on historical records and current conditions. The server reuses existing records in each generation process, avoiding repetitive manual input and static analysis, thereby improving computational efficiency and data utilization.

[0094] In this embodiment of the invention, the server improves the input quality of the generative artificial intelligence model through a structured prompt generation method. The server employs different combinations of natural language fragments and templates for different types of communication devices and different scenario conditions, making the prompts more stable and explicit in terms of vocabulary selection and sentence structure. By doing so, the server reduces the model's attention expenditure on unnecessary details, allowing the model to focus on processing tags related to network topology and constraints during self-attention operations, thereby improving the relevance and accuracy of the generated results within the same inference time. This structured data-driven prompt generation method significantly reduces the bias and inconsistency of the model's output compared to users directly inputting free text prompts.

[0095] In this embodiment of the invention, the server automatically parses the model output to construct a visual data structure, reducing the manual work of transcribing the model output into graphical format. The server performs pattern matching and syntactic analysis on the model output, distinguishing between descriptive text and structural information, and converting the structural information into node and connection information. This processing avoids performing complex text parsing operations on the terminal side, concentrating the load in a centralized processing environment within the server, which is beneficial for coordinating processing resources, improving the overall system response speed, and reducing the processing burden on the terminal.

[0096] In an embodiment of this invention, the server constructs a traceable network design version chain by recording the prompts and output results of each model call. The server can store the correspondence between prompts and topology data in a database, and record timestamps and user identifiers. In subsequent iterations, the server can perform differential updates based on existing versions instead of completely regenerating the topology. The server can adjust only some nodes and connections according to user requests, thereby reducing the inference length and computational load of the generative artificial intelligence model, resulting in shorter processing time and reduced server computational load.

[0097] In another embodiment of the invention, the server can select generative artificial intelligence models of different scales or accuracies based on communication environment conditions and system resource status. For example, the server can use a model with smaller parameter scales to improve response speed under high load, and a model with larger parameter scales to improve generation accuracy under low load. The server can adjust the generation parameters according to the complexity of the current request, for example, limiting the generation length when processing simple topologies, and relaxing the generation length and adding contextual information when processing complex topologies spanning multiple regions and multi-layer structures. This adaptive model selection and parameter adjustment mechanism enables the server to dynamically balance processing speed and generation accuracy, improving the overall system's technical performance.

[0098] In this embodiment of the invention, the server transforms the communication resource design task into an efficient data flow processing process within the computer, rather than simply mechanizing the manual drawing process, through the aforementioned techniques such as structured data processing, automatic generation of prompts, generative artificial intelligence model reasoning, and visualization data construction. The server utilizes a neural network model to internally model large-scale historical cases and topological patterns, and comprehensively evaluates complex constraints in vector space through multi-layer self-attention operations, generating statistically superior topological structure suggestions. Because the server normalizes the model input and performs structured parsing of the output, this invention not only improves the rationality and consistency of network design results but also achieves technical improvements in processing speed, data management, computational efficiency, and communication load.

[0099] use Figure 11 The processing procedure is explained.

[0100] Step 1: Users input communication resource information and processing conditions through the terminal.

[0101] Users select or fill in fields such as communication scenario, type and quantity of communication devices, business requirements, security requirements, and redundancy requirements on the terminal interface. Input includes natural language descriptions (e.g., "Deploy 1 router, 2 switches, and 1 firewall in the new office, supporting 50 people working together and VPN access") and structured options (such as drop-down lists and checkboxes). Based on the user's actions, the terminal converts these inputs into internal data objects and performs preliminary validation of required fields locally. The terminal then prepares the processed data as a request body to send to the server.

[0102] Step 2: The terminal packages the user-input data and sends it to the server.

[0103] The terminal serializes the internal data object formed in step 1 into a request format, uses it as input, and generates a request message containing the scenario, device list, and required conditions. The terminal sends this request message to the server's preset interface address via a communication protocol. The terminal appends user identifier and session information to the request message. The output is a network request message, which is delivered to the server with the support of transport and application layer protocols.

[0104] Step 3: The server receives and parses the request data sent by the terminal.

[0105] The server takes network requests from the terminal as input. The server's network interface module parses the transport layer message and extracts the request body data from the application layer payload. The server uses a data parsing library to deserialize the request body into an internal data structure, obtaining scene fields, device fields, requirement fields, etc. The server performs field integrity checks and data type validation on the parsed results. The output is a set of parsed and preliminarily validated communication resource information and processing conditions.

[0106] Step 4: The server standardizes and structures communication resource information and processing conditions.

[0107] The server takes the raw fields from the previous step as input and performs data cleaning and standardization operations: mapping device names in different languages ​​or writing styles to a unified type, converting text-based numbers to numeric types, and converting Boolean requirements (such as whether a VPN is needed) into logical fields. The server constructs a unified format record object in memory, which includes scenario category, a list of device types and quantities, business scale parameters, security level, redundancy flags, etc. The server completes the above data processing through string manipulation and type conversion operations. The output is a structured network import status record.

[0108] Step 5: The server writes structured network import status records into the information storage device.

[0109] The server takes the structured records obtained in the previous step as input and calls the database access module to generate data insertion instructions. The server writes the records into a pre-designed storage table via a database connection. The table fields correspond to the scenario, device type list, requirement parameters, timestamp, and user identifier. The database performs the insertion operation and returns the primary key identifier. The server saves this primary key as an identifier for this network import status in the session context. The output is the persisted network import status record and its corresponding record identifier.

[0110] Step 6: The server reads network import status information related to the current session from the information storage device.

[0111] Using the record identifier obtained in the previous step as input, the server issues a query command to the database, reading all fields corresponding to that identifier. The server can simultaneously read related records previously submitted by the same user to construct a more comprehensive context. The database returns a record set containing the scenario, device list, and requirement conditions, which the server reconstructs into an operable data structure in memory. The output is one or more complete network import status data objects.

[0112] Step 7: The server generates prompt messages based on network import status.

[0113] The server takes the network import status data object from step 6 as input, iterates through the device list to generate descriptive fragments, and generates uniform natural language phrases for different device types (e.g., "1 enterprise-grade router", "2 Layer 2 switches", "1 firewall"), and generates constraint descriptions based on business requirement fields (e.g., "supports approximately 50 users", "requires VPN remote access", "requires basic security access control"). The server combines these phrases into multiple statement elements through string concatenation and template filling operations. The output is a set of natural language elements representing the scenario, devices, requirements, etc.

[0114] Step 8: The server combines the elements of the prompt statement into a complete prompt statement.

[0115] The server takes the natural language elements from the previous step as input, selects a predefined prompt template, and inserts scene descriptions, equipment descriptions, requirement constraints, and output format requirements into placeholder positions within the template. The server generates continuous natural language paragraphs, forming a complete prompt statement instructing the generative AI model to perform the network topology generation task. Through string concatenation and formatting operations, the server obtains the final text. The output is a single prompt statement text, for example: "You are a senior network architect."

[0116] Based on the following network import status, please generate the optimal network topology scheme for a new office.

[0117] - Scene: New office; - Planned equipment to be introduced: 1 enterprise-grade router, 2 Layer 2 switches, and 1 firewall; - Number of employees: Approximately 50; - Requirement: Support VPN remote access and provide basic secure access control.

[0118] please: 1) Describe the recommended network layer structure (core layer / aggregation layer / access layer); 2) Explain the role of each piece of equipment; 3) Describe in detail the connection relationships between the devices; 4) If possible, please output a topology description including nodes and connections for subsequent visualization. Step 9: The server will input the prompt statement into the generative artificial intelligence model and obtain the model's output.

[0119] The server takes the prompt text from step 8 as input, sets parameters such as model identifier, maximum generation length, and temperature coefficient, and constructs a model invocation request. The server sends the prompt text to the generative AI model through the model interface. The model internally performs operations such as vector embedding, attention weight calculation, and sequence generation to produce output text and possible structured fragments. The server receives the model's return results and loads the complete output text into memory. The output is the model output text containing a network topology description and / or structured topology fragments.

[0120] Step 10: The server extracts structural information related to the network composition from the model output.

[0121] The server takes the model's output text as input and uses pattern matching and parsing rules to find parts representing communication devices and their connections, such as phrases containing "node," "device," or "from...to...". The server performs word segmentation and syntactic analysis on these phrases, extracting elements such as device name, device type, and connection endpoints. The server then constructs an intermediate data structure containing a list of devices and a list of connections. The output is a set of candidate nodes and candidate connections before classification.

[0122] Step 11: The server organizes the candidate set into standardized node and connection information.

[0123] Taking the candidate set from the previous step as input, the server assigns a unique identifier to each candidate device, determines its type label and display name, and verifies whether the start and end points of each candidate connection exist in the node set. The server eliminates incomplete or conflicting candidates and completes missing information according to preset rules (e.g., generating default identifiers for unnamed devices). The server organizes the results into two lists: a list of node information and a list of connection information. The output is a structured topology data object, in the form of a unified data structure containing node and connection information.

[0124] Step 12: The server converts topology data objects into structured data for visualization.

[0125] The server takes a list of node information and a list of connection information as input, assigns hierarchical attributes or initial layout coordinates to each node, and divides the nodes into layers according to their roles (such as edge devices, security devices, and switching devices). The server calculates the approximate inter-layer relationships based on the connection directions and adds layout auxiliary attributes to the connection records. The server then serializes the above data into a format compatible with the front-end rendering logic. The output is structured topology data that can be directly used by the terminal for drawing.

[0126] Step 13: The server constructs a response message and sends it to the terminal.

[0127] The server takes the text description from step 9 and the structured topology data from step 12 as input, and constructs a response object containing the text description and topology data. The server packages this object into a response message using application layer protocols and sends it to the terminal with the support of the transport and network layers. The output is a response message containing a description of the network structure and visual data.

[0128] Step 14: The terminal receives the server's response and parses and displays the data.

[0129] The terminal takes the server's response message as input, separating the text description and the topology structured data. The terminal uses local data parsing functions to restore the structured data to a set of nodes and connections in memory, and stores the text description in the display buffer. The terminal then prepares to call the graphics rendering module to generate a graphical object based on the node and connection information. The output consists of node data, connection data, and the descriptive text to be displayed, all suitable for rendering.

[0130] Step 15: The terminal draws a network diagram on the display unit and presents a text description.

[0131] The terminal takes node and connection data as input, calls the graphics rendering engine, and draws icons and connections for each communication device on the display canvas according to hierarchy and relationship. The terminal uses different graphic markers and colors for different types of devices and displays device names or role labels in appropriate locations. The terminal also displays text descriptions provided by the server in the interface area, enabling users to understand the network topology by combining graphics and text. The output is a network diagram and accompanying text descriptions presented on the terminal display unit.

[0132] Step 16: Users can input additional requests or corrections based on the displayed results.

[0133] The user uses the network diagram and description seen in step 15 as a reference to consider any shortcomings or new requirements. The user enters natural language text such as "Add an internet backup line and add a redundant firewall" into the terminal interface, or selects the "Add Redundancy" option through interface controls. The terminal collects this additional or corrected input as new request data. The output is the user input data containing the additional requirements or corrections.

[0134] Step 17: The terminal sends appended or corrected input to the server to initiate iterative processing.

[0135] The terminal uses the additional input data generated in the previous step and the existing network import status as input to construct an update request message, including the information "This request is for a modification of the existing scheme" in the message. The terminal sends this message to the server via the communication protocol. The output is an update request message, which, upon reaching the server, triggers the update processing of existing records and prompt statements.

[0136] Step 18: The server updates the network import status records and regenerates the prompt statements.

[0137] The server takes the update request from step 17 as input, reads the original network import status record from the information storage device, and merges the appended or corrected inputs into the corresponding fields, such as setting the redundancy requirement flag to true, increasing the number of security devices, or modifying capacity parameters. The server updates the database record and generates a new structured network import status. Subsequently, the server repeats the logic of steps 7 and 8, but adds the explanation "incremental optimization based on the existing solution" to the prompt statement. The output is the updated network import status record and the new prompt statement text, for example: "Based on the network topology previously generated for the new office (including 1 router, 2 switches, and 1 firewall), please perform the following incremental optimizations:" - Add an internet backup line; - Add a redundant firewall to form a high-availability cluster with the existing firewalls; - Explain the failover strategy for primary / backup lines and dual firewalls.

[0138] Please output the updated network topology description and provide an explanation of the new topology structure, making the newly added devices and links clearly visible. Step 19: The server invokes the generative AI model again based on the new prompt and updates the visualization data.

[0139] The server takes the new prompt from step 18 as input, repeats the processing flow from steps 9 to 13, obtains the updated model output, re-extracts and reorganizes node and connection information, generates new visualized structured data, and sends it to the terminal. The output is a new network structure graph reflecting the additional requirements.

[0140] Step 20: The terminal refreshes its display to show the updated network structure diagram.

[0141] The terminal takes the latest structured topology data and text description returned in step 19 as input, replaces the previous node and connection data, and re-executes the graphics rendering process. The terminal displays a new network configuration diagram including backup lines and redundant firewalls on the display unit and updates the text description. The output is a latest network configuration diagram reflecting the results of iterative optimization, which the user can use to continue with subsequent operations or end the design process.

[0142] Application Example 1 The process flow corresponding to the specific processing in Use Case 1 will be described below. The various parts of the system described below are implemented by the data processing device 12 and the intelligent device 14. Furthermore, the data processing device 12 is referred to as the "server" and the intelligent device 14 is referred to as the "terminal".

[0143] The specific processing performed by the specific processing unit 290 of the data processing apparatus 12 in Application Example 1 is achieved by the following means.

[0144] In this invention, the server includes: means for storing information including industrial equipment status information and communication link operation status as communication environment information in an information storage device; means for generating prompt statements as input to a generative artificial intelligence model based on the stored communication environment information and device identification information and connection request information obtained from the terminal; means for inputting the generated prompt statements into the generative artificial intelligence model to calculate communication configuration information including the connection relationship between communication devices and information processing devices and candidate communication paths; means for generating configuration graph data with communication devices and industrial equipment as nodes and physical or logical connection relationships as edges based on the calculated communication configuration information, and determining the optimal communication path among multiple communication paths included in the configuration graph data according to latency and load-related evaluation indicators; means for sending the determined optimal communication path and corresponding connection interface setting information to the terminal, and presenting it on a display device in the form of a network configuration graph through the terminal; and means for generating communication device setting data corresponding to the determined optimal communication path based on confirmation information from the terminal, and sending the setting data to the communication device. This enables the formation of an end-to-end processing flow within the server, encompassing communication environment information collection and modeling, automatic generation of prompt statements, generative artificial intelligence model reasoning, graph structure optimization calculation, and automatic issuance of configuration instructions. This achieves intelligent and automated generation of industrial network structure graphs and selection of communication paths, reducing reliance on manual network design experience while improving the accuracy, scalability, and adaptability to changes in production status. Ultimately, this improves the overall performance and resource utilization efficiency of industrial communication systems at the computer technology level.

[0145] A "system" refers to a collection of devices consisting of multiple functional components used to perform industrial communication network design, optimization, and configuration processing. It may include servers, terminals, communication equipment, industrial equipment, and information storage devices.

[0146] "Processing device" refers to a hardware or logic unit that executes program instructions in a system, processes and operates on input data to achieve a predetermined function, including but not limited to a central processing unit, a graphics processing unit, a programmable logic device, or a combination thereof.

[0147] "Information storage device" refers to a storage component used to store communication environment information, configuration data and other related data in a read-write manner, including but not limited to semiconductor memory, magnetic storage medium, optical storage medium or a combination thereof.

[0148] "Industrial equipment" refers to production-related devices used in industrial settings that exchange data with other devices through communication networks, including but not limited to industrial robots, sensors, actuators, production line equipment, and their control units.

[0149] "Communication equipment" refers to network devices used in communication networks for forwarding, exchanging or routing data, including but not limited to switching equipment, routing equipment, gateway equipment, access equipment and wireless access units.

[0150] "Information processing device" refers to a computing device that is connected to communication equipment through a communication network and performs data processing tasks, including but not limited to server devices, control devices, monitoring devices and user terminal devices.

[0151] "Terminal" refers to a human-computer interaction device operated by a user for inputting device information, receiving and displaying network configuration diagrams and related settings information, including but not limited to portable terminals, desktop terminals or industrial operation terminals.

[0152] "Communication environment information" refers to a comprehensive set of information representing the operating status and structural characteristics of a communication network, including but not limited to industrial equipment status information, communication link operating status, network topology, configuration parameters, and historical monitoring data.

[0153] "Industrial equipment status information" refers to data related to the current or historical operating status of industrial equipment, including but not limited to operating mode, start / stop status, data transmission frequency, real-time requirements, and fault status.

[0154] "Communication link operating status" refers to data related to the current or historical performance of the communication link, including but not limited to bandwidth utilization, transmission latency, packet loss rate, error rate, and link availability.

[0155] "Equipment identification information" refers to the marking information used to uniquely identify industrial or communication equipment, including but not limited to equipment number, name, address, location identifier, and type identifier.

[0156] "Connection request information" refers to information input by the terminal or user to indicate the access or change requirements of a new or existing device to the communication network, including but not limited to the target access node, expected bandwidth, latency requirements, and communication protocol type.

[0157] "Generative artificial intelligence models" refer to artificial intelligence models that are trained to automatically generate text, structured data, or other forms of output based on input prompts, including but not limited to deep learning-based language models, graph generation models, or multimodal models.

[0158] "Prompt statements" refer to text or equivalent representations that are constructed by the processing device and input into the generative artificial intelligence model to describe the current communication environment information, equipment information, and design requirements, in order to guide the generative artificial intelligence model to output results that meet expectations.

[0159] "Communication constituent information" refers to a set of data obtained by generative artificial intelligence models or subsequent calculations, representing the connection relationships and configuration schemes between communication devices, information processing devices and industrial equipment, including but not limited to the connection relationships between nodes, interface allocation information and link attributes.

[0160] "Communication path candidates" refers to a set of one or more possible transmission paths from a source node to a destination node under a given network topology. Each path consists of several communication devices and their interfaces in sequence.

[0161] “Graph data” refers to graphical information that represents the structure of a communication network or system in the form of a data structure, including graph structures with communication equipment and industrial equipment as nodes and physical or logical connections as edges, as well as their additional attribute information.

[0162] "Physical connection relationship" refers to the connection relationship directly realized by the transmission medium in a communication network, including but not limited to the device connection formed through wired links, wireless links or fieldbus.

[0163] "Logical connection relationship" refers to the virtual connection relationship established at the protocol layer or configuration layer, including but not limited to virtual local area networks, tunnels, virtual private networks and logical routing relationships.

[0164] "Communication path" refers to the ordered sequence of nodes and links that data passes through in a communication network from the source node to the target node, including intermediate communication devices and their interfaces.

[0165] "Optimal communication path" refers to the communication path with the best performance selected from multiple communication paths after comprehensive evaluation based on predetermined evaluation indicators (including but not limited to latency, bandwidth utilization, load balancing and reliability).

[0166] "Latency-related evaluation metrics" refer to quantitative indicators used to evaluate the latency performance of communication paths, including but not limited to end-to-end latency, single-hop latency, jitter and their statistics.

[0167] "Load-related evaluation metrics" refer to quantitative indicators used to assess the load of communication paths or links, including but not limited to bandwidth utilization, traffic occupancy, queue length, and congestion level.

[0168] "Connection interface settings information" refers to the parameter information used to configure ports or interfaces on communication devices, including but not limited to port enable status, speed settings, virtual LAN identifier, address information, and access control rules.

[0169] "Communication equipment configuration data" refers to the set of configuration data generated by the processing device and used to send to the communication equipment so that the communication equipment can operate according to the target network configuration, including but not limited to command sequences, configuration files, parameter tables and scripts.

[0170] The "communication address system" refers to the address structure and allocation rules used in a communication network to uniquely identify each node and interface and to perform data addressing, including but not limited to network addresses, subnetting, and address allocation strategies.

[0171] "Path control methods" refer to the control mechanisms and strategies used to select and maintain communication paths in a network, including but not limited to routing protocols, forwarding rules, load balancing strategies, and failover mechanisms.

[0172] "Design guidance information" refers to advisory information generated based on communication environment information and the output results of generative artificial intelligence models, used to assist in network planning and configuration, including but not limited to address planning suggestions, path control strategy suggestions, and device access schemes.

[0173] "Monitoring data" refers to data generated and collected by communication or industrial equipment during operation to reflect the operating status of the equipment and network, including but not limited to performance indicators, alarm information, log information, and statistical data.

[0174] "Generation parameters" refer to the parameters used to control the output content and form when calling a generative artificial intelligence model, including but not limited to temperature coefficient, generation length, sampling strategy, and filtering conditions.

[0175] In one embodiment of the present invention, the server serves as the core computing node, the terminal serves as the human-computer interaction node, and the user serves as the main operator in the industrial field. The three work together to achieve the automatic design, optimization, and configuration of the industrial communication network.

[0176] The server can utilize general-purpose computer equipment in terms of hardware, including multi-core CPUs, graphics processing units, main memory, and solid-state storage. In terms of software, the server can run a Unix-like operating system and deploy a web service framework, a database management system, and a program environment for executing generative artificial intelligence models and graph algorithms. The server can use Python as its primary programming language and leverage deep learning frameworks (such as frameworks based on tensor operations libraries), graph computing libraries (such as algorithm libraries supporting graph structure operations), and HTTP service frameworks to build the entire system.

[0177] During system initialization, the server creates a database schema in the storage device to store communication environment information. The server stores industrial equipment status information, communication link operating status information, network topology information, device identification information, and connection request information as relational data tables and document data. Specifically, the server uniformly encodes industrial equipment, communication equipment, and information processing devices as node identifiers, and encodes information such as port, bandwidth, latency, and error rate as edge attributes. This structured storage allows the data to be directly loaded into a graph data structure by subsequent algorithm modules, thereby reducing the overhead of repeated parsing and transformation.

[0178] In collecting communication environment information, the server interacts with communication devices and industrial equipment in the industrial field through communication interfaces. The server can use standardized management protocols to collect operational status data such as port traffic, packet loss rate, and real-time latency from switching and routing devices. The server can also collect status parameters such as operating mode, data transmission frequency, and real-time requirements from industrial equipment via fieldbus or industrial Ethernet gateways. The server performs timestamp alignment, unit conversion, and outlier filtering on the collected data, persisting the processed data as part of the communication environment information.

[0179] When constructing generative AI models, the server can employ a transformer-based generative AI model. During the training phase, the server uses historical network topology examples, real-world monitoring data, and corresponding network configuration schemes designed by human experts as training datasets. The server encodes the input as a sequence containing a list of devices, link attributes, and demand constraints, and the output as a network configuration description and communication path description. The server uses a cross-entropy loss function during model training, iteratively updates model parameters using backpropagation, employs an adaptive learning rate optimization algorithm to improve convergence speed, and reduces overfitting through gradient pruning and regularization. During training, the server can perform data augmentation on the input data, such as equivalent renaming of topology structures and noise perturbation of link loads, to improve the model's generalization ability across different scenarios.

[0180] The server uses a pre-trained generative artificial intelligence model during the inference phase. The server converts communication environment information and connection request information input by the user through the terminal into prompt statements. When constructing prompt statements, the server encodes graph structure information into natural language descriptions, such as enumerating existing switching devices, their port utilization, and new device requirements. The server can generate prompt statements like the following: "You are an industrial network architecture expert. The current factory network topology is as follows:" 1. Switching device A: Port 1 is connected to the existing robot, with a traffic utilization rate of 70%; Port 2 is idle.

[0181] 2. Switching device B: Port 5 is connected to the visual inspection device, with a traffic utilization rate of 30%; Port 10 is idle.

[0182] New device information: Device ID: Robot-W01 Location: Production Line A-3 - Protocol: EtherCAT - Expected bandwidth: 100 Mbps - Latency requirement: less than 10 ms Based on the information above, please generate a network configuration diagram design, select appropriate switching devices and ports to connect to the device, and output a recommended network configuration diagram description, as well as the optimal communication path from Robot-W01 to the control device Ctrl-01 (indicated by the order of the devices and ports traversed). The server inputs the prompt statement into the generative artificial intelligence model. Internally, the server segments and encodes the prompt statement to obtain a context vector representation, and then progressively generates the output sequence through a multi-layered self-attention network and a feedforward network. During the generation process, the server employs a beam search or temperature sampling strategy to balance diversity and determinism. The server's model output includes device access location suggestions, VLAN IDs, address planning suggestions, and a list of candidate communication paths.

[0183] After receiving the model output, the server parses the output text, transforming the access suggestions and path descriptions into structured data. Using a graph algorithm library, the server constructs a graph object from the current network topology and suggested access nodes. The server maps the bandwidth, current utilization, and monitored latency of each link to edge weights. It then forms a comprehensive cost function by performing a non-linear combination of these edge weights (e.g., weighted summation of latency and load according to preset weight coefficients). The server executes a shortest path algorithm or a multi-constraint path search algorithm on this graph object to further refine the communication path candidates given by the generative AI model, thereby obtaining the optimal communication path under the comprehensive cost function.

[0184] Instead of simply repeating manual calculations performed by human engineers, the server evaluates candidate paths based on high-dimensional feature vectors and global topological information. By utilizing the candidate space provided by a generative AI model, the server avoids blindly searching the entire path space, improving the computational efficiency of path optimization. Simultaneously, the server further optimizes the model output using graph algorithms, effectively reducing the impact of model generation errors on the final result and enhancing its verifiability and stability.

[0185] The terminal serves as the user interface in this system. Hardware-wise, the terminal can be a tablet, industrial operating terminal, or general-purpose computer; software-wise, it runs an operating system and a browser or dedicated application. The terminal encapsulates user-inputted device identification information, location information, bandwidth requirements, and latency requirements into a request and sends it to the server. Upon receiving the structural graph data and optimal communication path information returned by the server, the terminal converts it into a visual graphic and draws nodes and edges on the display device using a graphics library. The terminal highlights the links of the optimal communication path in the interface and displays detailed port configuration and address planning suggestions in the side area.

[0186] Users can view network diagrams and path suggestions via a terminal, and then perform physical wiring and preliminary configuration of industrial and communication equipment on-site based on these suggestions. After completing the configuration, users send confirmation information to the server via the terminal. Upon receiving the confirmation information, the server automatically generates communication equipment configuration data based on the final network scheme. The server can encode this configuration data into a sequence of configuration commands or a configuration file structure and distribute it to the communication equipment through the management interface. Before distributing the configuration data, the server can simulate the effective routing table and forwarding table to check for address conflicts or routing loops, thereby identifying potential errors before actual equipment reconfiguration and reducing the risk of network outages.

[0187] After the configuration is distributed, the server continues to collect monitoring data from communication and industrial equipment. The server compares this monitoring data with the initial design goals, such as comparing actual end-to-end latency with estimated latency, and comparing link utilization with expected load range. When the server detects performance deviations from thresholds, it updates the communication environment information and reconstructs the prompts, triggering joint processing by generative artificial intelligence models and graph algorithms to propose new optimization solutions. Through this closed-loop approach, the server continuously corrects and improves the network configuration, enabling the network to maintain optimal performance even when faced with additions or removals of devices, changes in service models, and fluctuations in link quality.

[0188] Internally, the server can be divided into modules such as a communication environment information management module, a prompt statement generation module, a generative artificial intelligence reasoning module, a graph algorithm optimization module, a visualization data generation module, and a configuration distribution module. The server exchanges information between modules using predefined data structures. For example, it uses a unified list of nodes, edges, and attributes to represent the topology, and a unified path object structure to represent communication paths. This unified data structure design reduces redundant calculations during data conversion and improves overall processing speed.

[0189] The generative AI model employed by the server differs from traditional rule-based design approaches. Instead of assigning a unique solution to each device type and load condition using fixed IF-THEN rules, the server learns implicit topology patterns and path selection strategies from large-scale historical samples. This allows the model to automatically generate reasonable configuration solutions when faced with unseen device combinations and load scenarios. By explicitly embedding current monitoring data (e.g., "Link X current utilization 85%, average latency 12ms") in prompts, the server enables the model to consider real-time conditions when generating solutions, thus achieving adaptation to dynamic environments.

[0190] After adopting the aforementioned structure and algorithms, the server can bring about multiple technical benefits. By centrally modeling and rapidly querying communication environment information, the server improves the efficiency of network design-related data management; through the synergy of generative artificial intelligence models and graph algorithms, the average computation time for network graph generation and path optimization is significantly shortened; by encoding detailed constraints and real-time monitoring indicators in prompt statements, the server improves the accuracy of generated results in meeting bandwidth, latency, and reliability requirements; and by automatically generating and distributing configuration data, the server reduces errors caused by manual configuration and lowers the probability of network outages. These technical benefits are not simply about automating manual operations, but rather about enhancing the server's efficient computing capabilities for complex industrial networks through improved data structure design, model structure design, and algorithm flow design.

[0191] In another implementation, the server can replace the internal structure of the generative AI model. The server can replace the text generation model based on a transformer structure with a multimodal model that simultaneously processes graph structures and text. In this structure, the server encodes the communication network topology as graph structure vectors, encodes the user request description as text vectors, jointly models the network through a graph-text interaction attention mechanism, and then generates a network configuration scheme. This approach can further improve the ability to model complex topological relationships. In different implementations, the server can adjust the number of model layers, the dimension of hidden vectors, and the number of attention heads to balance different hardware resource constraints and real-time requirements.

[0192] In another implementation, the server may not directly send configurations to the communication device. Instead, it may provide the terminal with detailed configuration scripts and operation steps, which the user can choose to execute as needed. The server still internally performs the same prompt generation, model inference, and path optimization processes, but in the final stage, it does not automatically write the configuration to the device; instead, it exports the configuration results as a file for user reference. This implementation is suitable for scenarios with high security and manual review requirements.

[0193] In summary, the server performs structured modeling of the communication environment information, utilizes generative artificial intelligence models and graph algorithms to collaboratively optimize the network structure graph and communication paths, and automatically maps the results to executable configurations for communication devices. This achieves a complete technical solution from data acquisition, intelligent generation, algorithm optimization to device control. The terminal provides users with a visual and operational interface, allowing users to confirm and monitor system outputs when necessary. Through the above configuration, the system of this invention achieves overall improvement in processing speed, solution accuracy, and resource utilization in the field of industrial communication network design and operation by coordinating the internal storage structure, processing flow, and algorithm combination of the computer.

[0194] use Figure 12 The processing procedure is explained.

[0195] Step 1: Users input device information and connection requirements using a terminal.

[0196] Users fill in fields such as device identifier, installation location, expected bandwidth, latency requirements, protocol type, and planned communication equipment in the terminal interface, and then click the submit button after confirmation.

[0197] The terminal's input consists of structured form data filled in by the user on the interface. The terminal performs basic format checks on the input data (e.g., whether it is empty or the values ​​are valid), then encapsulates the form data into a request message and sends it to the server over the network. The terminal's output is a request message containing device identification information and connection request information.

[0198] Step 2: The terminal sends device information and connection requests to the server.

[0199] After receiving form data from the user, the terminal converts each field into standard key-value pairs and serializes them into request data. The terminal appends authentication information and a timestamp to the request data for the server to verify the request's origin and timing.

[0200] The terminal's input consists of form data filled out by the user. The terminal performs data processing locally, packaging the multi-field form data into a single message with a unified structure. The terminal's output is request data encapsulated using network protocols, which is then sent as input to the server.

[0201] Step 3: The server receives requests from the terminal and performs data verification and standardization.

[0202] The server reads the request data sent by the terminal from the network interface and parses out the device identification information and connection request information. The server checks whether the required fields exist, whether the data type is correct, and whether the value is within the preset range. For numerical fields such as bandwidth and latency, the server converts different units (such as "100M" and "0.1G") into a unified numerical form and performs unified encoding on the location description.

[0203] The server's input is the request data sent by the terminal. The server generates a standardized internal representation (e.g., a unified field structure and data type) through data processing operations such as data parsing, type conversion, and range checking. The server's output is a validated and standardized device information record, which is written to the information storage device and used as input for subsequent processing modules.

[0204] Step 4: The server collects communication environment information from communication equipment and industrial equipment.

[0205] Based on a pre-configured list of communication devices and industrial equipment, the server sends status query requests or subscribes to monitoring data to each device. The server obtains bandwidth utilization, error counts, and round-trip latency for each port from switching and routing devices, and current operating status, data transmission frequency, and real-time requirements from industrial equipment. The server performs time alignment and anomaly filtering on the collected data, summarizing it into a snapshot representing the current state of the communication environment.

[0206] The server's input consists of raw monitoring data streams from multiple communication and industrial devices. The server performs data processing operations such as timestamp normalization, unit conversion, and outlier removal, mapping the multi-source data to a unified set of fields. The server's output is a set of structured communication environment information records, which are saved to an information storage device and used as input for subsequent prompt generation and path optimization.

[0207] Step 5: The server integrates device information and communication environment information to construct internal topology and demand representation.

[0208] The server reads stored communication environment information from the information storage device, including existing communication device nodes, link relationships, link attributes, and current monitoring metrics. The server inserts the device identification information and connection request information of the new device into the topology data structure, designating it as a node to be connected. The server constructs a graph data structure, using devices and communication equipment as nodes and links as edges with attributes.

[0209] The server takes standardized device information records and communication environment information records as input. It performs a graph construction operation, mapping multi-table data into a single graph structure, where each node and edge carries an attribute vector. The server outputs an internal topology graph object and a set of constraint parameters representing new device requirements; this data will serve as input for generating prompt statements.

[0210] Step 6: The server generates prompts for generative artificial intelligence models based on topology and requirements.

[0211] The server extracts the load information of key nodes (such as candidate communication devices) and their ports from the topology graph object, and reads the bandwidth and latency requirements of new devices from the constraint parameters. The server organizes this information into natural language text according to a predetermined template, which includes an overview of the current network structure, new device requirements, and design goals. The server uses a consistent description method in the text to ensure that the generative artificial intelligence model can accurately parse this information.

[0212] The server takes a topology diagram object and new equipment requirement parameters as input. It processes the structured diagram data into sequential text, or prompt statements, using methods such as string concatenation, formatting, and template filling. The server outputs a complete prompt statement, which serves as input to a generative artificial intelligence model.

[0213] Step 7: The server inputs the prompts into the generative artificial intelligence model and generates communication structure information and path candidates.

[0214] The server invokes a locally deployed generative AI model, using the generated prompts as input. Internally, the server segments and embeds the prompts, representing them as vector sequences, which are then processed through a multi-layered self-attention network for forward propagation. At each layer, the model weights and combines word vectors at different positions to extract features reflecting network structure and demand constraints. The server progressively generates text sequences from the output layer, containing recommended access communication devices, access ports, virtual network identifiers, and a set of candidate communication paths.

[0215] The server's input is a prompt text. The server performs data computation operations such as matrix multiplication, attention weighting, and nonlinear transformations within the generative AI model, obtaining the text-based output of the model. The server's output is a text containing communication structure information and candidate communication path descriptions, which is then parsed into structured data.

[0216] Step 8: The server parses the output of the generative artificial intelligence model and extracts the structured communication components.

[0217] The server performs sentence segmentation and key phrase matching on the text output by the model, identifying recommended access communication device identifiers, port numbers, virtual network numbers, address schemes, and the devices and interfaces traversed in the path. The server then populates this information into data structures, such as device access records and path records, according to predefined formats.

[0218] The server takes as input the text description output by the generative artificial intelligence model. It performs data processing operations such as text parsing, keyword matching, and field extraction, transforming the natural language results into structured communication components and a set of candidate communication paths with field names and values. The server outputs a set of structured access suggestion records and path candidate records, which are provided to the subsequent graph algorithm optimization module.

[0219] Step 9: The server uses a graph algorithm to finely optimize candidate communication paths and determine the optimal communication path.

[0220] The server maps candidate paths to an internal topology graph, marking the nodes and edges on each candidate path. The server calculates the overall cost for each edge based on communication environment information, such as summing the base latency and load factor by weights to form the cost value. The server accumulates the cost value for each candidate path and compares it with other reachable paths in the graph, selecting the path with the lowest overall cost as the optimal communication path. The server can also perform constrained shortest path search on the graph, directly filtering paths based on latency upper limits and bandwidth lower limits.

[0221] The server takes as input a topology graph object, path candidate records, and real-time link performance parameters. It performs data computation operations such as edge weight calculation, path cost accumulation, and optimal path selection, choosing the path from the candidate set that satisfies the constraints and has the lowest total cost. The server outputs one or more optimal communication path records, including the specific communication devices traversed, interface identifiers, and estimated latency.

[0222] Step 10: The server generates the structural diagram data and connection interface settings for display on the terminal.

[0223] Based on the structured communication composition information and optimal communication path records, the server adds access edges to the topology graph for new devices and sets highlight markers and additional attributes (such as estimated latency and bandwidth) for edges on the optimal path. The server converts the graph structure into a list of nodes and edges for visualization. The node list contains display parameters such as coordinates and icon type, while the edge list contains connection endpoints and display styles. The server also generates port configuration suggestions and address configuration suggestions.

[0224] The server takes topology graph objects, access suggestion records, and optimal path records as input. Through data processing operations such as graph transformation, attribute annotation, and visualization parameter appending, the server converts the internal graph objects into visual structural graph data and corresponding connection interface settings. The server outputs visual structural graph data and configuration information, which are encapsulated into response messages and sent to the terminal.

[0225] Step 11: The terminal receives network structure data and settings information, and presents the network structure diagram to the user.

[0226] The terminal receives response data from the server, containing a list of nodes, a list of edges, and configuration information. The terminal parses the attributes of the nodes and edges locally and uses a graphics drawing component to draw icons of industrial and communication equipment on the display screen, representing connections with lines. The terminal highlights or bolds the lines corresponding to the optimal communication path and displays the recommended access port, virtual network identifier, and address configuration parameters in text format in an adjacent panel.

[0227] The terminal's input consists of the network structure data and configuration information returned by the server. Through visualization rendering, the terminal transforms the abstract data structure into a graphical and textual interface, outputting a network structure diagram and detailed configuration suggestions displayed on the screen for user viewing.

[0228] Step 12: Users perform on-site wiring and equipment configuration according to the suggestions displayed on the terminal, and send confirmation information back to the server.

[0229] Users observe the network diagram and optimal communication path displayed on the terminal and, following the recommendations, connect the physical interface of the new industrial equipment to the designated communication device port on-site. Users then enter the suggested virtual network identifier and address parameters in the communication device's management interface and configure the network address and communication parameters in the industrial equipment's human-machine interface. Upon completion, users select "Configuration Complete" or a similar option on the terminal interface to trigger a confirmation message to the server.

[0230] The user's input consists of the actual operational results of physical wiring and equipment parameter settings. The user then abstracts this result into an acknowledgment signal via the terminal. The terminal's output is a message containing the acknowledgment flag, which the server receives as a trigger condition for subsequent verification and configuration issuance.

[0231] Step 13: The server generates and sends communication device configuration data based on the confirmation information.

[0232] After receiving confirmation from the terminal, the server generates configuration data for the communication device based on the final determined access device and optimal communication path. This data includes port enabling configuration, virtual network partitioning, access control rules, and static routing. The server performs consistency checks on this configuration data to avoid address conflicts or routing loops. The server then uses its management interface to call the communication device's configuration interface, distributing the configuration data item by item to the corresponding communication devices.

[0233] The server's inputs include access suggestion records, optimal path records, and user confirmation information. The server performs data calculations and processing operations such as configuration template filling, parameter checking, and device mapping to obtain configuration data specific to each device. The server's outputs are the sequence of configuration commands issued to the communication devices and the configuration status records registered in the information storage device.

[0234] Step 14: The server monitors the communication status after the configuration takes effect and updates the communication environment information based on the monitoring data.

[0235] After the server distributes the configuration, it periodically collects new monitoring data from communication and industrial equipment to check whether the latency, bandwidth utilization, and error rate on the optimal communication path are within the predetermined range. The server compares this monitoring data with the original estimates, and if the deviation is large, it marks the object that needs optimization. The server writes the new monitoring data to the information storage device to update the communication environment information.

[0236] The server's input is the newly collected monitoring data. Through statistical analysis and threshold comparison, the server evaluates the communication quality after the configuration takes effect and updates the communication environment information record with the analysis results. The server's output is the latest communication environment information status, which serves as the basis for subsequent construction of prompts and path optimization, enabling the system to continuously adjust and improve communication paths as the environment changes.

[0237] Alternatively, an emotion engine for inferring user emotions can be combined. That is, the specific processing unit 290 can also use the emotion-specific model 59 to infer user emotions and perform specific processing using user emotions.

[0238] Example 2 The flow of a specific process in Example 2 will be described. Each part of the system described below is implemented by the data processing device 12 and the smart device 14. The data processing device 12 will be referred to as the "server," and the smart device 14 as the "terminal."

[0239] In existing communication network design technologies, computing devices typically generate network topology schemes by simply matching and combining network device lists based on static parameter configuration templates or fixed rules. This approach has significant technical shortcomings in the following aspects: (1) On the input side, the computing device has difficulty in automatically extracting the truly critical elements for generating the network composition graph from unstructured or semi-structured imported status data. This often requires professional engineers to manually interpret and re-input the data, resulting in information loss and redundant processing between the human-computer interaction interface and the back-end computing process, which reduces the overall data processing efficiency.

[0240] (2) On the model calling side, traditional systems only input a small number of structural parameters such as the number and model of devices directly into the algorithm or rule engine. They lack the ability to uniformly model and encode multi-source data such as network scale, business requirements, and previous design history. They cannot dynamically adjust the instructions input to the intelligent model or algorithm for different import conditions, resulting in insufficient diversity, adaptability and optimization of the network composition graph generation results.

[0241] (3) On the data management side, existing systems usually only save the final topology results in the form of files or images, without storing the prompts, inference context, historical design process, etc. used to generate the results in a structured way. These results cannot be effectively reused by the computing device or used to continuously optimize the model input strategy, resulting in a waste of computing and data resources.

[0242] (4) In terms of the overall system architecture, the server side often separates the processing of network design tasks into multiple links such as database management, natural language processing, model calling, and visualization output. It lacks a technical solution for integrated optimization of the pipeline of "prompt statement - generative artificial intelligence model - topology data", resulting in long end-to-end processing path, large processing latency, poor scalability, and is not conducive to efficient operation on general computing platforms.

[0243] Therefore, a new computer implementation is needed to enable the server to: (a) The communication network import status is efficiently stored as structured constituent information in the relational data management device, and prompt statements for driving generative artificial intelligence models are automatically constructed on this basis. (b) By jointly managing and analyzing historical composition information and model output results, the content and structure of prompt statements are dynamically optimized, thereby improving the output quality and computational efficiency of generative artificial intelligence models in network composition graph generation tasks. (c) Convert the generated topology data into a standardized data format for front-end display control, reduce the rendering burden on the terminal side, and realize an integrated computing process from importing status data, generating prompt statements, model reasoning to visual output of the composition graph, thereby improving the overall technical performance of computer systems in the field of communication network design.

[0244] The specific processing performed by the specific processing unit 290 of the data processing apparatus 12 in Embodiment 2 is achieved by the following means.

[0245] In this invention, the server includes an information storage unit for storing the import status of the communication network as constituent information in a relational data management device; a prompt generation unit for obtaining the constituent information from the relational data management device and automatically generating prompt statements based on the constituent information to indicate the components required for generating a communication network constituent diagram; a model calling unit for inputting the prompt statements into a generative artificial intelligence model and obtaining constituent data representing the communication network constituent diagram based on natural language processing and data generation processing performed by the generative artificial intelligence model; and a display data generation unit for converting the constituent data into display control data to generate output information of the communication network constituent diagram that is visually displayed on a terminal device. This allows for the formation of an end-to-end computing pipeline within the server, encompassing structured storage of imported status data, automatic construction of prompt statements, generative artificial intelligence model inference, and standardized output of topology composition data. This not only reduces reliance on manual intervention and fixed rule configurations but also, by combining historical composition information stored in the relational data management device with model output results, dynamically optimizes the content and structure of prompt statements. Consequently, it improves the inference quality and processing efficiency of the generative artificial intelligence model in network composition graph generation tasks, thereby enhancing the overall performance and scalability of the computer system in communication network design applications.

[0246] A "system" refers to an integrated information processing unit composed of multiple functional units working together through hardware and software to perform data processing, model calling, and result output related to communication network design.

[0247] A "communication network" refers to a data transmission and exchange structure formed by the interconnection of multiple information processing devices, transmission devices, and connection media, used for sending and receiving data between different terminals.

[0248] "Import Status" refers to comprehensive status information related to the construction or expansion of communication networks, including existing or planned equipment configuration, deployment location, quantity, and service requirements.

[0249] "Construction information" refers to the set of network design-related data extracted from the imported status and represented in a structured manner, including equipment types, parameters, quantities, deployment areas, hierarchical relationships, and business constraints, which are used to support the subsequent network structure diagram generation process.

[0250] "Relational data management device" refers to an information processing device or software system that stores, queries, updates and manages data based on a relational data model. It is used to manage data such as constituent information, prompt statements and historical records in the form of tables.

[0251] "Information processing device" refers to a computing device that performs operations such as data reception, storage, calculation, conversion and output through a processor, memory and program instructions, including physical servers, virtual machines or cloud computing nodes.

[0252] "Prompt statements" refer to instructional information composed of natural language or structured text, used to explicitly indicate the target task, input conditions, and output requirements to generative artificial intelligence models, in order to guide the models to generate data related to the communication network structure diagram.

[0253] "Generative artificial intelligence models" refer to artificial intelligence models trained through machine learning that can automatically generate text, structured data, or other content based on input prompts, and have the ability to understand natural language and generate data.

[0254] "Natural language data" refers to information expressed in human-readable language (such as written sentences, paragraphs, etc.) that can be directly used as input or output content for generative artificial intelligence models.

[0255] Natural Language Processing (NLP) refers to a class of computational processing procedures that parse, understand, encode, and transform natural language data, including operations such as word segmentation, semantic analysis, intent recognition, and text generation.

[0256] "Data generation and processing" refers to the process by which generative artificial intelligence models automatically generate new text, structured data, or other information content based on input data and internal parameters through reasoning and computation.

[0257] "Construction data" refers to the structured data set output by generative artificial intelligence models, used to characterize the nodes, connections, and related attributes in the construction diagram of a communication network.

[0258] "Display control data" refers to the control data obtained by converting the constituent data, which is suitable for driving the terminal device to perform graphic rendering and interface display, including node coordinates, connection information, style parameters and layout instructions, etc.

[0259] "Output information" refers to data generated by the server and sent to the terminal device to present a diagram of the communication network structure and related explanatory content on the display interface, including display control data and associated text descriptions.

[0260] "Terminal device" refers to an information device operated by a user for interacting with a server via a network and displaying the results, including computing devices, mobile devices, or other devices with display and input functions.

[0261] "Design information" refers to the set of technical configuration suggestions or parameters generated during the communication network planning process, regarding address allocation, path control, security policies, resource allocation, and other aspects.

[0262] "Address allocation" refers to the process and result of assigning identification addresses to each node or subnet in a communication network, in order to achieve data forwarding, routing selection and node identification.

[0263] "Path control" refers to the control mechanisms and strategies for planning, selecting, and adjusting data packet transmission paths in communication networks, used to achieve traffic engineering, redundancy protection, and performance optimization.

[0264] "Additional information" refers to supplementary data generated by the information processing device based on the model results and import status, in addition to the constituent data. This includes design information, explanatory text, or annotation information for subsequent analysis.

[0265] "Historical information" refers to the data set recorded in the relational data management device, which includes compositional information, prompts, model output results, and time stamps related to previous network design requests, and is used for tracking and reuse.

[0266] "User attributes" refer to the classification or characteristic information related to the main user of the system, including industry type, organization size, previous design preferences, security level requirements, etc., which can be used to personalize the prompts and design schemes.

[0267] A “model invocation unit” refers to a software or hardware functional module within a server that is used to construct generative artificial intelligence model requests, send prompt statements, and receive model output results.

[0268] The "prompt generation unit" refers to a functional module within the server that extracts key elements from constituent and historical information and automatically generates prompt statements.

[0269] The “display data generation unit” refers to a functional module within the server that converts the constituent data output by the model into display control data suitable for visualization rendering on the terminal device and generates output information.

[0270] In one embodiment of the invention, the server operates as an information processing device, comprising a processor, main storage, non-volatile storage, and a network interface. Under the control of an operating system (e.g., a general-purpose server operating system), the server executes network design programs, database management programs, and generative artificial intelligence model invocation programs. The terminal is operated by a user and can be a computing device or a mobile device. The terminal communicates with the server via a network. The user inputs the communication network import status through the terminal's graphical user interface and views the communication network configuration diagram and related design information generated by the server.

[0271] The server stores relational data management software, such as a relational database management system, in non-volatile storage. Specifically, it can use general-purpose relational database software, such as a specific type of SQL database. The server reads and writes data to this relational data management system through a database driver. The server also stores client libraries for calling generative artificial intelligence models in the non-volatile storage, such as HTTP-based API client libraries, used to send prompts to external generative artificial intelligence services and receive model outputs.

[0272] The generative AI model invoked by the server in this embodiment can be implemented as a deep learning-based generative language model, internally employing a multi-layer Transformer neural network architecture. This model includes embedding layers, multi-head self-attention layers, feedforward network layers, and normalization layers. The server, through an interface provided by an external service, specifies the model name as a general large-scale language model and specifies inference parameters such as maximum output length, temperature coefficient, and sampling strategy. The server does not retrain the model locally; instead, it uses a pre-trained generative AI model that can be optionally fine-tuned by calling the interface.

[0273] The server defines various data structures within the relational data management device. For example, the server creates a "Design Master Table" in the database to store fields such as design identifier, user identifier, timestamp, number of users, number of floors, and business requirement descriptions for each network design request; the server creates an "Equipment Information Table" in the database to store fields such as type, model, manufacturing category, quantity, and installation location for each type of equipment; the server creates a "Prompt Statement Table" in the database to store the prompt statement text generated by the server and its associated design identifier, version number, and other fields; the server creates a "Model Output Table" in the database to store fields such as constituent data, text descriptions, address design information, and evaluation indicators returned by the generative artificial intelligence model; the server can also create a "Historical Feature Table" to store statistical features extracted from historical design records, such as common topology patterns, common equipment combinations, and typical user size ranges.

[0274] When the server executes the network design program, it first formats the import status data entered by the user through the terminal. The user enters text information in the terminal's browser, such as device type ("router," "switch"), device model, planned quantity, installation floor, and business requirements. The terminal uses a front-end script to package these input fields into structured data and sends it to the server over the network. Upon receiving this structured data, the server uses a data validation module to check if the device quantity is a positive integer, if any fields are missing, and if the text length exceeds a preset range. Then, the validated data is written to the design master table and device information table in the relational data management device.

[0275] During the generation of the prompt statement, the server reads the structural information corresponding to a specific design identifier from the relational data management device, including the number of users, the number of floors, the types and quantities of each piece of equipment, their installation locations, and the business requirement text. The server can also read historical statistical information related to similar scales or similar business types from the historical feature table, such as the core access layer structure pattern typically used for a certain range of user numbers. Through the prompt generation unit, the server combines these structured fields into a complete prompt statement according to a predefined natural language template. This natural language template contains fixed role setting descriptions and output format constraints. The server fills the variable positions with the current structural information, thereby obtaining a prompt statement specifically associated with the design request.

[0276] For example, the server can combine the constituent information into the following prompt statement: "You are a senior network engineer. Please design the topology of the office communication network based on the following conditions, and output a structural description of the network structure diagram:" - Number of office users: 50 - Number of office floors: 2 - Planned equipment to be used: - Router: AX-1000 (brand unknown), 1 unit, Location: 1st floor server room - Switches: 3 SW-24G units (brand unknown), located in the 2nd floor office area. - Business requirement: Provide high-speed wired and wireless networks for 50 employees. Please provide: 1. The role of each device in the network (core layer, aggregation layer, access layer, etc.); 2. The connections between devices (shown in a clear list or structured description); 3. Recommended VLAN segmentation and IP address planning suggestions. The server records the above prompt statement in the prompt statement table, and at the same time, through the model calling unit, it uses the prompt statement as the input of the "user role" and constructs a separate "system role" description, such as "You are an expert in enterprise network design and need to output structured results that can be parsed by the program according to the instructions", and sends it to the generative artificial intelligence model interface.

[0277] When the server invokes the generative AI model, it sends a request to the model server using an encrypted communication protocol via a network interface device. In the request, the server sets model parameters, such as a lower temperature value to reduce randomness and a maximum output length to avoid response truncation. The server can specify an output style of "first providing a brief text description, then a structured list," facilitating subsequent parsing. Internally, the generative AI model encodes the prompts using a multi-layer transformer structure, encoding each word or subword into a vector representation. It calculates the association weights between different words using a multi-head self-attention mechanism, mapping key phrases such as "number of users," "number of floors," "device type," and "location" in the prompts into high-dimensional features. The model then progressively generates output tags in the decoding phase, using autoregression to predict each subsequent word and mapping the prompt requirements to specific topological descriptions and configuration suggestions using parameters learned during pre-training. During training, the model uses cross-entropy loss as the error function, iteratively updating the weight matrix of each layer through backpropagation, and can be fine-tuned on domain-specific corpora to enhance its ability to recognize and generate terminology and topological patterns in the communication network domain.

[0278] After receiving the output from the generative AI model server, the server saves the output text to the model output table. The server then performs structured parsing on the output text using a text parsing module. This module uses regular expressions to match segment headings such as "Device Role List," "Connection Relationship List," "VLAN List," and "IP Planning Description," breaking each section down into structured records. The server generates fields such as node identifier, type, role attributes, and floor number for each device node; start and end identifier fields for each connection relationship; and subnet address, mask, and purpose description for each address planning entry. During parsing, the server uses a rule set, prioritizing the matching of standard phrases like "core layer" and "access layer." If non-standard descriptions appear, the server attempts to map them back to standard role categories using a thesaurus, thereby reducing parsing errors.

[0279] When generating display control data, the server maps structured component data into layout data suitable for front-end graphics rendering. The server can employ a hierarchical layout algorithm, placing core layer devices at the top of the layout space and access layer devices at the bottom, partitioning nodes vertically using the floor information of the devices. Based on the number of nodes and connections, the server calculates the relative coordinates between nodes and the paths of connecting lines, generating a node coordinate table and a connection table. The server can also specify icon types and color codes for each device type, for example, marking routers as circular nodes and switches as square nodes. The server combines these node coordinates, connection information, and display style parameters into display control data and encapsulates it along with text descriptions into output information.

[0280] After the terminal obtains the output information from the server, it renders the display control data in the browser using a drawing script library. The terminal displays a diagram of the communication network structure on the screen, showing the location and connection relationships of each node. Users can zoom, drag, and perform other operations on the terminal to intuitively view the topology. If the user finds that the number or structure of devices needs to be adjusted, they can modify the import status in the terminal's form. The terminal then sends the modified data back to the server, and the server repeats the above process of data storage, prompt statement generation, model invocation, and display control data generation, forming a closed-loop iterative design process.

[0281] In terms of data management, the server associates and stores prompt statements and model output results with design identifiers, enabling statistical analysis of historical data. The server can periodically extract features from the prompt statement table and model output table, such as statistically analyzing the most frequent topology patterns for a certain user scale or the common VLAN partitioning strategies for a certain type of device combination. The server can use simple clustering algorithms or frequency statistics methods to build a "common pattern library" internally. When processing new design requests, during the prompt generation process, the server can dynamically select different natural language templates based on the similarity between the current constituent information and common patterns, or add constraint statements such as "refer to the topology used in similar past scenarios" to the prompt statements, thereby guiding the generative artificial intelligence model to converge to a reasonable solution more quickly in the search space. This approach makes the construction of prompt statements no longer a fixed template, but a dynamic adjustment process based on historical data and feature analysis, thus improving the consistency and accuracy of model output.

[0282] By employing the aforementioned structured data management and adaptive prompt generation mechanism, the server achieves fine-grained control over the input of generative AI models compared to the traditional method of manually writing prompts or generating simple parameter lists based on fixed rules. By embedding structured context, historical pattern information, and explicit output format constraints into the prompts, the server enables the attention mechanism within the generative AI model to operate in a more focused feature space. This reduces the generation of irrelevant content, shortens the generation length, lowers communication load, and improves the consistency between the output results and network design requirements. Since the prompts and constituent data are uniformly stored in the relational data management device, the server can reuse some context in subsequent requests, thereby reducing redundant computational calls to the model and further improving overall computational efficiency.

[0283] In terms of technical effectiveness, the server, through the aforementioned modular data flow and algorithm design, not only automates the design process of human experts but also optimizes the internal processing path of the computer. The server restructures the originally fragmented "data input—template matching—graphics rendering" process into a pipeline of "structured storage—prompt statement generation—deep model inference—structured parsing—display control generation." This pipeline uses a relational data management device for state management, avoiding multiple format conversions and redundant calculations. Because the server optimizes the generation of prompt statements based on historical information, the generative AI model can perform probability calculations within a smaller search space during the inference phase, reducing the number of calculations for invalid branches, thereby accelerating inference speed and reducing the probability of error generation. Through structured parsing and rule mapping mechanisms, the server confines the model output within a specific semantic and format range, helping to reduce the risk of parsing errors and topological inconsistencies, thus improving the overall reliability and maintainability of the system.

[0284] In other implementations, the server can employ different relational data management devices, such as column-oriented databases or graph databases, to more efficiently store and query topology data. Regarding generative artificial intelligence models, the server can also use models further fine-tuned on communication network datasets, enabling them to achieve higher accuracy in recognizing device terminology and topology patterns. The server can also introduce additional verification modules to perform simple connectivity checks and basic capacity constraint checks on the generated constituent data. If discrepancies are found, the server automatically adjusts the prompts and re-invokes the generative artificial intelligence model to form a closed-loop correction mechanism. Alternatively, the server can combine the topology output by the model with traditional heuristic algorithms (such as minimum spanning tree algorithms and shortest path algorithms) to perform secondary optimization on critical connections, further improving the technical performance of the network design.

[0285] In other implementations, the terminal can not only display static configuration diagrams but also perform interactive simulations based on display control data provided by the server. For example, it can display metrics such as link bandwidth and device port utilization, helping users evaluate the quality of network design from a technical perspective. Users can select different design versions on the terminal, which requests the corresponding configuration data and prompts from the server. The server then provides inter-version difference analysis information, forming an iterative optimization process based on technical performance evaluation.

[0286] Through the above implementation forms, the collaboration between servers, terminals and users makes the generation process of communication network structure diagram no longer just a simple automation of manual design, but improves the technical performance of computer systems in terms of information representation, model input control, inference efficiency and result availability through the combination of generative artificial intelligence models and structured data management. Thus, it realizes the improvement of computer technology itself in the specific technical field of communication network design.

[0287] use Figure 13 The processing procedure is explained.

[0288] Step 1: The user enters the communication network import status on the terminal. Users access the form interface provided by the server through a browser within the terminal's graphical user interface. Users enter text and values ​​in input boxes such as "Equipment Type," "Equipment Model," "Manufacturing Category," "Quantity," "Installation Floor," and "Business Requirements." The input data is a mixture of natural language strings and numerical values.

[0289] Input: Raw text and numerical data entered by the user in various form controls.

[0290] The terminal reads the values ​​of these form controls, mapping each input field to a key-value pair, such as "device_type", "model", "vendor", "quantity", "floor", "requirement_text", etc. The terminal then constructs a structured data object in local memory.

[0291] Output: A structured import status data object stored in the terminal memory.

[0292] Step 2: The terminal will encapsulate the imported status data and send it to the server. The terminal uses a front-end script to serialize the structured data object generated in step 1 into a string format, such as JSON text. During serialization, the terminal performs type checking on numeric fields and fills empty fields with default values ​​or deletes invalid keys.

[0293] Input: A structured import status data object in terminal memory.

[0294] The terminal invokes a network communication interface (such as XMLHttpRequest or fetch), constructs an HTTP / HTTPSPOST request, places the serialized string into the request body, and sets the content type header. The terminal then sends this request over the network to the interface path specified by the server.

[0295] Output: A network request message containing import status data is transmitted to the server.

[0296] Step 3: The server receives the import status data and performs format validation. The server receives HTTP / HTTPS requests from the terminal via a network interface. Within the web application framework, the server parses the request body, restoring the serialized string to its internal data structure (such as a map or dictionary).

[0297] Input: The serialized import status string in the network request message.

[0298] The server uses a data validation module to check the parsed fields, including: whether the number of devices is a positive integer, whether required fields exist, whether the string length is within a preset range, and whether the device type is in the allowed list. Internally, the server performs conditional checks and simple arithmetic operations (e.g., quantity > 0), recording error messages for fields that do not meet the rules. If validation fails, the server generates an error response and returns it to the terminal; if validation passes, the server retains the normalized import status data in memory for subsequent processing.

[0299] Output: A normalized import status data object that has been validated, or response data containing error information.

[0300] Step 4: The server will import status data into the relational data management device. After the imported status data is verified, the server uses the database driver to connect to the relational data management device. The server extracts fields such as design identifier, number of users, number of floors, and business requirement description from the normalized data object, and constructs an insert statement to write these fields into the design master table. At the same time, the server traverses the equipment list, constructs a record for each device, and writes fields such as equipment type, model, manufacturing category, quantity, and installation location into the equipment information table.

[0301] Input: Normalized import status data object in server memory.

[0302] The server performs database insert operations, mapping each record's key-value pairs to table columns and generating a unique design identifier in the database or using an auto-incrementing primary key. After insertion, the server retains this design identifier in memory for subsequent prompt generation and model invocation.

[0303] Output: Import status record stored in the relational data management device, and design identifier associated with the record.

[0304] Step 5: The server reads the constituent information from the relational data management device and combines it into an internal representation. Based on the design identifier obtained in step 4, the server queries the design master table and equipment information table from the relational data management device. The server reads all records corresponding to the design identifier through database query statements and converts the query results into internal data structures, such as a combination of lists and dictionaries.

[0305] Input: Design identifiers and import status records already stored in the relationship data management device.

[0306] The server constructs a "composition information" object in memory, directly assigning values ​​to scalar fields such as the number of users and the number of floors, and combining multiple device records into a device list structure. At this time, the server can calculate some derived characteristics, such as the total number of devices, the number of devices per floor, and the distribution of different device types, as auxiliary information for generating subsequent prompt statements.

[0307] Output: A constituent information object containing basic fields and derived features.

[0308] Step 6: The server generates a prompt statement based on the configuration information and historical records. The server reads the number of users, number of floors, equipment list, and business requirements from the constituent information objects. At the same time, it reads statistical information (such as common topology patterns) related to similar user scales or business types from the historical feature table or prompt statement table in the relational data management device.

[0309] Input: The current constituent information object, as well as historical constituent information and prompt statement records in the database.

[0310] In the prompt generation unit, the server fills the current constituent information into a predefined natural language template to form the basic text, and dynamically adjusts the level of detail and output requirements based on historical statistical results. For example, the server may insert a recommended structure description for a specific scale or specify output format restrictions into the template. The server then combines these contents into a complete prompt text through string concatenation, placeholder replacement, and conditional selection statements.

[0311] Output: The prompt text generated for the current design request, and optional system role description text.

[0312] Step 7: The server inputs the prompts into the generative AI model and sets the inference parameters. The server constructs a model request through the model invocation unit. The model request includes system role descriptions, user role prompts, and inference parameters (such as temperature, maximum output length, and sampling strategy).

[0313] Input: Prompt text, system role description text, and preset or dynamically determined model inference parameters.

[0314] The server encodes these inputs into a request data structure and sends the request using the interface provided by the external model service. Before sending, the server can adjust the inference parameters based on the complexity of the current constituent information, such as increasing or decreasing the temperature value, or setting a higher maximum output length to accommodate complex topology descriptions. The server transmits the request data to the generative artificial intelligence model server via a network interface.

[0315] Output: A model request containing prompts and inference parameters is sent to the generative artificial intelligence model server.

[0316] Step 8: The server receives the output of the generative artificial intelligence model and parses it into constituent data. The server receives response text from the generative artificial intelligence model server. The response text typically includes network topology descriptions, device role lists, connection relationship lists, VLAN segmentation, and IP planning suggestions.

[0317] Input: The response text returned by the generative artificial intelligence model.

[0318] In the text parsing module, the server segments and tags the response text. The server can first segment the text according to expected titles or key phrases (such as "device role," "connection relationship," "VLAN segmentation," "IP planning"), and then use regular expressions and string splitting functions to parse each device description and connection relationship into a structured record. During the parsing process, the server maps role descriptions to standard categories according to predefined rules and extracts port descriptions as attribute fields.

[0319] Output: A structured data object representing the communication network graph, including a list of nodes, a list of connections, and address planning data.

[0320] Step 9: The server stores the model output, prompts, and composition information together as historical information. After the data is generated, the server retrieves the corresponding prompt text, original structure information, current timestamp, and design identifier from memory.

[0321] Inputs: structured data, prompt text, information objects, and design identifiers.

[0322] The server uses a database interface to write the prompt text to a prompt table, writes the original text and parsed constituent data of the model output to a model output table, and records statistical characteristic indicators, such as the number of devices, the number of topology levels, and the generated response length, in a historical feature table. During the recording process, the server performs simple counting, aggregation operations, and tag generation, providing foundational data for subsequent statistical analysis and prompt optimization.

[0323] Output: Historical information records stored in the relational data management device, including prompts, model outputs, and related features.

[0324] Step 10: The server generates display control data based on the constituent data. The server extracts a list of nodes and a list of connections from the constituent data objects. Within the data generation unit, the server performs layout calculations for the nodes, assigning them to different logical layers based on their roles (core layer, aggregation layer, access layer) and installation floor information.

[0325] Input: Structured data objects.

[0326] The server executes a layout algorithm, uniformly distributing the horizontal positions of nodes in each layer and assigning appropriate connection paths to avoid excessive line overlap. The server specifies display attributes (such as shape, color, and label text) for each node and line type and direction attributes for each connection, generating node coordinate tables and connection tables. The server combines these results into display control data, representing it as structured data suitable for terminal rendering.

[0327] Output: Display control data including node coordinates, connection information, and display style parameters.

[0328] Step 11: The server generates output information and sends it to the terminal. After generating the display control data, the server combines it with the textual descriptions generated by the model (such as network structure descriptions and address planning descriptions) to form an output information object.

[0329] Input: Display control data and text descriptions.

[0330] The server uses a web application framework to serialize the output information object into a response format (such as JSON text with accompanying descriptive fields) and sends it as the HTTP / HTTPS response body to the requesting terminal. The server specifies the content type and encoding format in the response header so that the terminal can parse it correctly.

[0331] Output: A response message containing display control data and explanatory text is transmitted to the terminal.

[0332] Step 12: The terminal parses the output information and renders a diagram of the communication network in the interface. The terminal receives the response message from the server, and the script running in the browser parses the response body, restoring the display control data and explanatory text into the internal data structure.

[0333] Input: Display control data and explanatory text in the response message sent by the server.

[0334] The terminal uses a front-end drawing library to convert node coordinates and connection information into graphical elements, drawing icons for nodes such as routers and switches in the drawing area, and setting their colors, shapes, and labels according to display style parameters. The terminal also displays network structure descriptions and configuration suggestions provided by the model in a text area. The terminal allows users to zoom and drag the view using a mouse or touch input to view the topology of different areas.

[0335] Output: A diagram of the communication network structure and a text description interface displayed on the terminal screen.

[0336] Step 13: Users can evaluate the displayed results and choose to modify the import status. Users can observe the communication network structure diagram and text description on the terminal to assess whether the device role allocation, connection relationship and address planning meet their needs.

[0337] Input: The diagram and explanatory text displayed in the terminal interface.

[0338] Users can modify the import status in the input form on the terminal, such as increasing the number of switches on a certain floor, changing the location of core equipment, or adjusting the description of business requirements. After the user completes the modifications and submits them again, the terminal re-executes the input and sending process of steps 1 and 2, and the server then re-executes the subsequent steps to generate a new configuration diagram.

[0339] Output: The updated import status input, and the new round of server-side data processing and network design generation process triggered by it.

[0340] Application Example 2 The process flow corresponding to the specific processing in Use Case 2 will be described below. The various parts of the system described below are implemented by the data processing device 12 and the intelligent device 14. In addition, the data processing device 12 is referred to as the "server" and the intelligent device 14 is referred to as the "terminal".

[0341] In traditional network design and management, network topology generation and optimization primarily rely on technical personnel with specialized knowledge. Computer systems are typically used only as static drawing tools or configuration entry tools, performing only fixed rule matching and simple visualization at the data processing level. They cannot automatically generate reasonable topologies based on dynamic network conditions, nor can they adaptively adjust output content according to different users' understanding and emotional states. This leads to the following specific problems at the computer technology level: (1) In the data processing link, although the server can receive network device information and configuration parameters, it lacks a mechanism to automatically convert these heterogeneous data into high-quality prompts that can be directly used by the generative artificial intelligence model. As a result, the output of the generative artificial intelligence model in the network scenario is unstable and uncontrollable. The server cannot form a reusable automated data to topology generation pipeline, resulting in low efficiency of computing resources and inconsistent quality of generation results.

[0342] (2) In terms of topology generation and visualization, traditional systems rely on fixed templates or manual drawing. The server side lacks a technical path to automatically extract network nodes and connection relationships from the natural language output returned by the generative artificial intelligence model and automatically generate structured topology data. This makes it difficult to automate the network topology generation process, and the server cannot give full play to the ability of the graphics rendering program in automatic layout and complex structure expression.

[0343] (3) In terms of human-computer interaction and output control, existing systems generally output network structure and technical descriptions in a fixed format. The server cannot perceive the user's knowledge level and emotional state, and can only provide results with a single level of complexity. For users who lack network knowledge or are under high pressure, a large amount of detailed information and technical terms will increase the burden of understanding. From the perspective of computer systems, there is a lack of dynamic scheduling mechanism for "output content granularity and complexity". The system lacks the ability to use user status signals to provide feedback control for subsequent computing tasks (prompt statement generation, topology simplification / refinement).

[0344] (4) In terms of the overall system architecture, the server and the terminal only perform linear data transmission of “form → configuration → result”, without forming a closed-loop adaptive computing process: the user input and emotion-related information collected by the terminal are not systematically used in the server to adjust the calling strategy, topology visualization strategy and explanatory information expression strategy of the generative artificial intelligence model, thus limiting the computer system’s ability to continuously optimize output quality in multiple rounds of interaction.

[0345] Therefore, a new computer implementation is needed to enable the server to: - Automatically converts network device information and communication requirements into high-quality prompts, stably driving generative artificial intelligence models to generate results suitable for the network topology; - Automatically construct machine-readable network topology data from natural language output and efficiently collaborate with graphics rendering programs to automatically generate visual topology maps; - Based on user input and emotion-related information, the complexity and display content of prompts and network topology diagrams are dynamically adjusted on the server side to achieve fine-grained control over the output; - By collecting data from the terminal and inferring from the server, an adaptive computing closed loop is formed with user status as the feedback signal, thereby improving the automation level, resource utilization efficiency, and human-computer interaction effect of network topology generation and display at the computer technology level.

[0346] The specific processing performed by the specific processing unit 290 of the data processing apparatus 12 in Application Example 2 is achieved by the following means.

[0347] In this invention, the server includes: a data receiving and preprocessing module for receiving network device information and communication requirement information from external terminals and generating structured data; an information management and requirement extraction module for storing the structured data in an information management area and aggregating it to extract communication topology requirements; a prompting statement generation module for automatically generating prompting statements for input to a generative artificial intelligence model based on the communication topology requirements and user attribute information; a model calling and result parsing module for sending a request containing the prompting statements to the generative artificial intelligence model, receiving natural language output, and parsing the natural language output into network topology data containing connection object elements and connection relationships; a topology visualization generation module for calling a visualization processing program based on the network topology data to generate network topology map drawing data and maintaining it in a format that can be output to an external terminal; and an emotion and comprehension adaptive control module for inferring the user's emotional state and comprehension level based on dialogue input and emotion-related information obtained from the user, and dynamically adjusting the level of detail and display content of the prompting statements and network topology map according to the inference results, and further for sending the adjusted network topology map and explanatory information to an external terminal for presentation. This allows for the formation of an end-to-end computing pipeline within the server, encompassing everything from raw network data to prompt generation, generative AI model inference, topology data extraction, automatic graphics rendering, and output adjustment based on user emotion and understanding. This not only improves the efficiency and stability of automatic network topology generation and visualization but also enhances the computer system's processing power and resource utilization in human-computer interaction scenarios by introducing a user-state-driven adaptive control mechanism. Consequently, it effectively addresses the issues of low automation in network design, uncontrollable output, and difficulty in adapting to different user comprehension levels in traditional systems from a computer technology perspective.

[0348] A "system" refers to a whole consisting of one or more information processing devices, external terminals, and software and hardware for data interaction between them, which is a technical solution for achieving automatic generation, visualization, and adaptive interactive processing of network topology.

[0349] "Information processing device" refers to a computing device that has a processor, storage unit and communication interface, and is capable of executing programs to perform calculations, storage and output control on input data. It can be a server, computing node or other computing platform.

[0350] "External terminal" refers to a user-side device that is connected to the information processing device through a communication network and is used to input network-related information and receive network topology diagrams and explanatory information, including but not limited to computing terminals, mobile terminals, or industrial field terminals.

[0351] "Network device information" refers to attribute data related to various devices used in the network, including device type, model, number of interfaces, address information, deployment location and functional role, etc., which are used to describe the technical information of network components.

[0352] "Communication requirements information" refers to the constraints and preferences related to the performance and functions required by the target network, including bandwidth requirements, latency tolerance, reliability level, security requirements, and scalability requirements, which are used to guide network topology design.

[0353] "Structured data" refers to network device information and communication requirements information organized according to a predefined data format or data model, which facilitates storage, retrieval, and programmed processing in information processing devices.

[0354] "Storage unit" refers to a storage resource used to permanently or temporarily store programs and data, including database systems, file storage systems, or other forms of data storage devices.

[0355] The "information management area" refers to the logical storage area within the storage unit used to centrally store network-imported data, user attribute data, and generated result data, in order to support subsequent aggregation analysis and retrieval processing.

[0356] "Network Import Status" refers to comprehensive information such as the deployment status, connection relationships, and operating status of network devices in the target environment, which is used to reflect the actual composition and usage of the current network.

[0357] "Communication topology requirements" refers to a set of design requirements extracted from network device information and communication requirement information, used to describe the connection structure, hierarchical relationship and path constraints between nodes in the network.

[0358] "User attribute information" refers to characteristic data related to the user, including but not limited to technical proficiency, role type, and historical operation records, which are used for personalized and adaptive control when generating prompts and output content.

[0359] "Generative artificial intelligence models" refer to artificial intelligence models trained through machine learning methods that can automatically generate text content or other output results based on input text or structured data, and are used to generate network topology schemes and explanatory information.

[0360] "Prompt statements" refer to text instructions or descriptions input into generative artificial intelligence models, used to describe the current task context, input data summary, and expected output format, thereby guiding the model to generate results that meet the requirements.

[0361] "Natural language output" refers to the results generated by generative artificial intelligence models based on prompts and expressed in natural language text, including descriptions of network topology, connectivity, and design rationale.

[0362] "Network topology data" refers to network structure information parsed from natural language output and represented in a machine-readable manner, including network nodes, node attributes, connection relationships, and related parameters.

[0363] "Connection object elements" refer to entities in network topology data that represent network components, including routing devices, switching devices, terminal devices, and other network functional entities.

[0364] "Connection relationship" refers to the association information in network topology data that describes the physical or logical connections between different connected object elements, including connection direction, connection type and related attributes.

[0365] A "visualization processor" refers to a program component used to convert network topology data into a graphical representation, generating network topology diagrams through graphical layout algorithms and drawing instructions.

[0366] A "network topology diagram" is a graphical representation of network nodes and their connections, used to visually demonstrate the network structure hierarchy and communication paths.

[0367] “Drawing data” refers to graphical description data or image files generated by visualization processing programs for presenting network topology diagrams on display devices, including vector graphics data, bitmap data, or renderable graphical description languages.

[0368] "Conversational input" refers to interactive content that users send to an information processing device via an external terminal in the form of natural language or text, including questions, instructions, feedback, and modification requests.

[0369] "Emotion-related information" refers to input data related to a user's emotions or psychological state, including emotional features in the user's text, voice tone features, facial expression features, or emotion tags uploaded by the terminal.

[0370] "Emotional state" refers to the category of user emotion inferred by the information processing device after analyzing emotion-related information, such as tension, confusion, calmness, or happiness, which is used to adjust the system's output strategy.

[0371] "Understanding level" refers to the depth of a user's grasp of the current network topology and explanatory information, as inferred by the information processing device based on the user's interaction history, question content, and emotional state. It is used to control the complexity of the output content.

[0372] "Level of detail" refers to the granularity of presentation of technical details, technical terms, and structural elements in the output of network topology diagrams and explanatory information, including different levels such as simplified and detailed presentations.

[0373] "Displayed content" refers to the collection of graphical elements and text descriptions presented to the user on an external terminal, including nodes, connecting lines, annotations, and corresponding explanatory text in the network topology diagram.

[0374] "Explanatory information" refers to natural language text generated based on the output of generative artificial intelligence models and internal system analysis, which describes the network topology, design rationale, and usage precautions to help users understand the network structure.

[0375] The "output control module" refers to a functional component in an information processing device that comprehensively controls the output format, level of detail, and presentation method of network topology diagrams and explanatory information based on the user's emotional state, level of understanding, and internal system strategies.

[0376] In the following embodiments, the server, terminal, and user are described as the main actors. The system of the present invention can be deployed in factory network planning environments, enterprise office network design environments, and other environments that require automatic generation and adaptive display of network topology.

[0377] I. System Overall Structure The server is the subject: A server includes a processor, storage units, and a network interface. The processor can be a multi-core general-purpose processor or a processor equipped with vector processing units. The storage units include main memory and persistent storage. The server runs multiple software components on an operating system, including: - Web service framework components used to provide HTTP / HTTPS communication interfaces with the terminal; - Database management components for interacting with relational database management systems; - Text processing component for natural language processing of network device information and user input; - Generative AI model invocation component, used for data interaction with externally or locally deployed generative AI models; - Topology parsing and data structure building components are used to convert natural language output into network topology data structures; - A graphical visualization component used to invoke graphical layout algorithms and generate network topology diagrams; - Sentiment and understanding estimation component, used to infer user state based on user input and sentiment-related information; - Output control component, used to adjust the content of prompts and network topology diagrams according to user status.

[0378] Terminal as subject: The terminal includes hardware components such as a display device, input device, camera, and microphone, and runs a browser or dedicated client program. The terminal connects to a server via a communication network to input network-related information, display network topology diagrams, and collect user-related sentiment information.

[0379] User as subject: Users operate the terminal, inputting network device information, communication requirements, problems, and feedback into the interface, and viewing the network topology diagram and explanatory information generated by the server.

[0380] II. Data Structure and Storage Method The server is the subject: The server creates multiple logical tables in the storage unit to store structured data. Exemplary table structures include: - Device table: Records fields such as device identifier, device type, device model, number of interfaces, logical role, physical location, and address information; - Requirements Table: Records the communication requirements associated with each network design task, including fields such as bandwidth target, latency constraints, reliability level, and security policy requirements; - Topology table: Records candidate or final network topology data, including fields such as node identifier, connection endpoint, link type, bandwidth suggestion, and logical segmentation information; - User Attribute Table: Records fields such as user role identifier, historical interaction statistics, and preference parameters (such as simplified or detailed descriptions of preferences); - Sentiment Status Table: Records sentiment tags, confidence levels, and corresponding summaries of input content, linked by time.

[0381] After receiving raw input from the terminal, the server parses it into structured data objects and writes them into the aforementioned table structure. This explicit data structure allows the server to quickly retrieve the necessary information in subsequent processing stages, thereby shortening the query time required for prompt generation and topology generation, and ultimately improving processing speed.

[0382] III. Prompt Statement Generation and Generative Artificial Intelligence Model Structure The server is the subject: The server uses a text processing component to summarize and transform structured data. When constructing prompts for each network design task, the server performs the following data processing: - The server groups the devices in the device table by type and counts the number of devices of each type, typical parameters (such as number of ports and speed), and their logical roles in the network. - The server extracts key fields related to latency, bandwidth, and security from the requirements table and generates concise natural language constraint descriptions. - The server retrieves the user's technical proficiency tag from the user attribute table to determine whether to request the generative AI model to output a highly or unprofessionally professional result in the prompt statement; - The server determines the style of the prompt statement (e.g., guiding the model to output a detailed or simplified statement) based on the sentiment label in the sentiment state table or within the weighted time window.

[0383] Generative AI models are used as the subject (their structure is described by the server): In this implementation, the generative artificial intelligence model can employ a multi-layer self-attention neural network based on the Transformer architecture. The model includes: - Word embedding layer, used to map discrete word symbols in input prompts into high-dimensional vectors; - A multi-layer encoder and decoder stacked structure, each layer containing a multi-head self-attention sub-layer and a feedforward sub-layer; - Layer normalization and residual connections are used to stabilize training and improve convergence speed; - Output probability distribution layer, used to predict the probability of the next word based on the preceding context.

[0384] When training this generative AI model, the server uses a corpus containing a large amount of network design documents, network architecture descriptions, and related explanatory texts as training data. The server defines the loss function as cross-entropy loss and uses gradient descent-type optimization algorithms and their variants to update the model parameters. During training, the server can perform data augmentation, such as paraphrasing network device descriptions, randomly inserting or deleting secondary information, to improve the model's robustness under different representation styles.

[0385] When the server invokes a generative AI model, it inputs prompts into the model's encoding, controlling the generation length and temperature parameters to balance the diversity and stability of the output. In this way, the server obtains natural language output containing network design schemes and explanations.

[0386] IV. Natural Language Output Parsing and Topology Data Construction The server is the subject: After receiving the natural language output from the generative artificial intelligence model, the server processes it using topology parsing and data structure construction components. The server can pre-define the output format in the prompt, such as requiring the model to output in a segmented list or semi-structured form. The server performs the following specific operations during parsing: - The server uses a pattern matching method to divide the text into device description fragments and connection description fragments based on a predefined set of keywords (such as "device", "connection", "from...to...", "link speed", etc.); - The server extracts the model, type, and role appearing in the device description fragment and matches them with existing device records in the database to establish a device identifier mapping; - The server extracts information such as "source device", "target device", "port number", and "bandwidth" from the connection description fragment, constructs an edge data structure, and writes it into the topology table; During the parsing process, the server merges and checks for conflicts among multiple descriptions of the same pair of devices. If inconsistent bandwidth recommendations are found, a record is selected based on preset priority or confidence rules.

[0387] Through this parsing strategy, the server transforms the topological information contained in the output of the generative artificial intelligence model into explicit sets of nodes and edges, and stores them in memory and a database in the form of a graph structure. Compared with traditional graph editing based on fixed templates, this parsing process can automatically adapt to different expression styles, improve the coverage and flexibility of topology generation, and thus technically improve the accuracy and robustness of automatic network topology generation.

[0388] V. Graphic Visualization Processing and Technical Effects The server is the subject: The server invokes the graphics visualization component to pass topology data to the graphics layout algorithm. The server can use hierarchical layout algorithms, force-directed layout algorithms, or other graphics layout methods. During the visualization process: - Map each device node to a primitive in the graph, and use different shapes or colors depending on the node type; - Map each link to an edge primitive and adjust the line thickness or style according to bandwidth or importance; - Based on the network hierarchy, core devices are placed in the upper layer and access devices are placed in the lower layer to reduce overlap and improve readability; For sections with redundant links, the server automatically increases the spacing between nodes using a layout algorithm, making the redundant structure clearly identifiable.

[0389] This automatic layout and attribute mapping process, compared to manual drawing or drawing based on simple rules, can significantly reduce layout time in large-scale network scenarios while maintaining the structural clarity of the graphics. This allows the server to maintain an acceptable response time when processing complex topologies, thereby achieving technical improvements in processing speed and readability.

[0390] VI. Mechanism for Emotional and Comprehension Inference Terminal as subject: The terminal uses a camera and microphone to capture the user's facial images and voice signals, compresses and encodes the data locally, and then sends it to the server. The terminal can also record the user's text input on the interface, including complaints, questions, or affirmative statements.

[0391] The server is the subject: The server uses two sub-models in the sentiment and understanding inference component: - Text sentiment classification sub-model: The user's text sequence is encoded using a convolutional neural network or a bidirectional recurrent neural network, and the sentiment category (such as tension, confusion, calm, happiness, etc.) is output through a fully connected layer and Softmax.

[0392] - Speech and facial expression emotion recognition sub-model: Convolutional neural network is used to extract features from image frames, and a temporal model is used to analyze speech features, thereby outputting emotion labels and confidence scores.

[0393] The server aligns the outputs from the two sub-models on the timeline and uses a weighted fusion strategy to obtain a comprehensive sentiment state. It also estimates the user's level of understanding by combining metrics such as the frequency of historical user questions and the proportion of repeated questions. The server stores the sentiment state and level of understanding in a sentiment state table for subsequent modules to access.

[0394] This structured emotion and understanding inference mechanism enables the server to input user status as a feedback signal into subsequent calculation paths, thereby achieving adaptive control in the prompt generation and topology visualization stages, avoiding a one-size-fits-all fixed output strategy, and improving the overall human-computer interaction quality of the system.

[0395] VII. Adaptive Output Control and Technical Improvements The server is the subject: The server performs the following technical processing based on the emotional state and level of understanding in the output control component: - If a user is judged to be "nervous" and has a low level of understanding, the server explicitly requires the generative AI model to use concise language when generating prompts, and automatically hides non-critical nodes, redundant links and low-level details during the visualization stage. - If the user is judged to be "calm" and has a high level of understanding, the server will request the model output in the prompt statement to include detailed descriptions of address design, routing policies, and redundancy schemes, and display information such as interface identifiers and VLAN segmentation marks in the topology diagram.

[0396] For example, the server can generate the following prompt: Example 1 (for users with lower comprehension levels): "Please assume that the user asking the question has almost no basic knowledge of the internet."

[0397] Based on the following network equipment and requirements: 1. A routing device, model R-1; 2. One switching device, model S-1; 3. One automated device, model A-1, used for production line control; Please explain how to connect these devices in very plain simplified Chinese, and provide a simplified network topology description. Avoid using technical abbreviations, and keep the output length within 300 characters. Example 2 (for users with a higher level of understanding): "You are a network architect."

[0398] Existing equipment: 1. Router R-1 serves as the network exit point; 2. Switching equipment S-1, equipped with 24 gigabit ports; 3. Automated equipment A-1, located in the production line area, is sensitive to time delays.

[0399] Please design the network topology and provide the following: 1. Suggested subnetting and address allocation scheme; 2. Routing control strategies for production line flow; 3. Recommendations for redundant link design.

[0400] Please output the results in item form, along with a brief technical explanation. By structurally adjusting the prompt statements, the server effectively constrains the output format and content granularity of the generative artificial intelligence model without changing its internal parameters, making subsequent parsing and visualization processing more stable, thereby achieving overall optimization of the computation process.

[0401] VIII. Examples of Specific Application Scenarios User as subject: When deploying a new production line in a factory, users enter equipment information in the terminal interface, such as "Adding a new robotic arm, model RA-2023, deployed on production line 1; existing switching equipment model SW-100; existing routing equipment model RT-200 for upper-layer network connection," and select the "Prioritize real-time control" option. After the user submits the information, the terminal converts it into structured data and sends it to the server.

[0402] The server is the subject: The server generates prompts based on the structured data and invokes a generative artificial intelligence model. The model's results include statements such as "The connection between the robot and the switching equipment needs to use a separate VLAN" and "It is recommended to set up redundant links between the switching equipment and the routing equipment." The server parses these results and generates the corresponding topology data structure, then uses a graphical visualization component to generate a network topology diagram. Simultaneously, the server uses sentiment and understanding inference components to analyze the user's historical interaction records. If the user repeatedly provides feedback such as "I don't understand" or "It's too complicated," the server selects a simplified mode in the current output, displaying only core devices and major links.

[0403] Terminal as subject: The terminal receives drawing data and explanatory information from the server, renders a simplified network topology diagram, and displays explanatory text written in plain language on the side panel. For example, the explanatory text may only mention that "the robot connects to the routing device through the switching equipment to form a secure and stable control network," without showing specific address planning. In this way, users can quickly understand the network structure on-site and perform physical cabling and equipment configuration accordingly.

[0404] IX. Technical Effects and Causal Relationships The server is the subject: By connecting structured data, prompt generation, generative AI model inference, natural language parsing, topology data construction, visualization, and emotion-adaptive control into an end-to-end computing pipeline, the server brings the following effects at the computer technology level: - In terms of processing speed, the server avoids manual data entry and drawing by using predefined data structures and parsing strategies, which significantly shortens the time for large-scale topology generation. - In terms of accuracy, the server uses generative artificial intelligence models to model complex design knowledge and improves the consistency and parsability of the output through structured parsing. Compared with traditional static template-based systems, it can obtain network design results that are closer to the best practices. - In terms of data management, the server stores information such as network import status, design requirements, and user status in a unified information management area, providing a data foundation for subsequent version comparison, historical review, and automatic optimization; - In terms of communication load, the server reduces the need for continuous transmission of large-scale multimedia data by performing preliminary compression and structuring on the terminal side, transmitting only the necessary structured data and emotion-related features.

[0405] This improvement is not merely a simple automation of manual operations, but rather introduces a series of unconventional computational steps within the server, such as adaptive prompt statement construction, parsable natural language output, and automatic topology layout. By combining the generative capabilities of generative artificial intelligence models with structured parsing algorithms, it forms a high-frequency iteration and large-scale trial capability that is difficult to achieve in human expert workflows, thereby achieving a technical improvement over traditional computer systems in terms of computational accuracy, stability, and scalability.

[0406] 10. Other Implementation Methods and Variations The server is the subject: In other implementations, the server can deploy the generative AI model on dedicated local acceleration hardware to reduce external network call latency and further improve response speed. The server can also use other types of neural network architectures, such as adding graph neural network layers to process intermediate topology data for evaluating topology reliability or cost metrics.

[0407] Terminal as subject: In another implementation, the terminal can run a lightweight emotion recognition model locally, sending only emotion tags to the server, thus achieving a better balance between privacy and bandwidth.

[0408] User as subject: In another implementation, users can choose "Expert Mode," where the server no longer restricts the use of technical terminology in the prompts, but instead requires the output to be detailed, similar to the technical specifications, so that experienced network engineers can directly generate configuration scripts based on the output.

[0409] The above-mentioned embodiments can be freely combined and replaced according to the specific usage environment. As long as the core technical elements such as data flow between the server, terminal and user, prompt statement generation, generative artificial intelligence model invocation, topology parsing, visualization processing and emotion adaptive control are realized, they are considered to fall within the protection scope of this invention.

[0410] use Figure 14 The processing procedure is explained.

[0411] Step 1: The user enters network-related information on the terminal. User as subject: Users can input network device information and communication requirements through form controls or dialog windows in the graphical interface of the terminal.

[0412] Input: Device name, device type (e.g., routing device, switching device, terminal device), device model, number of ports, deployment location, IP address (if any), bandwidth requirements, latency requirements, reliability level, security requirements, and other text or option data.

[0413] Specific actions: Users can type text in different input boxes or select parameters from drop-down menus, and can also enter additional descriptions in the free text area, such as "I want to control network priority in real time" or "Please explain the result in a simple way".

[0414] Output: The terminal receives a set of raw input data objects, which are used for subsequent preprocessing and sent to the server.

[0415] Step 2: The terminal preprocesses the input data and packages it into a structured format. Terminal as subject: The terminal uses local scripts to perform format checks and structure transformations on user input.

[0416] Input: The raw text and option data entered by the user on the interface in step 1.

[0417] Data processing and calculation: The terminal maps each field to a predefined data key, such as mapping "device type" to "type", "device model" to "model", and "bandwidth requirement" to "bandwidth_requirement". The terminal uses regular expressions to validate the IP address format, convert the number of ports from a string to an integer, and parse the bandwidth requirement from a string (such as "100Mbps") into a value and unit. If a required field is missing, the terminal displays an error message on the interface and does not send a request.

[0418] Output: Validated and formatted structured data objects (such as key-value pair collections) are stored in memory with a uniform structure, ready to be sent to the server.

[0419] Step 3: The terminal sends a network design request to the server. Terminal as subject: The terminal sends structured data to the interface provided by the server through the network communication interface.

[0420] Input: The structured data object generated in step 2.

[0421] Data processing and data computation: The terminal serializes structured data objects into text format (such as JSON strings), constructs HTTP / HTTPS request messages, writes content type and authentication information into the request header, and writes the serialization result into the request body; the terminal sends the request to the server address through the network protocol stack.

[0422] Output: The network request message sent by the terminal is received by the server port and enters the processing flow; at the same time, the terminal switches to the "processing" state on the interface and waits for the server response.

[0423] Step 4: The server parses the request and stores the input data. The server is the subject: The server receives requests from the terminal through a web service framework, parses them, and persists them.

[0424] Input: The HTTP / HTTPS request message sent by the terminal in step 3, which contains a serialized structured data string.

[0425] Data processing and computation: The server uses the network stack to receive requests and parses the request body into a string; the server uses a parsing library to deserialize the string into internal data structures (such as dictionaries or objects); the server validates each field (checking whether the device type is in the supported list, whether the IP format is correct, and whether the number of ports is positive); the server maps the data to database table fields, generates several insert statements, and interacts with the relational database through the database driver to insert the device information and requirement information into the corresponding tables; after the transaction is committed, the server generates a unique task identifier as an index for subsequent processing.

[0426] Output: Standardized records stored in the database, and a task identifier corresponding to this network design task; the server internally returns this identifier to subsequent modules.

[0427] Step 5: The server aggregates task data and extracts communication topology requirements. The server is the subject: The server reads data associated with task identifiers from the database, aggregates it, and abstracts the communication topology requirements.

[0428] Input: Task identifier, and the device table record and requirement table record associated with the task in the database.

[0429] Data processing and calculation: The server categorizes devices by role from the device table (e.g., core layer, aggregation layer, access layer, industrial terminal), and calculates the number of devices and total ports for each category; the server reads numerical fields such as latency requirements, bandwidth thresholds, and reliability levels from the requirements table, and normalizes them into internal requirement parameters (e.g., upper limit of latency, lower limit of bandwidth, redundancy level); the server establishes a communication topology requirement object based on these parameters, including mandatory connection relationships (e.g., control devices must be directly connected to switching devices), optional redundancy relationships, isolation requirements (e.g., isolation between office and production networks), and records it as intermediate data.

[0430] Output: A formalized communication topology requirement object containing several structured fields used to drive the construction of subsequent prompt statements.

[0431] Step 6: The server generates prompts for generative artificial intelligence models. The server is the subject: The server translates communication topology requirements and user attribute information into natural language prompts.

[0432] Input: The communication topology requirement object from step 5, user attribute records stored in the database (such as user professional level, historical preferences), and recent emotional state records (such as emotion category, level of understanding).

[0433] Data processing and calculation: The server determines whether to require the generation of details (such as subnetting and routing strategies) in the prompt statement based on the user's professional level; it determines the tone and output length constraints based on the emotional state; the server uses string templates or text generation rules to combine the device list and requirement summary into a complete task description; the server explicitly requires the output format in the prompt statement, such as using an item list, device and connection segmented descriptions, to facilitate subsequent parsing.

[0434] Output: A complete text-based prompt statement containing the task background, equipment status, requirements and constraints, and output format instructions, which is prepared as input to the generative artificial intelligence model.

[0435] Step 7: The server invokes a generative artificial intelligence model to generate network design text. The server is the subject: The server sends prompts to the generative AI model via the model call component and receives the natural language output generated by the model.

[0436] Input: The prompt statement generated in step 6.

[0437] Data processing and computation: The server encodes the prompts into model input vectors (through word embedding or sub-word encoding) and sends inference requests to the deployed generative AI model; the model internally uses a multi-layer attention network to calculate contextual representations based on the input sequence and gradually generates each label of the output sequence; the server sets generation parameters (such as maximum length, temperature, penalty coefficient) to control the diversity and stability of the output; the server receives the label sequence output by the model and decodes it into natural language text.

[0438] Output: A natural language output text containing network topology suggestions, device connectivity relationships, possible address planning, and design rationale.

[0439] Step 8: The server parses the model output and constructs the network topology data structure. The server is the subject: The server performs semantic parsing on the natural language output to generate a machine-readable network topology.

[0440] Input: The natural language output text returned by the generative artificial intelligence model in step 7.

[0441] Data processing and computation: Based on the pre-defined output format in the prompt statement, the server uses techniques such as pattern matching, keyword recognition, and simple syntax analysis to divide the text into device description and connection relationship sections. The server extracts node names, role types, and possible hierarchical information from the device description and matches them with existing device records in the database. The server identifies source nodes, target nodes, link types, bandwidth suggestions, etc., from the connection relationship description and constructs edge objects. The server combines the set of node objects and the set of edge objects into a graph data structure and detects duplicate or conflicting connections (such as multiple descriptions between a pair of nodes), merging or marking them according to predetermined rules.

[0442] Output: A standardized network topology data structure, including a list of nodes, a list of edges, and related attributes, stored in a topology table and passed to the visualization module.

[0443] Step 9: The server generates drawing data for the network topology diagram. The server is the subject: The server calls a visualization processing program to convert the topology data into visual plotting data.

[0444] Input: The network topology data structure constructed in step 8.

[0445] Data processing and calculation: The server selects an appropriate layout algorithm (such as hierarchical layout) to distribute nodes hierarchically in a planar coordinate system; the server assigns different graphic styles (such as shape and color) to different types of nodes, and assigns line styles and labels (displaying bandwidth or purpose) to edges; the server performs layout calculations to minimize intersecting lines and balance node spacing, thereby obtaining the coordinates of each node and the path of the edge; the server uses a graphics library to encode these graphic elements into drawing data (such as vector graphics descriptions or bitmap pixel data) and saves them as image files or graphic documents.

[0446] Output: The drawing data corresponding to the network topology diagram and its storage path, for terminal download and display.

[0447] Step 10: The terminal collects user emotion-related information and sends it back to the server. Terminal as subject: The terminal collects users' emotional cues through a multimodal approach.

[0448] Input: Text input by the user when viewing the results (questions, comments, etc.), facial images captured by the camera, and voice signals captured by the microphone.

[0449] Data processing and computation: The terminal compresses and encodes images and audio to reduce the amount of data; the terminal can extract simple emotional features (such as speech rate, volume, and facial expression intensity) locally and convert them into tags or values; the terminal packages these emotion-related data or tags with timestamps and sends them to the server through a secure connection.

[0450] Output: Feedback data messages containing user sentiment cues, which the server uses to perform sentiment and comprehension estimations.

[0451] Step 11: The server estimates the user's emotional state and level of understanding. The server is the subject: The server uses sentiment analysis models and statistical methods to infer user states.

[0452] Input: Emotion-related data uploaded by the terminal in step 10, the user's recent text interaction records, and historical usage statistics.

[0453] Data processing and computation: The server uses sentiment classification algorithms to extract emotional tendencies (such as confusion, dissatisfaction, and satisfaction) from text; it uses feature extraction and classifiers to calculate sentiment categories and confidence levels for speech and images; the server performs weighted fusion of results from multiple modalities to calculate the overall sentiment state; the server estimates the level of understanding using rules or simple models based on information such as user question frequency, number of repeated questions, and usage duration; and the server writes the current sentiment state and level of understanding into a sentiment state table.

[0454] Output: A structured record representing the user's emotional state and level of understanding, used to control subsequent output adjustments.

[0455] Step 12: The server adaptively adjusts the prompts and topology display strategy based on the user's status. The server is the subject: The server uses user status information to modify the complexity of subsequent prompts and visual content.

[0456] Input: Emotional state and comprehension level record from step 11, original topological data structure, and explanatory text.

[0457] Data Processing and Calculation: If the server detects that the user is nervous and has a low level of understanding, it reduces technical jargon and limits the output length in the new prompts, and requests the generative AI model to explain in a simplified form; the server filters the topology data, removing secondary nodes and detailed attributes, retaining only the core structure to generate a simplified diagram; if the server detects that the user is in a good mood and has a high level of understanding, it requests the model to supplement advanced information such as address planning and routing strategies in the prompts, and adds VLAN tags, port numbers, etc. to the topology diagram; the server regenerates the explanatory text or topology diagram for use in the next round of interaction.

[0458] Output: A customized prompt template for the current user status, updated topology drawing data, and explanatory text.

[0459] Step 13: The server returns a network topology diagram and explanatory information to the terminal. The server is the subject: The server encapsulates the generated results into a response message and sends it to the terminal.

[0460] Input: The plotting data path, explanatory text, and structured topology summary obtained in step 9 or step 12.

[0461] Data processing and data computation: The server constructs a response body, embeds the image access address or graphic data into the response, and appends explanatory text and structured information in the form of fields; the server sends the response to the terminal via HTTP / HTTPS; the server records the response time and result status to the log for performance monitoring and subsequent optimization.

[0462] Output: A response message containing a network topology diagram and explanatory information, which is received by the terminal for display.

[0463] Step 14: The terminal displays a network topology diagram and provides interactive operations. Terminal as subject: The terminal presents the results returned by the server to the user and allows further operations.

[0464] Input: The response data returned by the server in step 13, including topology drawing data or its access path, descriptive text, and topology summary structure.

[0465] Data processing and calculation: The terminal downloads images from the server or directly renders drawing data according to the image path and displays the network topology diagram on the interface canvas; the terminal displays explanatory text in the side panel; the terminal uses the topology summary structure to establish interactive hotspots for each node and connection, and displays detailed attributes when the user clicks; the terminal selects different display levels according to the mode marked by the server (simplified / detailed).

[0466] Output: A visual representation of the network topology and explanatory information on a display device, along with a set of user-interactive interface controls.

[0467] Step 15: Users can raise questions or request modifications based on the displayed results. User as subject: After reviewing the topology diagram and explanation, users can decide whether to ask further questions or modify the network design.

[0468] Input: The topology diagram and explanatory text displayed on the terminal in step 14.

[0469] Data processing and data calculation: Users enter natural language questions in the terminal's dialog box, such as "Why is redundant link needed?" or "Can it be simpler?"; or enter modification requests, such as "This device should be connected to another switching device"; after the user submits the input, the terminal treats it as a new round of request data.

[0470] Output: New user input is collected by the terminal and sent back to the server, triggering the next round of prompt generation, model inference, and topology update process.

[0471] The specific processing unit 290 sends the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires sound representing user input regarding the result of the specific processing. The control unit 46A sends the sound data representing user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the sound data.

[0472] Data generation model 58 is a so-called generative AI (Artificial Intelligence). Examples of data generation models 58 include ChatGPT (registered trademark) (accessible via the internet (URL: https: / / openai.com / blog / chatgpt)). Data generation model 58 is obtained through deep learning on a neural network. Input to data generation model 58 are prompt words containing instructions, and inference data such as sound data representing sound, text data representing text, and image data representing images (e.g., still image data or animation data). Data generation model 58 infers from the input inference data based on the instructions represented by the prompt words and outputs the inference result in one or more data forms, such as sound data, text data, and image data. Data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or induction. The specific processing unit 290 performs the aforementioned specific processing while using data generation model 58. The data generation model 58 can also be a model finely tuned to output inference results from prompts that do not contain instructions. In this case, the data generation model 58 can output inference results based on prompts that do not contain instructions. The data processing apparatus 12, etc., includes various data generation models 58, including AI other than the generation AI. AI other than the generation AI can be, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and can perform various processes, but is not limited to this example. Furthermore, the AI ​​can also be an AI agent. Furthermore, when the processing of the above-mentioned parts is performed by AI, the processing can be performed partially or entirely by AI, but is not limited to this example. Furthermore, the processing performed by the AI ​​including the generation AI can be replaced by processing in the rule base, and the processing in the rule base can also be replaced by processing performed by the AI ​​including the generation AI.

[0473] Furthermore, the processing of the aforementioned data processing system 10 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart device 14, but it can also be performed by both the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart device 14. Additionally, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the smart device 14 or external devices, and the smart device 14 acquires or collects information required for processing from the data processing device 12 or external devices.

[0474] For example, the collection unit is implemented by the control unit 46A of the smart device 14 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit uses the camera 42 or communication I / F 44 of the smart device 14 to acquire step data, which is then processed by the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12, which analyzes the data from the collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12, which uses a generation AI to generate a menu. For example, the serving unit is implemented by the output device 40 of the smart device 14 or the specific processing unit 290 of the data processing device 12, which provides the generated menu to the user. The correspondence between each unit and the device or control unit is not limited to the above examples and various changes can be made.

[0475] In the above embodiments, examples of specific processing by the data processing device 12 are given, but the technology disclosed herein is not limited to this, and specific processing may also be performed by the smart device 14.

[0476] Second Implementation Method Figure 3 An example of the configuration of the data processing system 210 according to the second embodiment is shown.

[0477] like Figure 3 As shown, the data processing system 210 includes a data processing device 12 and smart glasses 214. A server can be cited as an example of the data processing device 12.

[0478] The data processing apparatus 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" as understood in this disclosure. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, the database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0479] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. Furthermore, the microphone 238, speaker 240, camera 42, and communication I / F 44 are also connected to the bus 52.

[0480] Microphone 238 receives instructions from user 20 by receiving sounds emitted by user 20. Microphone 238 captures sounds emitted by user 20 and converts the captured sounds into sound data, which is then output to processor 46. Speaker 240 outputs sound according to instructions from processor 46.

[0481] Camera 42 is a small digital camera equipped with an optical system such as a lens, aperture and shutter, and imaging elements such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, which captures images of the user 20's surroundings (e.g., the field of view defined by an angle equivalent to the field of vision of an average healthy person).

[0482] Communication I / F44 is connected to network 54. Communication I / F44 and 26 are responsible for the transmission and reception of various information between processor 46 and processor 28 via network 54. The transmission and reception of various information between processor 46 and processor 28 using communication I / F44 and 26 is performed in a secure state.

[0483] Figure 4 This illustrates an example of the main functions of the data processing device 12 and the smart glasses 214. For example... Figure 4 As shown, in the data processing device 12, specific processing is performed by the processor 28. The specific processing program 56 is stored in the memory 32.

[0484] The specific processing program 56 is an example of a "program" involved in the technology of this disclosure. The processor 28 reads the specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. Specific processing is implemented by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0485] The memory 32 stores a data generation model 58 and an emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290. The specific processing unit 290 can use the emotion-specific model 59 to infer the user's emotions and perform specific processing based on the user's emotions. In the emotion inference function (emotion-specific function) using the emotion-specific model 59, various inferences and predictions related to the user's emotions are performed, including inferences and predictions of the user's emotions, but this is not limited to this example. Furthermore, emotion inference and prediction may also include, for example, emotion analysis (parsing).

[0486] In the smart glasses 214, the processor 46 performs reception and output processing. The memory 50 stores the reception and output program 60. The processor 46 reads the reception and output program 60 from the memory 50 and executes the read reception and output program 60 on the RAM 48. The reception and output processing is implemented by the processor 46 operating as a control unit 46A according to the reception and output program 60 executed on the RAM 48. Furthermore, the smart glasses 214 has the same data generation model and emotion-specific model as the data generation model 58 and the emotion-specific model 59, and these models can also be used to perform the same processing as the specific processing unit 290.

[0487] Next, the specific processing of the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is implemented by the data processing device 12 and the smart glasses 214. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0488] Example 1 The process is the same as that of the specific process described in Embodiment 1 in the first embodiment above, so the description is omitted.

[0489] Application Example 1 The process is the same as that in the specific processing described in Application Example 1 of the first embodiment above, so the description is omitted.

[0490] Example 2 The process is the same as that of the specific process in Embodiment 2 described in the first embodiment above, so the description is omitted.

[0491] Application Example 2 The process is the same as that in the specific processing described in Application Example 2 of the first embodiment above, so the description is omitted.

[0492] The specific processing unit 290 sends the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A outputs the result of the specific processing to the speaker 240. The microphone 238 acquires sound input representing the user's input regarding the result of the specific processing. The control unit 46A sends the sound data representing the user's input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the sound data.

[0493] Data generation model 58 is a so-called generative AI (Artificial Intelligence). Examples of data generation models 58 include ChatGPT (registered trademark) (accessible via the internet (URL: https: / / openai.com / blog / chatgpt)). Data generation model 58 is obtained through deep learning on a neural network. Input to data generation model 58 are prompt words containing instructions, and inference data such as sound data representing sound, text data representing text, and image data representing images (e.g., still image data or animation data). Data generation model 58 infers from the input inference data based on the instructions represented by the prompt words and outputs the inference result in one or more data forms, such as sound data, text data, and image data. Data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or induction. The specific processing unit 290 performs the aforementioned specific processing while using data generation model 58. The data generation model 58 can also be a model finely tuned to output inference results from prompts that do not contain instructions. In this case, the data generation model 58 can output inference results based on prompts that do not contain instructions. The data processing apparatus 12, etc., includes various data generation models 58, including AI other than the generation AI. AI other than the generation AI can be, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and can perform various processes, but is not limited to this example. Furthermore, the AI ​​can also be an AI agent. Furthermore, when the processing of the above-mentioned parts is performed by AI, the processing can be performed partially or entirely by AI, but is not limited to this example. Furthermore, the processing performed by the AI ​​including the generation AI can be replaced by processing in the rule base, and the processing in the rule base can also be replaced by processing performed by the AI ​​including the generation AI.

[0494] Furthermore, the processing of the aforementioned data processing system 10 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart glasses 214, but it can also be performed by both the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart glasses 214. Additionally, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the smart glasses 214 or external devices, and the smart glasses 214 acquires or collects information required for processing from the data processing device 12 or external devices.

[0495] For example, the collection unit is implemented by the control unit 46A of the smart glasses 214 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit uses the camera 42 or communication I / F 44 of the smart glasses 214 to acquire step data, which is then processed by the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12, which analyzes the data from the collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12, which uses a generation AI to generate a menu. For example, the serving unit is implemented by the speaker 240 of the smart glasses 214 or the specific processing unit 290 of the data processing device 12, which provides the generated menu to the user. The correspondence between each unit and the device or control unit is not limited to the above examples and various changes can be made.

[0496] In the above embodiments, examples of specific processing by the data processing device 12 are given, but the technology disclosed herein is not limited to this, and specific processing may also be performed by the smart glasses 214.

[0497] Third Implementation Method Figure 5 An example of the configuration of the data processing system 310 according to the third embodiment is shown.

[0498] like Figure 5 As shown, the data processing system 310 includes a data processing device 12 and a head-mounted terminal 314. A server can be cited as an example of the data processing device 12.

[0499] The data processing apparatus 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" as understood in this disclosure. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, the database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0500] The head-mounted terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. Furthermore, the microphone 238, speaker 240, camera 42, display 343, and communication I / F 44 are also connected to the bus 52.

[0501] Microphone 238 receives instructions from user 20 by receiving sounds emitted by user 20. Microphone 238 captures sounds emitted by user 20 and converts the captured sounds into sound data, which is then output to processor 46. Speaker 240 outputs sound according to instructions from processor 46.

[0502] Camera 42 is a small digital camera equipped with an optical system such as a lens, aperture and shutter, and imaging elements such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, which captures images of the user 20's surroundings (e.g., the field of view defined by an angle equivalent to the field of vision of an average healthy person).

[0503] Communication I / F44 is connected to network 54. Communication I / F44 and 26 are responsible for the transmission and reception of various information between processor 46 and processor 28 via network 54. The transmission and reception of various information between processor 46 and processor 28 using communication I / F44 and 26 is performed in a secure state.

[0504] Figure 6 This illustrates an example of the main functions of the data processing device 12 and the head-mounted terminal 314. For example... Figure 6 As shown, in the data processing device 12, specific processing is performed by the processor 28. The specific processing program 56 is stored in the memory 32.

[0505] The specific processing program 56 is an example of a "program" involved in the technology of this disclosure. The processor 28 reads the specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. Specific processing is implemented by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0506] The memory 32 stores the data generation model 58 and the emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290.

[0507] In the head-mounted terminal 314, the processor 46 performs the acceptance / output processing. The memory 50 stores the acceptance / output program 60. The processor 46 reads the acceptance / output program 60 from the memory 50 and executes the read acceptance / output program 60 on the RAM 48. The acceptance / output processing is implemented by the processor 46 operating as a control unit 46A according to the acceptance / output program 60 executed on the RAM 48.

[0508] Next, the specific processing of the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is implemented by the data processing device 12 and the head-mounted terminal 314. In the following description, the data processing device 12 will be referred to as the "server" and the head-mounted terminal 314 will be referred to as the "terminal".

[0509] Example 1 The process is the same as that of the specific process described in Embodiment 1 in the first embodiment above, so the description is omitted.

[0510] Application Example 1 The process is the same as that in the specific processing described in Application Example 1 of the first embodiment above, so the description is omitted.

[0511] Example 2 The process is the same as that of the specific process in Embodiment 2 described in the first embodiment above, so the description is omitted.

[0512] Application Example 2 The process is the same as that in the specific processing described in Application Example 2 of the first embodiment above, so the description is omitted.

[0513] The specific processing unit 290 sends the result of the specific processing to the head-mounted terminal 314. In the head-mounted terminal 314, the control unit 46A outputs the result of the specific processing to the speaker 240 and the display 343. The microphone 238 acquires sound input representing the user's input regarding the result of the specific processing. The control unit 46A sends the sound data representing the user's input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the sound data.

[0514] Data generation model 58 is a so-called generative AI (Artificial Intelligence). Examples of data generation models 58 include ChatGPT (registered trademark) (accessible via the internet (URL: https: / / openai.com / blog / chatgpt)). Data generation model 58 is obtained through deep learning on a neural network. Input to data generation model 58 includes prompt words containing instructions, and inference data such as sound data representing sound, text data representing text, and image data representing images (e.g., still image data or animation data). Data generation model 58 infers the input inference data based on the instructions represented by the prompt words and outputs the inference result in one or more data forms such as sound data, text data, and image data. Data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or induction. The specific processing unit 290 performs the aforementioned specific processing while using data generation model 58. The data generation model 58 can also be a model finely tuned to output inference results from prompts that do not contain instructions. In this case, the data generation model 58 can output inference results based on prompts that do not contain instructions. The data processing apparatus 12, etc., includes various data generation models 58, including AI other than the generation AI. AI other than the generation AI can be, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and can perform various processes, but is not limited to this example. Furthermore, the AI ​​can also be an AI agent. Furthermore, when the processing of the above-mentioned parts is performed by AI, the processing can be performed partially or entirely by AI, but is not limited to this example. Furthermore, the processing performed by the AI ​​including the generation AI can be replaced by processing in the rule base, and the processing in the rule base can also be replaced by processing performed by the AI ​​including the generation AI.

[0515] Furthermore, the processing of the aforementioned data processing system 10 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the head-mounted terminal 314, but it can also be performed by both the specific processing unit 290 of the data processing device 12 and the control unit 46A of the head-mounted terminal 314. Additionally, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the head-mounted terminal 314 or external devices, and the head-mounted terminal 314 acquires or collects information required for processing from the data processing device 12 or external devices.

[0516] For example, the collection unit is implemented by the control unit 46A of the head-mounted terminal 314 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit uses the camera 42 or communication I / F 44 of the head-mounted terminal 314 to acquire step data, which is then processed by the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12 to analyze the data from the collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12 to generate a menu using a generation AI. For example, the serving unit is implemented by the speaker 240 and display 343 of the head-mounted terminal 314 or the specific processing unit 290 of the data processing device 12 to provide the generated menu to the user. The correspondence between each unit and the device or control unit is not limited to the above examples and various changes can be made.

[0517] In the above embodiments, examples of specific processing by the data processing device 12 are given, but the technology disclosed herein is not limited to this, and specific processing may also be performed by the head-mounted terminal 314.

[0518] Fourth Implementation Method Figure 7 An example of the configuration of the data processing system 410 according to the fourth embodiment is shown.

[0519] like Figure 7 As shown, the data processing system 410 includes a data processing device 12 and a robot 414. A server can be cited as an example of the data processing device 12.

[0520] The data processing apparatus 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" as understood in this disclosure. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, the database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0521] Robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. Furthermore, the microphone 238, speaker 240, camera 42, controlled object 443, and communication I / F 44 are also connected to the bus 52.

[0522] Microphone 238 receives instructions from user 20 by receiving sounds emitted by user 20. Microphone 238 captures sounds emitted by user 20 and converts the captured sounds into sound data, which is then output to processor 46. Speaker 240 outputs sound according to instructions from processor 46.

[0523] Camera 42 is a small digital camera equipped with an optical system such as a lens, aperture and shutter, and imaging elements such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, to photograph the area around robot 414 (e.g., the field of view defined by a perspective equivalent to the field of vision of an average healthy person).

[0524] Communication I / F44 is connected to network 54. Communication I / F44 and 26 are responsible for the transmission and reception of various information between processor 46 and processor 28 via network 54. The transmission and reception of various information between processor 46 and processor 28 using communication I / F44 and 26 is performed in a secure state.

[0525] The controlled object 443 includes a display device, LEDs (light-emitting diodes) for the eyes, and motors for driving the arms, hands, and feet. The posture or movement of the robot 414 is controlled by controlling the motors in the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. In addition, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0526] Figure 8 This illustrates an example of the main functions of the data processing device 12 and the robot 414. For example... Figure 8 As shown, in the data processing device 12, specific processing is performed by the processor 28. The specific processing program 56 is stored in the memory 32.

[0527] The specific processing program 56 is an example of a "program" involved in the technology of this disclosure. The processor 28 reads the specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. Specific processing is implemented by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0528] The memory 32 stores the data generation model 58 and the emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290.

[0529] In robot 414, the processor 46 performs the acceptance and output processing. The memory 50 stores the acceptance and output program 60. The processor 46 reads the acceptance and output program 60 from the memory 50 and executes the read acceptance and output program 60 on RAM 48. The acceptance and output processing is implemented by the processor 46 acting as the control unit 46A according to the acceptance and output program 60 executed on RAM 48.

[0530] Next, the specific processing of the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is implemented by the data processing device 12 and the robot 414. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 will be referred to as the "terminal".

[0531] Example 1 The process is the same as that of the specific process described in Embodiment 1 in the first embodiment above, so the description is omitted.

[0532] Application Example 1 The process is the same as that in the specific processing described in Application Example 1 of the first embodiment above, so the description is omitted.

[0533] Example 2 The process is the same as that of the specific process in Embodiment 2 described in the first embodiment above, so the description is omitted.

[0534] Application Example 2 The process is the same as that in the specific processing described in Application Example 2 of the first embodiment above, so the description is omitted.

[0535] The specific processing unit 290 sends the result of the specific processing to the robot 414. In the robot 414, the control unit 46A outputs the result of the specific processing to the speaker 240 and the controlled object 443. The microphone 238 acquires sound input representing the result of the specific processing. The control unit 46A sends the sound data representing the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the sound data.

[0536] Data generation model 58 is a so-called generative AI (Artificial Intelligence). Examples of data generation models 58 include ChatGPT (registered trademark) (accessible via the internet (URL: https: / / openai.com / blog / chatgpt)). Data generation model 58 is obtained through deep learning on a neural network. Input to data generation model 58 are prompt words containing instructions, and inference data such as sound data representing sound, text data representing text, and image data representing images (e.g., still image data or animation data). Data generation model 58 infers from the input inference data based on the instructions represented by the prompt words and outputs the inference result in one or more data forms, such as sound data, text data, and image data. Data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or induction. The specific processing unit 290 performs the aforementioned specific processing while using data generation model 58. The data generation model 58 can also be a model finely tuned to output inference results from prompts that do not contain instructions. In this case, the data generation model 58 can output inference results based on prompts that do not contain instructions. The data processing apparatus 12, etc., includes various data generation models 58, including AI other than the generation AI. AI other than the generation AI can be, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and can perform various processes, but is not limited to this example. Furthermore, the AI ​​can also be an AI agent. Furthermore, when the processing of the above-mentioned parts is performed by AI, the processing can be performed partially or entirely by AI, but is not limited to this example. Furthermore, the processing performed by the AI ​​including the generation AI can be replaced by processing in the rule base, and the processing in the rule base can also be replaced by processing performed by the AI ​​including the generation AI.

[0537] Furthermore, the processing of the aforementioned data processing system 10 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the robot 414, but it can also be performed by both the specific processing unit 290 of the data processing device 12 and the control unit 46A of the robot 414. Additionally, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the robot 414 or external devices, and the robot 414 acquires or collects information required for processing from the data processing device 12 or external devices.

[0538] For example, the collection unit is implemented by the control unit 46A of the robot 414 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit uses the camera 42 or communication I / F 44 of the robot 414 to acquire step data, which is then processed by the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12, which analyzes the data from the collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12, which uses a generation AI to generate a menu. For example, the serving unit is implemented by the speaker 240 of the robot 414 and the control object 443 or the specific processing unit 290 of the data processing device 12, which provides the generated menu to the user. The correspondence between each unit and the device or control unit is not limited to the above examples and various changes can be made.

[0539] In the above embodiments, examples of specific processing by the data processing device 12 are given, but the technology disclosed herein is not limited to this, and specific processing may also be performed by the robot 414.

[0540] Furthermore, the emotion-specific model 59, acting as an emotion engine, can determine a user's emotion based on a specific mapping. Specifically, the emotion-specific model 59 can determine a user's emotion based on an emotion graph that serves as a specific mapping (see [reference]). Figure 9 The emotion-specific model 59 can also determine the robot's emotion, and the specific processing unit 290 performs specific processing based on the robot's emotions.

[0541] Figure 9 This is a diagram representing an emotion map 400 that maps multiple emotions. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotion is. On the outer side of the concentric circles, emotions representing states or behaviors arising from mood are arranged. Emotions are concepts that include feelings and mental states. Emotions generated by reactions occurring in the brain are arranged roughly to the left of the concentric circles. Emotions derived from situational judgments are arranged roughly to the right of the concentric circles. Emotions generated by reactions occurring in the brain and derived from situational judgments are arranged roughly above and below the concentric circles. Furthermore, "pleasant" emotions are arranged above the concentric circles, and "unpleasant" emotions are arranged below them. Thus, in the emotion map 400, multiple emotions are mapped based on the structure that generates emotions, and emotions that are likely to occur simultaneously are mapped close to each other.

[0542] These emotions are distributed at the three o'clock position of the emotion map 400, typically fluctuating between peace and anxiety. In the right half of the emotion map 400, situational awareness dominates over internal sensation, thus resulting in an impression of calm.

[0543] The inner side of the emotion map 400 represents the inner state, while the outer side represents behavior. Therefore, the further outward you are from the emotion map 400, the more visible the emotion becomes (manifested in behavior).

[0544] Here, human emotions are based on various balances such as posture and blood sugar levels. When these balances deviate from an ideal state, it indicates an unpleasant state; when they approach the ideal state, it indicates a pleasant state. Emotions in robots, cars, motorcycles, etc., can also be created in the following way: based on various balances such as posture and remaining battery power, when these balances deviate from an ideal state, it indicates an unpleasant state; when they approach the ideal state, it indicates a pleasant state. Emotion maps can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a Brain Physiological Signal Analysis System for Voice Emotion Recognition and Emotion, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). In the left half of the emotion map, emotions belonging to the sensory-dominated region, called "response," are arranged. Furthermore, in the right half of the emotion map, emotions belonging to the situational cognition-dominated region, called "situation," are arranged.

[0545] In the emotion map, two types of emotions that promote learning are defined. One is a negative emotion on the situational side, in the middle or peripheral region of "repentance" or "reflection." This occurs when the robot experiences negative emotions such as "I don't want to experience this feeling again" or "I don't want to be blamed again." The other is a positive emotion on the response side, near the "desire" region. This occurs when there are positive feelings such as "wanting more" or "wanting to know more."

[0546] The emotion-specific model 59 inputs user input into a pre-trained neural network to obtain emotion values ​​representing each emotion shown in the emotion map 400, thereby determining the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network... Figure 10 As shown in the sentiment graph 900, it was trained in a way that sentiments that are configured close to each other have similar values. Figure 10 The text shows examples of emotions such as "peace of mind", "stability", and "reassurance" that have similar emotion values.

[0547] The above description focuses on the functions of the data processing device 12, but the system of this disclosure is not necessarily installed on a server. The system of this disclosure can also be installed as a general information processing system. This disclosure can also be installed, for example, as a software program running on a personal computer, an application running on a smartphone, etc. The method of this disclosure can also be provided to users in the form of SaaS (Software as a Service).

[0548] In the above embodiments, an example of a specific process being performed by a single computer 22 is given. However, the technology disclosed herein is not limited to this, and the specific process can also be distributed among multiple computers, including computer 22. For example, the data generation model 58 can be located on an external device of the data processing apparatus 12, where data is generated based on the input data.

[0549] In the above embodiments, examples of storing a specific processing program 56 in the memory 32 have been described, but the technology disclosed herein is not limited thereto. For example, the specific processing program 56 may also be stored in a portable computer-readable non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed into the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0550] Alternatively, a specific processing program 56 may be pre-stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 according to the requirements of the data processing device 12.

[0551] In addition, it is not necessary to store all the specific processing program 56 in the storage device such as the server connected to the data processing device 12 via the network 54 or in the memory 32; a portion of the specific processing program 56 may be stored in advance.

[0552] As hardware resources for performing specific processes, various processors, as shown below, can be used. For example, a CPU can be listed as a processor, which functions as a general-purpose processor that performs specific processes by executing software, i.e., a program. Furthermore, processors can be listed as special-purpose circuits such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application-Specific Integrated Circuits), which are processors with circuitry specifically designed to perform specific processes. Each processor has built-in or connected memory, and each processor executes specific processes using that memory.

[0553] The hardware resources for performing a specific process can consist of one of these various processors, or a combination of two or more processors of the same or different types (e.g., a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resources for performing a specific process can be a single processor.

[0554] As an example of a single processor, there are two approaches: First, a processor is composed of a combination of one or more CPUs and software, which functions as a hardware resource to perform a specific process; second, as represented by a SoC (System-on-a-chip), a processor is used to implement the functionality of the entire system, which includes multiple hardware resources for performing a specific process, using a single IC (Integrated Circuit) chip. In this way, the specific process is implemented by using one or more of the aforementioned processors as hardware resources.

[0555] Furthermore, the hardware architecture of these various processors, more specifically, can utilize circuits that combine semiconductor elements and other circuit components. Moreover, the specific process described above is just one example. Therefore, without departing from the main point, unnecessary steps can certainly be deleted, new steps added, or the processing order changed.

[0556] The descriptions and illustrations above are detailed explanations of a portion of the technology disclosed herein, and are merely one example of the technology disclosed herein. For example, the above descriptions of the structure, function, effect, and results are just one example of the structure, function, effect, and results of a portion of the technology disclosed herein. Therefore, without departing from the spirit of the technology disclosed herein, unnecessary parts may be deleted, new elements added, or replacements may be made to the descriptions and illustrations above. Furthermore, to avoid confusion and facilitate understanding of a portion of the technology disclosed herein, explanations of common technical knowledge that do not require special explanation under the premise of being able to implement the technology disclosed herein have been omitted from the descriptions and illustrations above.

[0557] All documents, patent applications and technical specifications set forth in this specification are incorporated herein by reference to the same extent that each document, patent application and technical specification is specifically and individually described therein and referenced by reference.

[0558] In addition, the following notes are provided in response to the above explanation.

[0559] Example 1 (Note 1) An information processing system, characterized in that it comprises: A device for obtaining information related to communication resources and processing conditions from a user via a communication terminal; A device for structuring and recording the information related to communication resources and the information related to processing conditions into an information storage device, and for generating prompt statements that express instructions for generative artificial intelligence models based on the recorded information. A device for inputting the generated prompt statement into the generative artificial intelligence model and obtaining, through the generative artificial intelligence model, structural information representing the connection and arrangement relationships of the communication resources; An apparatus for converting the acquired compositional information into structured data for visualization and outputting the structured data to the communication terminal; An apparatus for updating the recorded information and the prompt statement based on additional or corrected input from the user, and for repeatedly performing the process of generating the constituent information through the generative artificial intelligence model.

[0560] (Note 2) The information processing system according to Appendix 1 is characterized in that, The device for generating the prompt statement is configured to classify the information related to communication resources recorded in the information storage device, and change the description content and format of the prompt statement according to the classification result, thereby dynamically adjusting the instruction content sent to the generative artificial intelligence model.

[0561] (Note 3) The information processing system according to Appendix 1 is characterized in that, The apparatus for converting the constituent information into structured data for visualization is configured to extract elements representing communication devices and elements representing connection relationships between the communication devices from the output information obtained from the generative artificial intelligence model, and associate the elements representing the communication devices as node information and the elements representing the connection relationships as connection information into the structured data.

[0562] Application Example 1 (Note 1) An information processing system, characterized in that it comprises: This is a means by which a processing device stores information containing industrial equipment status information and communication link operation status as communication environment information into an information storage device. A means for generating prompt statements as input to a generative artificial intelligence model by a processing device based on stored communication environment information and device identification information and connection request information obtained from the terminal. A means for inputting generated prompt statements into the generative artificial intelligence model by a processing device to calculate communication structure information including the connection relationship between communication devices and information processing devices, as well as communication path candidates; A means for generating a structural graph data with communication devices and industrial equipment as nodes and physical or logical connection relationships as edges, based on the communication structural information obtained by the processing device, and determining the optimal communication path among the multiple communication paths contained in the structural graph data according to the latency and load-related evaluation indicators. A means for sending a specific optimal communication path and corresponding connection interface setting information from a processing device to the terminal, and for presenting the network structure diagram on a display device via the terminal. Means for a processing device to generate communication device setting data corresponding to a specific optimal communication path based on confirmation information from the terminal, and to send the setting data to the communication device.

[0563] (Note 2) The information processing system according to Appendix 1 is characterized in that, The processing device is also used to generate design guidance information related to the communication address system and path control method based on the communication environment information stored in the information storage device and the output results of the generative artificial intelligence model, and to provide the design guidance information to the terminal.

[0564] (Note 3) The information processing system according to Appendix 1 is characterized in that, The processing device is also used to update the communication environment information based on monitoring data obtained from the communication equipment and industrial equipment, and to dynamically adjust the content of the prompt statement and the generation parameters used to call the generative artificial intelligence model, so that the updated communication environment information is reflected in the prompt statement input to the generative artificial intelligence model.

[0565] Example 2 (Note 1) An information processing system, characterized in that it comprises: It operates as an information processing device and is used to store the imported status of the communication network as constitutive information into a relational data management device. A device for obtaining the composition information from the relational data management device and generating, based on the composition information, prompt statements as natural language data to indicate the components required for generating a communication network composition diagram. A device for inputting the prompt statement into a generative artificial intelligence model and obtaining constituent data representing the compositional graph of the communication network based on natural language processing and data generation processing performed by the generative artificial intelligence model; A device for converting the constituent data into display control data and generating output information for visually displaying the communication network constituent diagram on a terminal device.

[0566] (Note 2) The information processing system according to Appendix 1 is characterized in that, The information processing device is configured to generate design information on communication network address allocation and path control as supplementary information based on the constituent data obtained from the generative artificial intelligence model, and to record the prompt statement and the supplementary information as historical information in the relational data management device.

[0567] (Note 3) The information processing system according to Appendix 1 is characterized in that, The information processing device is configured to obtain past composition information and historical information from the relational data management device, and dynamically adjust the recorded content and structure of the prompt statement according to the import status of the communication network, user attributes and past design results, so as to optimize the content input into the generative artificial intelligence model.

[0568] Application Example 2 (Note 1) An information processing system, characterized in that it comprises: A device for receiving network device information and communication requirement information from an external terminal in an information processing apparatus and acquiring the information as structured data; An apparatus for storing the structured data in an information management area within a storage unit and for retrieving the structured data from the information management area for aggregation processing, thereby extracting communication topology requirements; A device for automatically generating prompt statements for input into a generative artificial intelligence model based on the communication topology requirements and user attribute information; A device for sending a network topology generation request to the generative artificial intelligence model using the prompt statement, and for parsing the natural language output obtained from the generative artificial intelligence model to convert it into network topology data containing connection object elements and connection relationships. A device for generating drawing data of a network topology map by calling a visualization processing program based on the network topology data, and for maintaining the drawing data in a format that can be output to the external terminal; An apparatus for inferring a user's emotional state and level of understanding based on dialogue input and emotion-related information obtained from the user, and for dynamically adjusting the level of detail and content of the prompt statements, the network topology diagram, and the displayed content based on the inference results. A device for sending the network topology diagram and explanatory information generated by the generative artificial intelligence model to the external terminal, and presenting the information to the user through the dialog interface of the external terminal.

[0569] (Note 2) According to the information processing system described in Appendix 1, the information processing device is configured to: add requirements related to address system design and routing control strategy to the prompt statements for the generative artificial intelligence model, thereby obtaining explanatory information about the address allocation structure and routing control structure from the generative artificial intelligence model, and providing the explanatory information to the external terminal.

[0570] (Note 3) According to the information processing system described in Appendix 1, the information processing device is configured to: adjust the content and constraints of the prompt statement based on the network import status and operation history stored in the information management area, and switch the display elements of the network topology map and the professional level of the explanatory information according to the inference result of the emotional state.

Claims

1. An information processing system, characterized in that, include: processor; The processor is configured as follows: Save the network import status to the database; Based on the saved data, a generative artificial intelligence model is used to generate prompt text indicating the components required to generate the network structure diagram; The prompt text is input into the generative artificial intelligence model to generate the optimal network structure graph.

2. The information processing system according to claim 1, characterized in that, The processor is configured to provide information related to network address design and routing control.

3. The information processing system according to claim 1, characterized in that, The processor is configured to adjust the prompt text based on the network import status.

Images