Information processing methods, programs, and information processing systems

Abstract

This invention provides an information processing method, program, and information processing system that integrally generate application code and infrastructure configurations based on requests in natural language. [Solution] The method involves receiving input in natural language S101, extracting application requirements and infrastructure requirements S102, generating code to generate application code S103, determining the necessary infrastructure configuration S104, adjusting the application code S105, performing security verification on the adjusted application code and infrastructure configuration S106, verifying the operation of the application code and infrastructure configuration in a simulated environment S107, determining whether correction is necessary S108, and if correction is determined to be necessary, correcting the application code or infrastructure configuration S109.

Description

Technical Field

[0001] The present invention relates to an information processing method, a program, and an information processing system.

Background Art

[0002] Patent Document 1 discloses a task management system that performs automatic design and execution of tasks through user interaction in natural language.

Prior Art Document

Patent Document

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the case of the above prior art, the use of generative AI and interactive applications enables the automation of business task design and execution in natural language, but the target is limited to a series of task processing. In actual application development and system infrastructure construction, it is necessary to integrally process multiple steps.

[0005] In consideration of the above facts, an object of the present invention is to obtain an information processing method, a program, and an information processing system that can integrally generate application code and infrastructure configuration based on requirements in natural language.

Means for Solving the Problems

[0006] According to an information processing method according to one embodiment, the method includes: a step of receiving input in natural language; a requirement extraction step of extracting application requirements and infrastructure requirements based on the content of the natural language input; a code generation step of generating application code based on the application requirements; an infrastructure configuration determination step of determining the necessary infrastructure configuration based on the infrastructure requirements; a code adjustment step of adjusting the application code based on the constraints of the infrastructure configuration; a security verification step of performing security verification on the adjusted application code and the infrastructure configuration; an operation verification step of verifying the operation of the application code and the infrastructure configuration in a simulated environment; a determination step of determining whether or not modifications are necessary based on the results of the security verification step and the operation verification step; and a modification step of modifying the application code or the infrastructure configuration if it is determined in the determination step that modifications are necessary.

[0007] According to a program of one embodiment, the information processing device is made to execute an information processing method that includes: a step of receiving input in natural language; a requirement extraction step of extracting application requirements and infrastructure requirements based on the content of the natural language input; a code generation step of generating application code based on the application requirements; an infrastructure configuration determination step of determining the necessary infrastructure configuration based on the infrastructure requirements; a code adjustment step of adjusting the application code based on the constraints of the infrastructure configuration; a security verification step of performing security verification on the adjusted application code and the infrastructure configuration; an operation verification step of verifying the operation of the application code and the infrastructure configuration in a simulated environment; a determination step of determining whether or not modifications are necessary based on the results of the security verification step and the operation verification step; and a modification step of modifying the application code or the infrastructure configuration if it is determined in the determination step that modifications are necessary.

[0008] According to an information processing system according to one embodiment, the information processing system includes an information processing device and performs the following steps: receiving input in natural language; extracting application requirements and infrastructure requirements based on the content of the natural language input; generating code based on the application requirements; determining the necessary infrastructure configuration based on the infrastructure requirements; adjusting the application code based on the constraints of the infrastructure configuration; performing security verification on the adjusted application code and the infrastructure configuration; verifying the operation of the application code and the infrastructure configuration in a simulated environment; determining whether or not modifications are necessary based on the results of the security verification step and the operation verification step; and modifying the application code or the infrastructure configuration if it is determined in the determination step that modifications are necessary. [Effects of the Invention]

[0009] According to one embodiment, application code and infrastructure configuration can be generated integrally based on requests in natural language. [Brief explanation of the drawing]

[0010] [Figure 1] This figure shows an example of the configuration of an information processing system according to the embodiment. [Figure 2] This figure shows an example of the server hardware configuration according to the embodiment. [Figure 3] This figure shows an example of the hardware configuration of a user terminal according to the embodiment. [Figure 4] This figure shows an example of the functional configuration of the server according to the embodiment. [Figure 5] This figure shows an example of the functional configuration of a user terminal according to the embodiment. [Figure 6] This flowchart shows an example of a process performed by the information processing system according to the embodiment. [Figure 7] This is a schematic diagram illustrating the implementation of each function of the information processing system according to this embodiment. [Modes for carrying out the invention]

[0011] An embodiment of the information processing system according to the present invention will be described below with reference to Figures 1 to 5. In each figure, identical or equivalent components and parts are given the same reference numerals. Furthermore, the dimensional ratios in the drawings are exaggerated for illustrative purposes and may differ from actual ratios.

[0012] First, an overview of the information processing system 10 according to this embodiment will be described. The information processing system 10 according to this embodiment is an integrated development system that automatically generates applications and infrastructure from natural language requests and builds a secure production environment. Users can develop applications and build infrastructure without requiring specialized knowledge simply by inputting ideas or business requirements in natural language from the user terminal 2. This makes it possible to autonomously perform a series of processes such as design, development, infrastructure construction, security checks, and testing, which were conventionally performed individually and manually, in a continuous loop within the system. In this embodiment, each functional unit that utilizes AI technology may be a software module that autonomously learns, judges, and processes in response to user instructions or environmental changes, using artificial intelligence technologies such as large-scale language models (LLMs) and reinforcement learning algorithms.

[0013] According to this embodiment, the process of building a production environment, which previously required specialized infrastructure engineers, can be automated using only natural language requests. In other words, by not separating application code generation and infrastructure construction, but processing them integrally while considering their interdependencies, the technical gap in the transition from prototype to production environment can be eliminated, and a practical system that meets security requirements can be built in a short time. In particular, by automatically and repeatedly adjusting the code due to infrastructure configuration constraints and re-determining the infrastructure configuration based on code requests, optimization that would be difficult to achieve manually can be realized, enabling cost reduction and a significant shortening of development time.

[0014] In this embodiment, the requirements extraction unit classifies and extracts functional, non-functional, security, and performance requirements from natural language requests entered by the user. The extracted requirements are passed in parallel to the code generation unit and the infrastructure configuration determination unit, each generating initial deliverables. Subsequently, the code is modified based on infrastructure constraints (e.g., session management when using a load balancer, database connection limits, etc.), and additional resources are added to the configuration as needed. This process is automatically repeated until consistency between the two is confirmed. Furthermore, comprehensive verification is performed based on the company's security checklist and industry standards, and actual operation is confirmed using a simulated environment. If areas requiring correction are found during verification, the correction unit automatically performs corrections by referring to past correction history and learned patterns. In addition, even after deployment to the production environment, user feedback is accepted in natural language, and the system is continuously improved through reinforcement learning, including DPO (Direct Preference Optimization). In this way, this embodiment provides an innovative platform that enables anyone to easily build a production-ready system through the integrated generation of applications and infrastructure, even if they lack specialized knowledge.

[0015] (System Configuration) Figure 1 shows an example configuration of the information processing system 10 according to this embodiment. As shown in Figure 1, the information processing system 10 according to this embodiment is a system that integrally generates and manages infrastructure and applications based on natural language instructions, and comprises a server device 1 connected to each other via a network N for communication, and a user terminal 2 operated by the user. The network N may consist of a wired LAN, a wireless LAN, the Internet, a public telephone network, a mobile data communication network, or a combination thereof. The server device 1 in this embodiment is an information processing device that has the function of performing various processes such as extracting application requirements and infrastructure requirements using natural language, generating and adjusting application code and infrastructure configuration, integrated security verification, operational verification in a simulated environment, and automatic correction and optimization based on the results thereof. The server device 1 is also connected via the network N to a simulated environment (not shown) for verifying the operation of applications, a production infrastructure environment (cloud service or on-premise environment, etc., not shown) where the final product is deployed, and various external API services. The user terminal 2 is equipped with an input / output unit for the user to input generation requests to the server device 1 using natural language, and to check and operate the progress of the generation process, management, notification information, etc. In the example shown in Figure 1, there is one server device 1 and one user terminal 2, but multiple devices may be provided.

[0016] Server device 1 is an example of an information processing device capable of performing the main processes of the present invention, which integrate and manage infrastructure and applications based on natural language instructions from the user and continuously improve them. Server device 1 is equipped with software and hardware to control the operation of each step of the information processing method (requirements extraction, code generation, infrastructure configuration determination, code adjustment, security verification, operation verification, judgment, correction, learning, etc.). Server device 1 has the functionality to automatically and continuously perform the extraction of application and infrastructure requirements using natural language, application code generation, infrastructure configuration determination, code adjustment based on infrastructure constraints, integrated security verification, operation verification in a simulated environment, automatic correction and optimization using reinforcement learning, and continuous learning based on user feedback, etc. Server device 1 may be a PC (Personal Computer), smartphone, tablet terminal, dedicated server machine, cloud server, microcomputer, or a combination thereof. Furthermore, the configuration of server device 1 can include PCs, dedicated server machines, virtual machines in a cloud environment, microcomputers, etc., and large-scale cluster configurations or distributed deployments are also permitted.

[0017] The user terminal 2 is an example of an information processing device for a user to access the server device 1 and perform operations such as inputting a generation request in natural language on the management screen, viewing the progress of the generation process, viewing notification information, using the deployed product, and inputting improvement feedback. The user terminal 2 is equipped with a display unit such as a display, an input unit such as a mouse, keyboard, touch panel, etc., and a network communication function. The user terminal 2 may be a PC, smartphone, tablet terminal, wearable terminal, or any combination thereof. In this embodiment, even a user without specialized knowledge of infrastructure can build a secure system by simply inputting requests such as "want to create an EC site" or "want an in-house attendance management system" in natural language. Continuous improvement is possible through natural language feedback even after construction. Also, in this embodiment, not only can the user grasp the state of the product generated on the management screen and check the content of the notification from the server device 1, but also the user can smoothly perform application development and operation management tasks such as accessing detailed information of the corresponding product, accessing recommended actions, and inputting feedback for continuous improvement.

[0018] (Hardware Configuration - Server Device) Figure 2 is a block diagram showing the hardware configuration of the server device 1. The server device 1 includes a processor 101, a memory 102, a storage 103, and a communication I / F 104, which are communicably connected to each other via a bus B.

[0019] The processor 101 controls each component of the server device 1 and realizes the functions of the server device 1 by expanding and executing various programs stored in the storage 103 in the memory 102. The programs executed by the processor 101 include, but are not limited to, the OS (Operating System) and various programs described later. By executing these programs by the processor 101, a part of the information processing method according to the present embodiment is realized. The processor 101 is, for example, a CPU (Central Processing Unit), MPU (Micro Processing Unit), GPU (Graphics Processing Unit), ASIC (Application Specific Integrated Circuit), DSP (Digital Signal Processor), or a combination thereof.

[0020] The memory 102 is, for example, a ROM (Read Only Memory), RAM (Random Access Memory), or a combination thereof. The ROM is, for example, a PROM (Programmable ROM), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), or a combination thereof. The RAM is, for example, a DRAM (Dynamic RAM), SRAM (Static RAM), MRAM (Magnetoresistive RAM), or a combination thereof.

[0021] The storage 103 stores the OS, various programs described later, and various data. The storage 103 is, for example, a flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), SCM (Storage Class Memories), or a combination thereof.

[0022] The communication interface 104 is an interface for connecting server device 1 to external devices, including user terminals 2 and 3, via network N, and for controlling communication. The communication interface 104 is, for example, an adapter compliant with Bluetooth®, Wi-Fi®, ZigBee®, Ethernet®, or optical communication (e.g., Fibre Channel).

[0023] The input device 105 is a device for inputting information into the server device 1. The input device 105 may be, but is not limited to, a mouse, keyboard, touch panel, microphone, scanner, camera, various sensors, or operation buttons.

[0024] The output device 106 is a device for outputting information from the server device 1. The output device 106 is, for example, a display device (display or monitor), a projector, a printer, a speaker, or a vibrator, but is not limited to these.

[0025] (Hardware configuration - User terminal) Figure 3 is a block diagram showing the hardware configuration of user terminal 2. User terminal 2 comprises a processor 201, memory 202, storage 203, communication I / F 204, input / output I / F 205, input device 206, and output device 207, all of which are connected to each other via bus B for communication. Communication I / F 204 is an interface for connecting user terminal 2 to an external device, including server device 1, via network N, and for controlling communication. Communication I / F 204 is, for example, an adapter compliant with Bluetooth®, Wi-Fi®, ZigBee®, Ethernet®, or optical communication (e.g., Fibre Channel).

[0026] The input / output interface 205 is an interface for connecting an input device 206 and an output device 207 to the user terminal 2. The input device 206 is, for example, a mouse, keyboard, touch panel, microphone, scanner, camera, various sensors, operation buttons, or a combination thereof. The output device 207, as a user interface, is, for example, a display, projector, printer, speaker, vibrator, or a combination thereof. In this embodiment, as an example, the output device 207 and the input device 206 are configured as an integrated touch panel display.

[0027] In this embodiment, the program may be written to memory 202 or storage 203 during the manufacturing stage of the server device 1, or it may be provided to the server device 1 via the network N. It may also be provided to the server device 1 via a non-temporary, computer-readable recording medium such as a disk medium (not shown).

[0028] (Functional Configuration - Server Device) Next, the functional configuration of the server device 1 will be described. Figure 4 is a diagram showing an example of the functional configuration of the server device 1 according to this embodiment. The server device 1 integrates the design, construction, verification, and optimization processes of applications and infrastructure based on natural language input using the hardware resources described above. In this embodiment, the server device 1 comprises a communication unit 11, a storage unit 12, and a control unit 13 as its functional configuration. Each functional unit is realized by the processor executing a program stored in memory or storage. The storage unit 12 includes requirement information 121 that stores information on various requirements extracted from the input natural language, code information 122 that stores generated code information, infrastructure configuration information 123 that stores infrastructure configuration information, security policy 124 that stores policy information that serves as a security standard, and revision history 125 that stores past revision history.

[0029] The control unit 13 includes a reception unit 131, a requirements extraction unit 132, a code generation unit 133, an infrastructure configuration determination unit 134, a code adjustment unit 135, a security verification unit 136, an operation verification unit 137, a determination unit 138, a correction unit 139, and a learning unit 140.

[0030] The reception unit 131 receives an application generation request in natural language from the user terminal 2 via the communication unit 11. The requirements extraction unit 132 extracts application requirements and infrastructure requirements based on the request received by the reception unit 131, while referring to the requirements information 121. The code generation unit 133 generates application code based on the extracted application requirements and stores it in the code information 122.

[0031] The infrastructure configuration determination unit 134 analyzes the application code in the code information 122 and determines the necessary infrastructure configuration while referring to the infrastructure configuration information 123. The code adjustment unit 135 adjusts the application code in the code information 122 based on the constraints of the infrastructure configuration information 123 and stores it in the code information 122. The security verification unit 136 refers to the security policy 124 on the code information 122 and the infrastructure configuration information 123 and performs integrated security verification.

[0032] The operation verification unit 137 works in conjunction with a simulated environment to verify the operation of the code information 122 and the infrastructure configuration information 123, and stores the results in the modification history 125. The determination unit 138 determines whether or not a modification is necessary based on the results of the security verification unit 136 and the operation verification unit 137. If the determination unit 138 determines that a modification is necessary, the modification unit 139 refers to the modification history 125 and modifies the code information 122 or the infrastructure configuration information 123. The learning unit 140 improves processes such as code generation or configuration determination through reinforcement learning based on feedback from each functional unit and feedback from the user. The server device 1 also has a function to notify the user of information that each functional unit has detected, such as abnormalities or changes, that it has decided should be notified, on the management screen displayed on the user terminal 2 via the communication unit 11.

[0033] In this embodiment, each functional unit within the control unit 13 may be implemented as an independent software module. These functional units may operate individually or cooperatively using specialized algorithms or decision logic, and may execute processing in parallel and asynchronously without waiting for each other to complete. In particular, during the code generation, verification, and modification cycle, each functional unit responsible for intelligent processing such as a large-scale language model (LLM) or reinforcement learning model can continuously monitor and analyze large amounts of information. Furthermore, these functional units share information with each other and work closely together; for example, the configuration constraints determined by the infrastructure configuration determination unit 134 are fed back to the code generation unit 133, and the results of the security verification unit 136 influence the modification strategy of the modification unit 139. The notification function via the communication unit 11 integrates the verification results and learning results obtained from each functional unit and autonomously determines the need for notification, providing only the necessary information at the appropriate time without user intervention.

[0034] In this embodiment, each functional unit within the control unit 13 is a group of software modules that handle intelligent processing such as natural language processing, inference, and reinforcement learning, and autonomously learn, judge, and process in response to user instructions, system events, and environmental changes. By having these functional modules work in coordination, delays in response due to information overload can be eliminated, and bottlenecks in development and operation can be removed. Furthermore, the results of continuous verification of the generated application code and infrastructure configuration, as well as the results of responses to user notification information (immediate response, later response, ignore, etc.), are collected as feedback by the learning unit 140 and used to retrain the reinforcement learning model. As a result, the accuracy of code generation, infrastructure configuration determination, and correction is continuously improved, and a safe and optimal product is provided to the user.

[0035] (Functional Configuration - User Terminal) Next, the functional configuration of the user terminal 2 will be described. Figure 5 is a diagram showing an example of the functional configuration of the user terminal 2. When executing various programs, the user terminal 2 uses the above hardware resources to realize various functions. The user terminal 2 has a communication unit 21, a storage unit 22 in which the program 221 is stored, and a control unit 23 as the functional configuration realized by the user terminal 2. Each functional configuration is realized by the processor 201 reading and executing the program 221 stored in the memory 202 or storage 203. The control unit 23 includes an information acquisition control unit 231 that acquires information sent from the server device 1, and a display unit 232 that displays the acquired information on the output device 207.

[0036] Next, the processing flow of the information processing system according to this embodiment will be explained using Figure 6. Figure 6 is a flowchart showing an example of the processing performed by the information processing system 10 according to this embodiment. In this embodiment, the information processing system 10 autonomously and integrally performs application and infrastructure design, configuration, verification, and modification processing based on natural language input transmitted from the user terminal 2.

[0037] First, in step S101, the reception unit 131 receives an input request in natural language sent from the user terminal 2. This input request contains a description in natural language of the system the user wants to implement, such as "I want to create an e-commerce site" or "I want an internal attendance management system."

[0038] Next, in step S102, the requirements extraction unit 132 analyzes the natural language input received by the reception unit 131 and extracts application requirements and infrastructure requirements while referring to the requirements information 121. Specifically, it uses a large-scale language model (LLM) to classify and extract functional requirements, non-functional requirements, security requirements, performance requirements, etc., and stores them in the requirements information 121. For example, the requirements extraction unit 132 may automatically derive requirements such as product management functions, payment functions, high availability requirements, and PCI-DSS compliance from the input "EC site".

[0039] Next, in step S103, the code generation unit 133 generates application code based on the extracted application requirements and stores the generation results in the code information 122. In parallel, in step S104, the infrastructure configuration determination unit 134 analyzes the application code generated by the code generation unit 133 and determines the necessary infrastructure configuration based on the infrastructure requirements, while referring to the infrastructure configuration information 123. At this time, the optimal configuration is determined by considering domain knowledge such as cloud environment, on-premises environment, and local environment, as well as the estimated costs of multiple configuration options.

[0040] Next, in step S105, the code adjustment unit 135 adjusts the application code by referring to the infrastructure configuration constraints stored in the infrastructure configuration information 123. The adjusted code is then re-stored in the code information 122. Note that the infrastructure configuration determination step S104 and the code adjustment step S105 are executed repeatedly with mutual feedback until consistency is achieved between the application code and the infrastructure configuration.

[0041] Next, in step S106, the security verification unit 136 automatically performs security verification on the code information 122 and infrastructure configuration information 123 while referring to the security policy 124. Then, in step S107, the operation verification unit 137 works in conjunction with the simulated environment to verify the operation of the code information 122 and infrastructure configuration information 123, and records the verification results in the modification history 125. Specifically, in a simulated environment that can reproduce operation equivalent to that of the production environment, the application's functional operation, response time, throughput, resource utilization, etc. are measured, and log data and performance data are recorded.

[0042] Subsequently, in step S108, the determination unit 138 determines whether correction is necessary based on the verification results of the security verification unit 136 and the operation verification unit 137. If it is determined that correction is necessary (YES), in step S109, the correction unit 139 corrects the code information 122 or infrastructure configuration information 123 while referring to the correction history 125 and records the contents in the storage unit 12. Specifically, if the determination unit 138 determines that correction is necessary, it analyzes past correction history, learned patterns, or log data generated during operation verification stored in the correction history 125 and applies or regenerates a correction template using an LLM or reinforcement learning model. After correction, the process returns to step S106, and the cycle from security verification is executed again. This continuous correction and verification loop ensures the safety and quality of the product. On the other hand, if it is determined that no correction is necessary (NO), the process of this flowchart ends, and the generated product is ready for deployment to the production environment.

[0043] The information, modifications, and feedback obtained at each stage of this process are stored and analyzed by the learning unit 140 and used to improve the processing accuracy of the code generation unit 133 or the infrastructure configuration determination unit 134 through reinforcement learning. Each processing step is divided and executed in cooperation with multiple functional units within the server device 1, and notifications are sent to the user terminal 2 as appropriate, allowing the user to easily and safely manage the operational status.

[0044] The series of processes described above are performed asynchronously or in parallel by each functional unit within the control unit 13 as an independent software module, enabling efficient system design, construction, verification, and operation management. Each functional unit shares information and cooperates with each other, enabling dynamic and collaborative system optimization. For example, if a change in the infrastructure configuration may affect security verification, it prompts the security verification unit 136 to perform additional verification. This eliminates delays in response due to information overload and removes bottlenecks in development and operation. Furthermore, the execution results of each functional unit and user feedback are aggregated in the learning unit 140, and the processing accuracy of each functional unit is continuously improved through reinforcement learning, thereby providing users with a truly safe and optimal product.

[0045] Furthermore, the processing performed by each of the above-mentioned functional units (code generation in step S103, infrastructure configuration determination in step S104, and code adjustment in step S105) can each operate in parallel as independent processes. This allows for efficient processing of application requirements and infrastructure requirements, and enables the determination of the initial configuration of the entire system in a short time. In addition, each functional unit shares information with each other and operates cooperatively; for example, if the code generated by the code generation unit 133 generates a new requirement for the infrastructure configuration, it prompts the infrastructure configuration determination unit 134 to consider additional configurations. Moreover, the control unit 13 monitors the number of mutual feedbacks or processing time between the infrastructure configuration determination unit 134 and the code adjustment unit 135, and can optimize the overall efficiency of the system by dynamically adjusting the processing priority or allocated resources as needed. This eliminates the technical gap in the transition from a prototype to a production-ready system and enables rapid development, allowing users to build high-quality systems without infrastructure expertise. Furthermore, the determination of whether or not correction is necessary by the judgment unit 138 is optimized to ensure sufficient quality while avoiding unnecessary correction loops by comprehensively evaluating information from the security verification unit 136 and the operation verification unit 137.

[0046] Furthermore, the correction unit 139 achieves highly accurate corrections by combining past successful patterns accumulated in the correction history 125, solutions generated by a large-scale language model (LLM), and correction strategies optimized through reinforcement learning. This correction process is executed iteratively until both security and operational verification are cleared, and the correction content and results in each iteration are analyzed and learned by the learning unit 140. The learning unit 140 continuously improves the parameters of each functional unit based on user feedback or indicators such as the correction success rate, using reinforcement learning algorithms such as DPO (Direct Preference Optimization). As a result, the system becomes smarter the more it is used, enabling the generation of a high-quality system with fewer corrections. In addition, even after deployment to the production environment, it can flexibly respond to changing business requirements or technological environments by accepting natural language feedback from users and maintaining a continuous improvement cycle.

[0047] The series of processes in this embodiment integrates application development and infrastructure construction, which were previously separate, enabling a significant reduction in development time. In particular, thorough pre-verification in a simulated environment minimizes the risk of failures in the production environment, and automating quality assurance during the transition from prototype to product allows developers to focus on core business logic. This provides an innovative development environment in which anyone can safely develop high-quality systems, regardless of their technical expertise.

[0048] Next, using Figure 7, the overall configuration of the information processing system 10 according to this embodiment, in which multiple AI agents work in cooperation, will be explained. Figure 7 is a schematic diagram showing how user U1, server device 1, user terminal 2, and multiple AI agents (coding agent A1, application generation agent A2, infrastructure construction agent A3) work together to realize each function of the information processing system 10 according to this embodiment. In this embodiment, an AI agent refers to a software module that uses artificial intelligence technology such as natural language processing or inference algorithms to autonomously learn, judge, and process information in response to instructions from the user.

[0049] First, user U1 sends a request in natural language to server device 1 via user terminal 2. This request describes an outline of the application or service that the user wants to realize. After receiving this input request, server device 1 autonomously assigns a series of processes, such as requirements extraction, design, and verification, to each AI agent. For example, user U1 inputs an application generation request in natural language, such as "I want to create an e-commerce site," via user terminal 2 and sends it to server device 1. In other words, coding agent A1 receives the input from user U1 (step S101 in Figure 6) at reception unit 131, analyzes the input natural language using requirements extraction unit 132, and extracts application requirements and infrastructure requirements (step S102 in Figure 6). At this time, coding agent A1 utilizes a large-scale language model (LLM) to automatically derive detailed requirements such as functional requirements, non-functional requirements, security requirements, and performance requirements, and stores them in requirements information 121. Then, coding agent A1 instructs application generation agent A2 with the extracted application requirements, and instructs infrastructure construction agent A3 with the extracted infrastructure requirements.

[0050] Application generation agent A2 generates application code using the code generation unit 133 based on application requirements from coding agent A1 and stores it in code information 122 (step S103 in Figure 6). Meanwhile, infrastructure construction agent A3 determines the infrastructure configuration necessary for the application code to operate using the infrastructure configuration determination unit 134 based on infrastructure requirements from coding agent A1 and stores it in infrastructure configuration information 123 (step S104 in Figure 6). At this time, infrastructure construction agent A3 utilizes specialized domain knowledge regarding cloud or on-premises environments to determine the optimal configuration that satisfies cost and performance requirements. These agents execute their respective processes independently and in parallel, while feeding information back to each other as needed.

[0051] Infrastructure construction agent A3 notifies application generation agent A2 of the constraints of the determined infrastructure configuration. Application generation agent A2, receiving these infrastructure configuration constraints, adjusts its application code using the code adjustment unit 135 and updates the code information 122 (step S105 in Figure 6). This infrastructure configuration determination and code adjustment process is repeated until consistency is achieved between the application code and the infrastructure configuration. Each agent utilizes its respective domain knowledge to perform operational optimizations based on constraints specific to the execution environment, such as cloud or on-premises, security standards, and industry guidelines. Furthermore, reinforcement learning technology is applied to the entire process, and by learning from the modifications and operational feedback, the processing accuracy or optimization capability for subsequent processes is continuously improved.

[0052] After consistent application code and infrastructure configuration are generated, coding agent A1 instructs application generation agent A2 and infrastructure construction agent A3 to start an integrated verification cycle. Application generation agent A2 and infrastructure construction agent A3 deploy the jointly generated product (application code and infrastructure configuration) to a simulated environment and perform security verification based on security policy 124 using security verification unit 136 (step S106 in Figure 6). Then, they perform functional and performance verification using operation verification unit 137 (step S107 in Figure 6). This verification cycle is autonomously repeated as a "simulated environment loop" and a "security check loop" for the purpose of robustness, cost optimization, cost minimization, etc.

[0053] The verification results are reported from application generation agent A2 and infrastructure construction agent A3 to coding agent A1. Coding agent A1 uses the determination unit 138 to determine whether corrections are necessary based on the reported security verification results and operational verification results (step S108 in Figure 6). If corrections are determined to be necessary, coding agent A1 uses the correction unit 139 to issue correction instructions to application generation agent A2 or infrastructure construction agent A3, while referring to the correction history 125. Each agent that receives the instructions corrects the code information 122 or infrastructure configuration information 123 using past patterns, LLM, and reinforcement learning models stored in the correction history 125 (step S109 in Figure 6), and returns to the integrated verification cycle.

[0054] This collaborative verification and correction loop between agents is repeated until it is guaranteed that the product operates safely and optimally. If it is determined that no corrections are needed, the product is ready for deployment to the production environment. Coding agent A1 uses the learning unit 140 to continuously learn from the information obtained from each stage of this generation, verification, correction, and deployment process. In particular, it utilizes reinforcement learning techniques such as DPO (Direct Preference Optimization) to continuously optimize the code generation, infrastructure construction, and correction strategies of application generation agent A2 and infrastructure construction agent A3 based on user feedback, verification results, and correction success rates. As a result, the system becomes smarter the more it is used, enabling it to generate higher-quality products more efficiently.

[0055] Thus, each AI agent in this embodiment possesses independent judgment capabilities and expertise, and autonomously and collaboratively performs assigned processing in response to requests from user U1, thereby realizing a process for automatically generating and optimizing secure and high-quality applications and infrastructure from natural language ideas. Each agent performs processing in a shared or coordinated manner within each functional unit of the server device 1 (reception unit 131, requirements extraction unit 132, code generation unit 133, infrastructure configuration determination unit 134, code adjustment unit 135, security verification unit 136, operation verification unit 137, judgment unit 138, correction unit 139, learning unit 140). Through a loop of repeatedly executed simulated environment verification and security checks, and an advanced self-optimization mechanism based on domain knowledge and reinforcement learning, end-to-end automation and quality assurance are realized from development to operation. As a result, users, regardless of their expertise, can repeatedly enjoy the rapid, secure construction and improvement of high-quality applications and infrastructure.

[0056] As described above, this embodiment enables the creation of an integrated development environment that allows for the reliable and secure construction of systems when necessary, while minimizing the development burden due to the user's lack of infrastructure expertise, even when automatically generating a production-ready application and infrastructure integrated system from natural language input. In particular, iterative quality improvement through operational verification and security verification in a simulated environment, along with continuous improvement through reinforcement learning, contributes to a significant improvement in system development efficiency and quality. This reduces the shortage of infrastructure engineers, high development costs, and the gap in transition from prototype to product that occurred with conventional development methods.

[0057] According to this embodiment, in system development from natural language input, each processing unit in the control unit 13 operates autonomously and collaboratively, continuously analyzing large amounts of requirements information and generating and providing only safe and high-quality systems to the user. In other words, when a user inputs development requests in natural language, requirements extraction, code generation, infrastructure configuration determination, and code adjustment are automatically performed sequentially in the background, and only those that pass security and operational verification are presented quickly and integrated as application infrastructure systems. This frees the user from complex infrastructure design work, allowing them to concentrate on the core business logic. By ensuring that only systems that have passed thorough pre-verification in a simulated environment are generated, even users without technical expertise can intuitively trust the quality and safety of the system. At each stage of system generation, modification history and learning data are dynamically accumulated, and continuous quality improvement is achieved through reinforcement learning. These can be flexibly applied to diverse development environments such as startup environments, enterprise operations, and various cloud platforms. Furthermore, because each processing unit operates independently, comprehensive development support is realized from the perspectives of application quality, infrastructure optimization, and security assurance. These processing units can also be implemented as so-called AI agents, enabling more advanced collaborative processing.

[0058] Each processing unit in this embodiment utilizes large-scale language models (LLMs) and reinforcement learning technologies, but is not limited to these. Depending on the application or system requirements, specialized machine learning models, rule-based engines, or hybrid generation systems may be used. For example, the code generation unit 133 may combine a highly accurate programming language-specific model with real-time syntax checking, the infrastructure configuration determination unit 134 may use a model specialized in natural language processing, and the security verification unit 136 may use a model excellent in vulnerability pattern recognition. By selecting appropriate technologies according to the development content, the overall efficiency and accuracy of the system can be optimized. Furthermore, each processing unit used in this embodiment can be selected according to the development scale, system requirements, and network configuration, regardless of whether it is large-scale, medium-scale, or small-scale, utilizing AI technologies such as LLMs. For example, depending on the type and complexity of each development request, the system configuration may involve distributing the generation and verification processes across multiple servers in the cloud or local environment to coordinate their operation. In addition, even in large-scale development environments that integrate and manage multiple users and multiple projects, scalability and efficiency can be achieved by flexibly allocating each processing unit.

[0059] Furthermore, this embodiment shows a configuration in which each processing unit is implemented in server device 1. However, it is not limited to this configuration, and for example, the functions of each processing unit may be distributed across a cloud distributed environment, an edge computing environment, or multiple on-premises servers, and operated in a coordinated manner via a network.

[0060] Furthermore, it is possible to implement a scalable configuration by placing each processing unit as a microservice in an independent container or serverless environment. In particular, when the number of development requests increases, horizontal scaling of each processing unit allows for large-scale development while maintaining performance. In addition, in an edge computing environment, it is possible to optimize the deployment according to the characteristics of the processing, such as performing real-time verification where latency is critical on a nearby edge server, and performing computationally intensive code generation in a cloud environment. Moreover, the configuration between server device 1 and user terminal 2 can be freely expanded according to various development forms such as on-premise development, cloud infrastructure, hybrid, multi-tenant, and distributed systems. For example, each processing unit can be externally implemented on a dedicated server or external service and linked via API from the core system, or individual functional modules can be combined and applied to external development tools. In addition, each processing flow and data transmission means such as requirements management, code generation, verification control, and user management can be configured to combine according to development requirements, such as real-time generation, delayed processing, and batch processing.

[0061] While this embodiment describes each processing unit, it is not limited to these. For example, a database design unit may be added to automatically generate data schemas, or a test generation unit may be added to automatically create unit tests or integration tests. A document generation unit may also be added to automatically create specifications or operation manuals for the generated system. Furthermore, a performance optimization unit may be added to identify bottlenecks and propose optimizations. By arranging multiple security verification units 136 in parallel, each performing specialized verification from different perspectives (code vulnerabilities, infrastructure security, data protection, etc.), a more comprehensive security guarantee can be achieved. These processing units can also be implemented as AI agents, as shown in Figure 7, to realize more flexible and advanced collaborative processing.

[0062] Furthermore, this embodiment is not limited to development operations by a single user, but may be integrated with collaborative development by multiple developers, access control using role-based access control (RBAC), and approval workflows. The flow shown in Figure 6 allows each processing unit to work cooperatively even in a team development environment, enabling parallel processing of multiple development requests. Various development processes and code generation methods may also be configured to combine real-time generation, periodic batch processing, event-driven processing, predictive generation, etc. In addition, logs such as development history, revision history, and learning history acquired at each processing stage can be accumulated in the learning unit 140 and used for quality analysis, trend identification, audit response, knowledge base construction, etc. Furthermore, the learning agent can optimize various verification processes and data collection methods, combining real-time monitoring, periodic batch processing, event-driven processing, predictive monitoring, etc. The learning agent accumulates logs such as acquired verification history, revision history, and response history in the revision history 125, which can be used for incident analysis, trend identification, audit response, knowledge base construction, etc. Furthermore, the learning agent can use machine learning to learn past response patterns and automatically suggest recommended solutions when similar problems occur.

[0063] Furthermore, this embodiment can be applied to a wide range of industries, including fintech app development in financial institutions, IoT system construction in manufacturing, medical support system development in healthcare institutions, e-commerce site construction in retail, and learning management system development in educational institutions. Following the flow shown in Figure 6, it can provide consistent automated processing from requirements extraction to final system generation for development requests in these various industries. These functions are expandable in line with the evolution of AI technology and the promotion of DX in development sites, and by adopting an agent configuration as shown in Figure 7, more advanced and flexible system development support can be achieved.

[0064] While this embodiment primarily describes visual notifications via a management screen, it is also possible to configure the system as a multimodal notification system, including voice notifications, vibration notifications, AR displays, and notifications to wearable devices, through a notification agent (not shown in the diagram, but responsible for integrating the notification results of each agent). Furthermore, by implementing an adaptive control function that dynamically adjusts the execution frequency of each AI agent according to the system state and network environment, stable operation can be achieved at various scales and in various environments.

[0065] As described above, according to this embodiment, users can quickly deliver high-quality products by using natural language input, without having to manually design and verify individual application code and infrastructure configurations. The sequential execution of each processing unit enables safe and high-quality system development even for users without infrastructure expertise. Furthermore, iterative quality improvement through operational verification and security verification in a simulated environment minimizes the risk of failure in the production environment. In addition, continuous improvement through reinforcement learning makes the system smarter the more it is used, resulting in more efficient and secure system development. These processes can also be implemented as an AI agent, as shown in Figure 7, in which case more advanced collaborative processing and flexible scalability can be achieved.

[0066] Although one embodiment of the present invention has been described above, the present invention is not limited to the embodiments described above, and any modifications, improvements, etc. that can achieve the objectives of the present invention are included in the present invention.

[0067] Furthermore, the series of processes described above can be executed by hardware or by software. In other words, the functional configuration is merely illustrative and not particularly limited. That is, it is sufficient for the information processing system to have the functionality to execute the series of processes described above as a whole, and there is no particular limitation on what functional blocks are used to realize this functionality. Also, the location of the functional blocks is not particularly limited and can be arbitrary. For example, the functional blocks of a server may be delegated to a user terminal, etc. Conversely, the functional blocks of a user terminal may be delegated to a server, etc. Moreover, a single functional block may be composed of hardware alone, software alone, or a combination of both.

[0068] For example, when a series of processes are executed by software, the programs that make up that software are installed on a computer or other device from a network or storage medium. The computer may be a computer built into dedicated hardware. Alternatively, the computer may be a computer capable of performing various functions by installing various programs, such as a server, a general-purpose smartphone, or a personal computer.

[0069] Furthermore, for example, a recording medium containing such a program may consist not only of removable media (not shown) distributed separately from the main unit of the device to provide the program to the user, but also of a recording medium provided to the user in a state where it is pre-installed in the main unit of the device.

[0070] In this specification, the step of describing a program to be recorded on a recording medium includes not only processes that are performed chronologically in that order, but also processes that are not necessarily performed chronologically, but are executed in parallel or individually. Furthermore, in this specification, the term "system" refers to an overall system composed of multiple devices, means, etc.

[0071] (Modification 1: Retraining and template application integrated automatic generation) Furthermore, looking at the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "continuously learning past deliverables, revision history, and operational results, recommending and automatically applying optimal templates and patterns to similar requirements, and rapidly generating highly accurate deliverables in a verification and correction loop."

[0072] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, A history reference step that extracts requirements from natural language input and searches for past deliverables, revision history, and operational performance, A similarity / type evaluation step to assess the similarity or type of the aforementioned requirements and past examples, and to identify a suitable template / design pattern, A template application and generation step that automatically applies the aforementioned template / pattern to generate the design, code, and infrastructure configuration, A self-optimization step that performs verification and automatic correction on the generated product and updates the results as history, A continuous learning step that continuously collects operational data and updates the learning model and template set. Information processing methods including [this].

[0073] According to the above configuration, first, the requirements extraction unit 132 analyzes the new input, and at the same time, the learning unit 140 automatically searches for similar projects from the memory unit 12 and identifies past cases with a success rate above a certain level (history reference step). The similarity of the requirements is quantified using vector similarity and semantic search, and the optimal design pattern is extracted from cases with a similarity above a certain level (similarity / type evaluation step). Based on the extracted template, the code generation unit 133 customizes it by reusing a certain amount of existing implementation, and the infrastructure configuration determination unit 134 automatically applies a proven configuration (template application generation step). The security verification unit 136 and the operation verification unit 137 perform automatic verification, and the correction unit 139 automatically corrects the necessary parts based on past correction patterns and stores them in the correction history 125 (self-optimization step). After the start of operation, performance indicators, user satisfaction, failure rate, etc. are continuously collected, and the accuracy of template selection is improved using the DPO (Direct Preference Optimization) algorithm (continuous learning step). As a result, the initial development phase achieved three times the normal development speed, and when addressing similar requirements, both speed and quality improvement could be achieved simultaneously.

[0074] (Variation 2: Parallel generation of multiple deliverables and collaborative development integration) Furthermore, looking at the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "automatically optimizing dependency relationships, task sharing, and collaboration design by coordinating simultaneous requests from multiple users while simultaneously generating, verifying, and managing multiple deliverables in parallel."

[0075] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, A parallel input / processing step that accepts multiple inputs at once and processes them in parallel on a deliverable-by-deliver basis, A dependency and commonality analysis step analyzes the dependencies between deliverables and requirements to determine the optimal order and common elements, The collaborative design and task allocation optimization steps involve coordinating the requirements of multiple users, assigning priorities, and suggesting appropriate task assignments. The integration design step involves sharing and applying common templates and infrastructure settings, and designing the integration interface. An integrated management output step that manages and outputs progress, history, and verification results in individual / integrated views. Information processing methods including [this].

[0076] According to the above configuration, first, the reception unit 131 receives multiple parallel development requests, and the control unit 13 starts worker processes in parallel for each request, initiating simultaneous processing in independent memory spaces (parallel reception and processing step). The requirements extraction unit 132 automatically analyzes the technical dependencies between each request (database sharing, API integration, etc.), determines the optimal processing order using graph theory, and identifies modules that can be commonly used (dependency and commonization analysis step). The learning unit 140 analyzes each developer's area of ​​expertise (frontend, backend, infrastructure, etc.) and past success rates, and proposes the optimal division of labor using a task assignment optimization algorithm (collaborative design and division of labor optimization step). The common database schema, authentication module, logging function, etc. are implemented in a unified manner, and integration APIs between microservices are automatically generated (integration design step). The progress rate, test results, and dependencies of each deliverable are visualized on a real-time dashboard, providing both individual project details and an overall overview (integrated management output step). As a result, overall development efficiency is improved compared to individual development, and operational costs can be reduced through commonization.

[0077] (Variation 3: Integrated change management with differential extraction and partial updates) Furthermore, looking at the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "achieving safe and efficient partial updates by consistently automating the process from extracting differences and analyzing impacts on existing deliverables to minimal change design, refactoring, verification, history management, and rollback."

[0078] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, A difference and impact analysis step involves extracting differences from existing deliverables and new requirements as input to identify the scope of impact, A minimum change design step generates differential design and refactoring proposals that satisfy requirements with the minimum changes for a specific range, The compatibility implementation step involves implementing fixes and feature additions while maintaining compatibility, A diff-focused validation step that automatically executes regression tests and integrated validation specific to the changes, Change history management steps include recording version history, rollback, and saving differential verification results. Information processing methods including [this].

[0079] According to the above configuration, the requirements extraction unit 132 first performs a detailed analysis of the existing system's components (database schema, API specifications, business logic, etc.), automatically extracts the differences with the new requirements, and identifies areas requiring change through impact propagation analysis (difference and impact analysis step). The code adjustment unit 135 generates a differential design that minimizes the impact on existing functions for the identified scope of impact, and simultaneously executes refactoring proposals (minimum change design step). The code generation unit 133 guarantees version compatibility with existing APIs and automatically implements a phased migration function when the database schema is changed (compatibility maintenance implementation step). The operation verification unit 137 automatically generates a regression test suite specialized for the changed areas and their dependencies, and prioritizes the execution of tests that detect adverse impacts on existing functions (difference-focused verification step). The modification unit 139 automatically records detailed snapshots before and after changes, code differences, and configuration change history, and provides a rollback function that allows recovery to the previous version with one click in case of a problem (change history management step). As a result, development time can be shortened and change risks reduced compared to a full redevelopment.

[0080] (Variation 4: Automatic execution environment switching and migration support) Furthermore, if we consider the aforementioned information processing system from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "automatically designing and executing migration from an existing on-premises environment to the cloud, etc."

[0081] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, An input step that accepts current deliverable information and new environment specification information as input, A design step that automatically designs migration procedures, risks, and configuration optimizations for the current configuration and the new environment, The migration, conversion, and generation steps for code and infrastructure configurations, Migration verification and migration history recording steps, Steps for automatically deploying deliverables to the new environment Information processing methods including [this].

[0082] According to the above configuration, first, the infrastructure configuration determination unit 134 performs a detailed analysis of the current on-premises configuration (server specifications, network configuration, storage capacity, etc.) and automatically selects the optimal migration destination (input step). The migration plan generation algorithm formulates a phased migration strategy, quantifies data migration time, downtime, migration cost, and risk assessment, and determines the optimal migration procedure (design step). The code adjustment unit 135 performs an automatic conversion to a cloud-native architecture, transforming the conventional monolith into microservices and the database into a managed service (migration conversion / generation step). The operation verification unit 137 performs a simulated production test in the new environment, completes performance comparison, functional verification, and security checks, and then records a detailed migration log (migration verification / migration history recording step). The Blue-Green deployment method maintains the old environment while putting the new environment into production, and provides an automatic rollback function in case of problems (automatic deployment of deliverables to the new environment step). As a result, downtime can be reduced and the risk of migration failure can be lowered compared to manual migration.

[0083] (Modification 5: Operational learning-driven predictive and dynamic optimization) Furthermore, looking at the aforementioned information processing system from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "achieving both planned optimization that considers future load based on operational data and growth forecasts, and dynamic optimization based on real-time performance monitoring."

[0084] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, A predictive analysis step that estimates future load based on past performance and business growth data, A planned optimization step involves designing a phased scaling plan and bottleneck avoidance measures, A real-time monitoring and analysis step to monitor and analyze performance in operation in real time and identify bottlenecks, The process involves generating proposed configuration changes and pre-verifying their effects in a simulated environment, as well as the following steps: The non-stop application and continuous learning step safely applies changes using a non-stop deployment method and reflects the results in learning. Information processing methods including [this].

[0085] According to the above configuration, the learning unit 140 first analyzes past traffic trends, user growth rates, and sales growth rates using machine learning to predict future load (predictive analysis step). The infrastructure configuration determination unit 134 formulates a phased scale-out plan based on the prediction results and determines the scaling settings (planned optimization step). The real-time monitoring function monitors response time, throughput, and resource utilization at predetermined intervals and detects bottlenecks in advance using an anomaly detection algorithm (real-time monitoring and analysis step). Optimization proposals (cache expansion, database optimization, CDN placement changes, etc.) are generated for the detected bottlenecks, and the performance improvement effect is pre-verified in a simulated environment (change proposal generation and pre-verification step). The verified change proposals are applied to production without interruption using Blue-Green deployment, and the application results are accumulated as learning data to improve future prediction accuracy (non-stop application and continuous learning step). As a result, unplanned downtime can be reduced and infrastructure operating costs can be optimized.

[0086] (Modification 6: Domain knowledge integration and autonomous domain extension) Furthermore, looking at the aforementioned information processing system from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "incorporating external knowledge, industry standards, and case studies to automatically extract industry requirements and design compliance, while also adapting and expanding to unfamiliar domains through self-learning."

[0087] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, Domain information acquisition and integration step involves identifying and acquiring information from the target industry and related external databases, and integrating that knowledge. Industry requirements extraction and integration steps involve extracting laws, regulations, standards, and practices and integrating them into requirements, Industry-standard application and verification steps that apply industry-standard architecture and technology stacks and reflect verification criteria, An autonomous domain extension step that continuously learns information from unknown fields and extends the model and template, A generation and output step that generates and outputs deliverables based on the aforementioned reflection results. Information processing methods including [this].

[0088] According to the above configuration, the requirements extraction unit 132 first identifies the target industry from the input content and automatically selects and acquires relevant legal regulations and technical standards databases for each industry (domain information acquisition and integration step). Using natural language processing, it extracts specific requirements (encryption strength, log retention period, access control, etc.) from legal documents and integrates them with user input requirements to generate a compliance requirements list (industry requirements extraction and integration step). The infrastructure configuration determination unit 134 and code generation unit 133 apply industry-proven architecture patterns and implement industry-specific verification items (vulnerability scanning, data encryption verification, etc.) (industry standard application and verification step). When the learning unit 140 detects input for a new industry, it crawls external databases (e.g., industry association sites, technical standards documents, case studies, etc.) and adapts and extends existing knowledge to the new domain through transfer learning (autonomous domain extension step). Based on the integrated industry knowledge and standards, it generates the final deliverable and outputs it along with a compliance checklist (generation output step). As a result, even developers without industry expertise can achieve expert-level design quality and reduce the risk of compliance violations.

[0089] (Modification 7: Multimodal input compatible extended system) Furthermore, looking at the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "to integrate and analyze various input formats other than natural language (charts, audio, images, etc.) to achieve more intuitive and comprehensive requirements extraction."

[0090] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, A multimodal input acceptance step that accepts multiple types of input data, including natural language, audio, images, and charts, A data integration and transformation step involves analyzing each input data with the corresponding recognition engine and integrating and transforming it into text data. The relationship analysis step involves analyzing the relationships between each input format from the integrated and transformed data to extract comprehensive requirements. A requirements consistency verification step that verifies the consistency of extracted requirements and automatically fills in any inconsistencies or omissions, An integrated generation step that generates application code and infrastructure configuration based on the aforementioned consistency-verified requirements. Information processing methods including [this].

[0091] According to the above configuration, the user can first input not only a description in natural language, but also handwritten screen design diagrams, supplementary voice explanations, images of reference systems, etc. (multimodal input acceptance step), and the data in each format is analyzed by the corresponding recognition engine (image recognition AI, voice recognition AI, OCR, etc.) (data integration and transformation step). The requirements extraction unit 132 analyzes the relationships between the information obtained from these different input formats and extracts comprehensive requirements that cannot be captured by a single input format (relationship analysis step). Furthermore, it automatically detects contradictory and missing requirements between each input format and supplements them after confirmation with the user (requirements consistency verification step). Through this multifaceted requirements extraction, a more accurate and complete system generation is achieved (integration generation step).

[0092] (Variation 8: Automated test case generation and execution system) Furthermore, if we consider the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "to automatically generate and execute comprehensive tests on the generated system and realize continuous quality assurance."

[0093] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, The test target extraction step involves analyzing the generated application code and infrastructure configuration to extract the elements to be tested, A test case automatic generation step that automatically generates various test cases for unit tests, integration tests, system tests, and performance tests for the aforementioned elements under test, The smart test plan generation step involves analyzing past defect history and focusing testing on areas prone to bugs, and The automated test execution step automatically runs the generated test cases in a simulated environment and collects and analyzes the results. A test result feedback step identifies areas that need correction based on test results and performs automatic correction or suggests corrections. Information processing methods including [this].

[0094] According to the above configuration, first, the operation verification unit 137 analyzes all components of the generated system and extracts the functions, APIs, data flows, etc. to be tested (test target extraction step). For each extracted element, unit tests based on functional specifications, integration tests verifying inter-module cooperation, end-to-end system tests, and load / performance tests are comprehensively and automatically generated (automatic test case generation step). The learning unit 140 analyzes past failure patterns accumulated and identifies processes or data patterns prone to bugs, generating a focused test plan (smart test plan generation step). The generated test cases are automatically executed in a simulated environment, and the result data is collected and analyzed (automatic test execution step). The correction unit 139 attempts to correct any test failures, and if correction is difficult, a detailed correction suggestion is presented to the user (test result feedback step).

[0095] (Experiment 9: Multilingual and multicultural automated localization) Furthermore, if we consider the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "automatically generating a multilingual and multicultural system based on global deployment, and realizing localization that conforms to the legal regulations and cultural requirements of each region."

[0096] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, A localization requirements extraction step involves obtaining language, cultural, and legal / regulatory information for the target market from a memory unit and extracting localization requirements. A multilingual UI generation step that automatically generates UI design and user experience tailored to the linguistic characteristics of each region, Legal compliance implementation steps to implement data processing functions that comply with the privacy laws and data protection regulations of each country, Cultural adaptation steps that incorporate region-specific business logic and workflows, including cultural considerations, The inter-regional collaboration step involves verifying the consistency of each regional system and implementing inter-regional data linkage functions. Information processing methods including [this].

[0097] According to the above configuration, first, the requirements extraction unit 132 obtains the language system, cultural characteristics, and legal and regulatory environment (GDPR, CCPA, Personal Information Protection Act, etc.) of the target deployment region from the storage unit 12 and extracts region-specific localization requirements (localization requirements extraction step). The code generation unit 133 generates a UI design that takes into account the character direction (left-to-right writing, right-to-left writing), date format, numerical notation, color culture, etc. for each language (multilingual UI generation step). The security verification unit 136 implements data encryption, access control, data deletion rights, etc. in accordance with the personal data protection laws of each country (legal and regulatory compliance implementation step). Business logic that reflects the business practices, payment methods, holiday calendars, business hours, etc. of each region is incorporated (cultural adaptation step). Data synchronization and consistency assurance functions between regional versions are implemented to ensure global consistency (inter-regional collaboration step).

[0098] (Variation 10: AI Model Integrated Development Platform) Furthermore, looking at the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "supporting the automated development of intelligent applications that integrate machine learning models and AI services, and realizing the construction of AI-native systems."

[0099] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, The AI ​​function identification step involves identifying the need for AI functionality from application requirements and classifying appropriate machine learning tasks, The AI ​​model selection step involves selecting the most suitable pre-trained model or API service for each machine learning task, An AI integrated code generation step that generates application code that integrates the selected AI model and business logic, The AI ​​optimization function implementation steps include automatically implementing continuous learning and A / B testing functions for the AI ​​model, The AI ​​model continuous improvement step involves continuously monitoring the prediction accuracy and processing performance of the AI ​​model and automatically executing model updates. Information processing methods including [this].

[0100] According to the above configuration, first, the requirements extraction unit 132 analyzes the application requirements and identifies the need for AI functions such as image recognition, natural language processing, predictive analytics, and recommendation (AI function identification step). From the pre-trained model library or API service information stored in the memory unit 12, the most suitable AI model for each task is selected (AI model selection step). The code generation unit 133 automatically generates integrated application code including the inference pipeline, pre-processing / post-processing, and error handling of the selected AI model (AI integrated code generation step). The learning unit 140 automatically implements the AI ​​model's continuous learning function, A / B testing environment, and model performance monitoring function (AI optimization function implementation step). After operation begins, prediction accuracy, response time, and resource consumption are continuously monitored, and automatic model updates and hyperparameter optimization are performed when performance deteriorates (AI model continuous improvement step).

[0101] (Variation 11: Zero Trust Security Automated Implementation System) Furthermore, looking at the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "to realize the construction of a system that automatically implements a zero-trust architecture and comes standard with advanced cybersecurity."

[0102] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, A security policy generation step that analyzes all communication paths within the system and automatically generates a zero-trust security policy, An encryption authentication implementation step that automatically implements end-to-end encryption and multi-factor authentication on all communication paths, A microsegmentation implementation step involves microsegmenting the network and implementing least privilege access control. The continuous authentication implementation step involves implementing continuous authentication of users and devices and anomaly detection through behavioral analysis, A security dynamic update step that dynamically updates security settings based on threat intelligence information, and Information processing methods including [this].

[0103] According to the above configuration, first, the security verification unit 136 analyzes all data flows, API communications, and user access patterns within the system and automatically generates a zero-trust security policy based on the "trust nothing" principle (security policy generation step). TLS encryption, certificate-based authentication, multi-factor authentication (MFA), etc., are implemented on all communication paths to guarantee the confidentiality and integrity of communications (encryption authentication implementation step). The infrastructure configuration determination unit 134 subdivides the network and implements access control rules based on the principle of least privilege between each segment (microsegmentation implementation step). The learning unit 140 learns normal user behavior patterns and continuously monitors and detects abnormal access attempts or privilege escalation attacks (continuous authentication implementation step). In conjunction with external threat intelligence information, firewall rules and intrusion detection system settings are dynamically updated based on new threat information (dynamic security update step).

[0104] (Variation 12: Automated design of edge-cloud integration system) Furthermore, looking at the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "optimally linking edge computing and the cloud to automatically construct a next-generation hybrid system that responds to real-time requests."

[0105] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, Processing distribution optimization steps involve analyzing data characteristics, latency requirements, and communication costs to determine the optimal processing distribution between the edge and the cloud. An edge system design step that automatically designs the edge-side system configuration considering the processing power and availability of edge devices, The cloud integration platform design step involves designing an integrated cloud platform for large-scale data processing and machine learning on the cloud side, A load balancing control implementation step that implements dynamic load balancing and failover functionality between edge and cloud, Implementation steps for implementing automatic cloud switching and data synchronization functions in the event of an edge failure. Information processing methods including [this].

[0106] According to the above configuration, first, the infrastructure configuration determination unit 134 analyzes the characteristics of IoT sensor data, video data, audio data, etc., and determines the optimal distribution in which data requiring real-time processing is processed at the edge and data requiring large-volume analysis is processed in the cloud (processing distribution optimization step). Considering the limited computing resources of edge devices and the risk of network disruption, an edge system configuration including lightweight applications and local storage is automatically designed (edge ​​system design step). On the cloud side, a platform integrating a large-scale data processing platform, machine learning pipelines, and long-term data storage is designed (cloud integrated platform design step). A failover function is implemented to dynamically distribute processing according to network conditions and edge device load, eliminating single points of failure (load balancing control implementation step). In the event of edge device failure or network disruption, processing automatically switches to the cloud, and a data synchronization function after recovery is also implemented (automatic failure switching implementation step).

[0107] (Variation 13: Sustainability Optimization System) Furthermore, if we consider the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "automatically realizing sustainable system design aimed at minimizing environmental impact, and supporting ESG management and SDGs achievement from a technical standpoint."

[0108] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, Sustainability assessment steps that evaluate the carbon footprint and renewable energy utilization rate of each data center, A green infrastructure selection step that automatically selects infrastructure configurations and resource allocations that minimize environmental impact, Energy-saving application implementation steps include designing applications and implementing processing algorithms to optimize energy efficiency, A carbon measurement step that continuously measures power consumption and CO2 emissions during system operation, An environmental optimization step that dynamically optimizes operational parameters based on environmental impact reduction targets. Information processing methods including [this].

[0109] According to the above configuration, first, the infrastructure configuration determination unit 134 evaluates the carbon footprint, renewable energy usage rate, and PUE (Power Usage Effectiveness) value of each cloud provider and data center (sustainability evaluation step). Based on the evaluation results, the data center with the least environmental impact is selected, and the system is designed with the minimum necessary resource configuration (green infrastructure selection step). The code generation unit 133 generates application code that emphasizes CPU-efficient algorithms, optimized memory usage, and elimination of unnecessary processing (energy-saving application implementation step). The operation verification unit 137 continuously monitors power consumption, network traffic, and storage usage during system operation and calculates CO2 emissions (carbon measurement step). The learning unit 140 dynamically optimizes processing scheduling, resource utilization, and data placement based on environmental impact reduction targets (e.g., a 30% reduction in CO2 emissions) (environmental optimization step).

[0110] (Modification 14: Quantum Computing Ready System) Furthermore, if we consider the information processing system described above from a different perspective, the problem (objective) that the information processing system according to this embodiment aims to solve can also be understood as "automatically constructing a post-quantum cryptography system and a quantum algorithm utilization platform that looks ahead to the quantum computing era, and adapting to the next-generation computing paradigm."

[0111] Given the problem as described above, an invention as a means to solve the problem might look like this: "An information processing method executed by an information processing device, A quantum vulnerability assessment step that analyzes current encryption methods and evaluates their vulnerability to quantum computer threats, A post-quantum cryptography implementation step that selects a post-quantum cryptography algorithm and automatically generates a step-by-step transition plan, The quantum advantage design step involves identifying problem domains where quantum algorithms are effective and designing future applications of quantum advantages, The quantum hybrid design step involves pre-designing a quantum-classical hybrid processing architecture, A quantum technology tracking step that automatically updates the system in accordance with advances in quantum computing technology. Information processing methods including [this].

[0112] According to the above configuration, first, the security verification unit 136 analyzes currently used public-key cryptography schemes such as RSA and ECC and quantitatively evaluates the risk of decryption by quantum computers (quantum vulnerability assessment step). Based on the evaluation results, it selects post-quantum cryptography algorithms such as lattice-based cryptography, code-based cryptography, and multivariate cryptography, and automatically generates a phased migration plan that maintains compatibility with existing systems (post-quantum cryptography implementation step). The requirements extraction unit 132 identifies problem domains where quantum algorithms such as optimization problems, machine learning, and cryptography are effective, and pre-designs an architecture that can utilize future quantum advantages (quantum advantage design step). The infrastructure configuration determination unit 134 designs a hybrid processing platform that efficiently links quantum computers and classical computers (quantum hybrid design step). The learning unit 140 continuously monitors the latest trends in quantum computing technology and automatically updates the system configuration in accordance with the practical application of new quantum algorithms or quantum devices (quantum technology tracking step). [Explanation of Symbols]

[0113] 10. Information Processing Systems 1 Server Device 2. User terminals 131 Reception Department 132 Requirements extraction part 133 Code Generation Unit 134 Infrastructure Configuration Decision Department 135 Code Adjustment Section 136 Security Verification Department 137 Operation Verification Section 138 Judgment Department 139 Correction Department 140 Learning Department

Claims

1. An information processing method performed by an information processing device, A step that accepts natural language input, A requirements extraction step that extracts application requirements and infrastructure requirements based on the aforementioned natural language input, A code generation step that generates application code based on the aforementioned application requirements, An infrastructure configuration determination step in which the necessary infrastructure configuration is determined based on the aforementioned infrastructure requirements, A code tuning step that adjusts the application code based on the constraints of the infrastructure configuration, A security verification step that performs security verification on the aforementioned adjusted application code and infrastructure configuration, A verification step for verifying the operation of the application code and the infrastructure configuration in a simulated environment, A determination step that determines whether or not correction is necessary based on the results of the security verification step and the operation verification step, If the determination step determines that a correction is necessary, the correction step involves correcting the application code or the infrastructure configuration. Includes, The infrastructure configuration determination step and the code adjustment step are repeatedly executed while providing mutual feedback to each other. The process is repeated until the application code and the infrastructure configuration are consistent. Information processing methods.

2. In the requirements extraction step, the natural language input is analyzed using a large-scale language model, and requirements including functional requirements, non-functional requirements, security requirements, and performance requirements are classified and extracted. The information processing method according to claim 1.

3. In the infrastructure configuration determination step, a configuration is selected that corresponds to at least one of the following: a cloud environment, an on-premises environment, or a local environment. The information processing method according to claim 1.

4. In the aforementioned infrastructure configuration determination step, multiple infrastructure configuration options are generated, the estimated cost of each option is calculated, and a configuration that is cost-effective while meeting performance requirements is selected. The information processing method according to claim 1.

5. The aforementioned security verification step involves performing verification based on an external vulnerability information database or the company's own security checklist. The information processing method according to claim 1.

6. The modification step involves regenerating the application code or the infrastructure configuration based on at least one of the results of the security verification step and the operational verification step. The information processing method according to claim 1.

7. The simulated environment in the aforementioned operational verification step is a simulated environment of a cloud service, and is capable of reproducing the same operation as the production environment. The information processing method according to claim 1.

8. In the aforementioned operation verification step, Measure at least one of the following: response time, throughput, and resource utilization. If the measured result does not meet a predetermined threshold, Perform scaling settings for the infrastructure configuration or optimize the application code. The information processing method according to claim 1.

9. In the correction step described above, corrections are made using a large-scale language model based on past correction history and trained patterns. The information processing method according to claim 1.

10. In the correction step, the log data generated in the operation verification step is analyzed, error patterns are identified, and correction templates corresponding to the error patterns are applied. The information processing method according to claim 1.

11. After deployment to the production environment, a feedback reception step is implemented to receive user feedback in natural language. Based on the aforementioned feedback, a learning step is performed to improve the processing of the code generation step and the infrastructure configuration determination step through reinforcement learning. Includes The information processing method according to claim 1.

12. In an information processing device, A step that accepts natural language input, A requirements extraction step that extracts application requirements and infrastructure requirements based on the aforementioned natural language input, A code generation step that generates application code based on the aforementioned application requirements, An infrastructure configuration determination step in which the necessary infrastructure configuration is determined based on the aforementioned infrastructure requirements, A code tuning step that adjusts the application code based on the constraints of the infrastructure configuration, A security verification step that performs security verification on the aforementioned adjusted application code and infrastructure configuration, A verification step for verifying the operation of the application code and the infrastructure configuration in a simulated environment, A determination step that determines whether or not correction is necessary based on the results of the security verification step and the operation verification step, If the determination step determines that a correction is necessary, the correction step involves correcting the application code or the infrastructure configuration. Includes, The infrastructure configuration determination step and the code adjustment step are repeatedly executed while providing mutual feedback to each other. The process is repeated until the application code and the infrastructure configuration are consistent. A program for executing information processing methods.

13. An information processing system equipped with an information processing device, A step that accepts natural language input, A requirements extraction step that extracts application requirements and infrastructure requirements based on the aforementioned natural language input, A code generation step that generates application code based on the aforementioned application requirements, An infrastructure configuration determination step in which the necessary infrastructure configuration is determined based on the aforementioned infrastructure requirements, A code tuning step that adjusts the application code based on the constraints of the infrastructure configuration, A security verification step that performs security verification on the aforementioned adjusted application code and infrastructure configuration, A verification step for verifying the operation of the application code and the infrastructure configuration in a simulated environment, A determination step that determines whether or not correction is necessary based on the results of the security verification step and the operation verification step, If the determination step determines that a correction is necessary, the correction step involves correcting the application code or the infrastructure configuration. Perform The infrastructure configuration determination step and the code adjustment step are repeatedly executed while providing mutual feedback to each other. The process is repeated until the application code and the infrastructure configuration are consistent. Information processing system.

Images