Method and device for realizing automatic development of intelligent agent learning system through computing power by intelligent computing center

Abstract

The invention provides a method and device for an intelligent computing center to realize automatic development of an intelligent agent learning system through computing power, and the method is executed by a learning system development intelligent agent, and comprises the steps: S1, receiving a task for constructing a to-be-developed intelligent agent learning system; s2, determining a development demand of a learning system with a user according to the task in combination with a large language model method; s3, determining a design scheme of the learning system according to the development requirements in combination with a large language model method; s4, according to the development requirements and the design scheme, codes of a learning system are generated in combination with a large language model method; s5, deploying a running environment of the learning system in combination with a large language model method, and debugging codes of the learning system; when the development requirements are not met, the design scheme and codes of the learning system are checked in combination with a large language model method, the design scheme is modified, the process returns to the step S3, the codes are modified, and the process returns to the step S4 until the development requirements are met or debugging or modification exceeds times; and S6, feeding back result information to the user.

Description

Technical Field

[0001] The present invention relates to the technical fields of intelligent computing centers, intelligent computing centers, and computing power infrastructure, and particularly relates to a method and device for automatically developing an intelligent agent learning system through computing power in an intelligent computing center. Background Art

[0002] With the rapid development of artificial intelligence technology, "intelligent computing centers" and "intelligent computing centers" have emerged as the times require.

[0003] An "intelligent computing center" refers to a facility that provides the required computing power, data, and algorithms for artificial intelligence applications (such as scenarios like artificial intelligence deep learning model development, model training, and model inference) by using large-scale heterogeneous computing power resources, including general computing power and intelligent computing power. An intelligent computing center covers facilities, hardware, and software, and can provide full-stack capabilities from underlying computing power to top-level application enabling.

[0004] An "intelligent computing center" includes, but is not limited to, an "intelligent computing center".

[0005] An "intelligent computing center", that is, an artificial intelligence computing center, is a type of computing power infrastructure that provides computing power services, data services, and algorithm services required for artificial intelligence applications based on artificial intelligence theory and using an artificial intelligence computing architecture.

[0006] "Computing power" is the core of "intelligent computing centers" and "intelligent computing centers". It is the ability of computer devices or computing / data centers to process information, the ability of computer hardware and software to cooperate to jointly execute a certain computing requirement, the computing ability to achieve the output of target results by processing information data, and a new type of productive force integrating information computing power, network carrying capacity, and data storage capacity, which mainly provides services to society through computing power infrastructure.

[0007] An "intelligent agent" is an agent that can perceive the environment and take actions to achieve specific goals. It can be software, hardware, or a system, and has autonomy, adaptability, and interaction capabilities. An intelligent agent perceives changes in the environment (such as through sensors or data input), makes judgments and decisions based on the knowledge and algorithms learned by itself, and then executes actions to affect the environment or achieve a predetermined goal. Intelligent agents are widely used in the field of artificial intelligence, commonly found in automation systems, robots, virtual assistants, and game characters, etc. The core lies in their ability to learn independently and evolve continuously to better complete tasks and adapt to complex environments.

[0008] Currently, the vast majority of agents lack the ability to learn and cannot solve new tasks through learning. The development cost of the learning ability of agents is high and the efficiency is low. Existing agents tend to develop specific task processing abilities manually or inject knowledge in advance to solve the problem of improving the ability of agents, and abandon the research and development of the ability to learn and solve new tasks. Summary of the Invention

[0009] The present invention provides a method and device for automatically developing an agent learning system by computing power in an intelligent computing center, which is used to solve the problems that existing agents lack learning ability and the cost of manually developing the learning system of agents is high and the efficiency is low.

[0010] In order to solve the above technical problems, the present invention is implemented as follows:

[0011] In a first aspect, the present invention provides a method for automatically developing an agent learning system by computing power in an intelligent computing center, which is executed by a learning system to develop an agent, and includes:

[0012] Step S1: Receive task information for constructing a learning system of an agent to be developed provided by a user through an interaction interface;

[0013] Step S2: Determine the development requirements of the learning system in combination with the large language model method and the user according to the task information;

[0014] Step S3: Determine the design scheme of the learning system in combination with the large language model method according to the development requirements;

[0015] Step S4: Generate the code of the learning system in combination with the large language model method according to the development requirements and the design scheme;

[0016] Step S5: Deploy the running environment of the learning system and debug the code of the learning system in combination with the large language model method; when the debugging result indicates that the learning system does not meet the development requirements, check the design scheme and code of the learning system in combination with the large language model method. When it is necessary to modify the design scheme, return to step S3. When it is necessary to modify the code, return to step S4 until the debugging result indicates that the learning system meets the development requirements or the number of debugging or modification times exceeds a preset threshold;

[0017] Step S6: Feedback result information to the user through the interaction interface.

[0018] Optionally, the development requirements of the learning system include at least one of the following: the input method of learning materials, whether to prompt the user to provide learning key points during learning, whether to require the user to provide knowledge tags during learning, the learning capabilities required by the learning system, the learning depth, whether to provide the ability to visually access the learned knowledge, the application scenarios of the learning system in the to-be-developed intelligent agent, whether to provide practice cases, test cases, acceptance criteria, output requirements, technical requirements, performance requirements, security and privacy.

[0019] Optionally, step S2 includes:

[0020] Step S21: According to the task information, communicate with the user through the interaction interface for at least one round by combining the large language model method;

[0021] Step S22: Determine the development requirements of the learning system according to the task information and the communication content with the user by combining the large language model method.

[0022] Optionally, step S3 includes:

[0023] Step S31: According to the development requirements of the learning system, obtain learning results by using at least one of the following learning methods by combining the large language model method: search relevant materials through a search engine, view and learn open-source code, learn the existing code and documents of the to-be-developed intelligent agent;

[0024] Step S32: Determine the design scheme of the learning system according to the learning results.

[0025] Optionally, the design scheme includes at least one of the following:

[0026] Modify the task processing module of the to-be-developed intelligent agent so that the task processing module can identify whether the user task is a learning task;

[0027] Modify the upload module of the to-be-developed intelligent agent to support the upload and archiving of learning materials;

[0028] Add a learning system;

[0029] The application scenarios of the learning system in the to-be-developed intelligent agent;

[0030] Modify the task processing module in the to-be-developed intelligent agent so that the task processing module can call the knowledge extraction module of the learning system and integrate the knowledge learned by the learning system into the prompt words of the task processing module;

[0031] Add a practice module;

[0032] General design document, detailed design document, interface design document, and / or database design document.

[0033] Optionally, the learning system has at least one of the following learning capabilities: knowledge learning, knowledge archiving, knowledge extraction, knowledge matching, knowledge application, knowledge practice, knowledge display, knowledge encoding, knowledge maintenance, knowledge version maintenance, multi-document knowledge learning, large-document knowledge learning, parallel learning.

[0034] Optionally, the acceptance criteria include: whether the learning system can apply the learned knowledge in the process of processing new tasks; step S5 includes:

[0035] Step S51: If there is a practice case, check whether the learned knowledge points are output after the to-be-developed intelligent agent runs the practice case. If the learned knowledge points are output, the debugging result indicates that the learning system meets the development requirements; if the learned knowledge points are not output, the debugging result indicates that the learning system does not meet the development requirements.

[0036] In a second aspect, the present invention provides an automatic development device for an intelligent agent learning system through computing power in an intelligent computing center, including:

[0037] A receiving module, configured to receive task information for constructing a to-be-developed intelligent agent learning system provided by a user through an interaction interface;

[0038] A requirements confirmation module, configured to determine the development requirements of the learning system in combination with the large language model method according to the task information and the user;

[0039] A design module, configured to determine the design scheme of the learning system in combination with the large language model method according to the development requirements;

[0040] An encoding module, configured to generate the code of the learning system in combination with the large language model method according to the development requirements and the design scheme;

[0041] A debugging module, configured to deploy the running environment of the learning system and debug the code of the learning system in combination with the large language model method; when the debugging result indicates that the learning system does not meet the development requirements, check the design scheme and code of the learning system in combination with the large language model method. When it is necessary to modify the design scheme, trigger the design module to continue working. When it is necessary to modify the code, trigger the encoding module to continue working until the debugging result indicates that the learning system meets the development requirements or the number of debugging or modification times exceeds a preset threshold;

[0042] A feedback module, configured to feedback result information to the user through the interaction interface.

[0043] In a third aspect, the present invention provides an electronic device, including: a processor, a memory, and a program stored on the memory and executable on the processor. When the program is executed by the processor, it implements the steps of the method for automatically developing an intelligent agent learning system through computing power in the intelligent computing center as described in the first aspect above.

[0044] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the method for automatically developing an intelligent agent learning system through computing power in the intelligent computing center as described in the first aspect above.

[0045] In a fifth aspect, the present invention provides a computer program product, including computer instructions. When the computer instructions are executed by a processor, it implements the steps of the method for automatically developing an intelligent agent learning system through computing power in the intelligent computing center as described in the first aspect above.

[0046] In the present invention, by developing an intelligent agent through a learning system running in an intelligent computing center, it can communicate with the user to confirm clear and detailed development requirements, automatically complete the development of the learning system of the intelligent agent to be developed, can automatically develop learning systems for different types of intelligent agents, has universality, and replaces the process of manually designing and developing the code of the learning system, thereby effectively saving development costs. Moreover, the intelligent computing center can provide sufficient computing power resources, greatly improving the development efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0048] Figure 1 is a schematic flow chart of the method for automatically developing an intelligent agent learning system through computing power in the intelligent computing center of the present invention;

[0049] Figure 2 is a schematic structural diagram of the device for automatically developing an intelligent agent learning system through computing power in the intelligent computing center of the present invention;

[0050] Figure 3 is a schematic structural diagram of the electronic device of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] Next, the technical solutions in the present invention will be clearly and completely described in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0052] First, the technical terms related to the present invention will be briefly described below.

[0053] The "computing power" referred to in the present invention means: the ability of a computer device or a computing / data center to process information, the ability of computer hardware and software to cooperate to jointly execute a certain computing requirement, the computing ability to achieve the output of a target result through processing information data, a new type of productive force integrating information computing power, network carrying capacity, and data storage capacity, and mainly providing services to society through computing power infrastructure.

[0054] The "computational power" (Computational Power, CP) referred to in the present invention means: the ability of a data center server to process data and achieve result output, a comprehensive indicator for measuring the computing ability of a data center, including general computing ability, supercomputing ability, and intelligent computing ability. The commonly used measurement unit is the number of floating-point operations per second (FLOPS, 1 EFLOPS = 10^18 FLOPS), and the larger the value, the stronger the comprehensive computing ability. It is estimated that 1 EFLOPS is approximately the computing power output of 5 Tianhe 2A or 500,000 mainstream server CPUs or 2 million mainstream laptops. The calculation formula is: CP = CP 通用 +CP 智能 +CP 超级 .

[0055] The "carrying capacity" (Network Power, NP) referred to in the present invention means: the manifestation of the data transmission ability of computing power facilities, a comprehensive ability including network architecture, network bandwidth, transmission delay, intelligent management and scheduling, etc., involving network transmission inside and between data centers, and a comprehensive indicator for measuring network transmission scheduling ability.

[0056] The "Storage Power" (SP) described in the present invention refers to the comprehensive ability of a data center in four aspects: data storage capacity, performance, security and reliability, and green and low-carbon. It is a comprehensive indicator for measuring the data storage ability of a data center, including external storage devices such as storage arrays and built-in storage devices of servers. The commonly used measurement unit for storage capacity is exabyte (EB, 1EB = 2^60 bytes), the commonly used measurement unit for performance is the number of read and write operations per second per unit capacity (IOPS / TB, Input / Output Operations Per Second / TB), and the disaster recovery ratio is an important manifestation of security and reliability.

[0057] The "computing power infrastructure" described in the present invention refers to a new type of information infrastructure that integrates information computing power, network carrying capacity, and data storage power, and can realize the centralized computing, storage, transmission, and application of information.

[0058] The "new type of information infrastructure" described in the present invention mainly includes network infrastructures such as 5G networks, fiber broadband networks, backbone networks, international communication networks, and satellite Internet, computing power infrastructures such as data centers, general computing power centers, intelligent computing centers, and supercomputing centers, and new technology facilities such as artificial intelligence, blockchain, and quantum computing.

[0059] The "computing power" described in the present invention includes general computing power, intelligent computing power, and super computing power.

[0060] The "general computing power" described in the present invention refers to the computing ability provided by servers based on CPU (Central Processing Unit) chips, which is used to support basic general computing such as cloud computing and edge computing.

[0061] The "intelligent computing power" described in the present invention refers to a computing platform that is deployed on a large scale for various artificial intelligence innovation applications based on dedicated chips such as GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), and ASIC (Application Specific Integrated Circuit), such as natural language processing and machine vision.

[0062] The "super computing power" described in the present invention mainly refers to the computing ability provided by high-performance computing clusters such as supercomputers. It utilizes the centralized computing resources of multiple computer systems working in parallel and processes extremely complex or data-intensive problems through a dedicated operating system. It is mainly used for computing in cutting-edge scientific fields, such as planetary simulation, drug molecule design, and gene analysis.

[0063] The "Intelligent Computing Center" described in the present invention refers to a facility that uses large-scale heterogeneous computing power resources, including general computing power (CPU) and intelligent computing power (GPU, FPGA, ASIC, etc.), and mainly provides the required computing power, data, and algorithms for artificial intelligence applications (such as scenarios like artificial intelligence deep learning model development, model training, and model inference). The intelligent computing center covers facilities, hardware, and software, and can provide full-stack capabilities from underlying computing power to top-level application enabling.

[0064] The "Intelligent Computing Center" described in the present invention includes, but is not limited to, the "Intelligent Computing Center".

[0065] The "Intelligent Computing Center" described in the present invention, namely the artificial intelligence computing center, is a type of computing power infrastructure based on artificial intelligence theory, adopting an artificial intelligence computing architecture, and providing computing power services, data services, and algorithm services required for artificial intelligence applications.

[0066] The "Computing Power Center" described in the present invention refers to a facility mainly composed of infrastructure such as wind, fire, water, and electricity, and IT software and hardware devices, and having computing power, carrying capacity, and storage capacity, including general data centers, intelligent computing centers, supercomputing centers, etc.

[0067] The "Supercomputing Center" described in the present invention refers to, namely the supercomputing data center, which is a data center based on supercomputers or large-scale computing clusters, and can provide functions such as large-scale computing, storage, and network services, and is widely used in application scenarios such as aerospace, national defense, oil exploration, climate modeling, and genome sequencing.

[0068] The "Computing Power Resources" described in the present invention refers to technologies and facilities with information computing, transmission, storage, and application capabilities required for the development of the digital society, including but not limited to computing resources such as CPU and GPU, network resources such as switches and routers, storage resources such as storage arrays and distributed storage, security resources such as firewalls and intrusion detection systems, and support and guarantee resources such as wind, fire, water, and electricity.

[0069] The "Large Language Model" described in the present invention refers to the large language model (LLM), which is a language model with a relatively large number of parameters, aiming to understand and generate human language, trained through a large amount of text data, and can perform a wide range of tasks including text summarization, translation, sentiment analysis, etc.

[0070] The "Agent" described in the present invention refers to an agent that can perceive the environment and take actions to achieve specific goals, and is an important part of the intelligent computing center. It can be software, hardware, or a system, and has autonomy, adaptability, and interaction capabilities. The agent perceives changes in the environment (such as through sensors or data input), makes judgments and decisions based on the knowledge and algorithms it has learned, and then executes actions to affect the environment or achieve the predetermined goals. Agents are widely used in the field of artificial intelligence and are commonly found in automation systems, robots, virtual assistants, and game characters. The core lies in the ability to learn independently and evolve continuously to better complete tasks and adapt to complex environments.

[0071] The "method of combining large language models" described in the present invention refers to organizing input information and designing precise prompt words to call large language models to generate various content forms, including thinking, responses, decisions, tool selection, code, and documents. This method supports tasks such as reasoning and thinking, design, conversation with users, generation of solution designs, and code generation by agents, and combines with code frameworks to execute instructions such as tool method calls and code program executions, achieving the completion of specific tasks such as document retrieval, information search, file processing, and script running.

[0072] To solve the problems that existing agents lack learning ability and the cost of manually developing the learning system of agents is high and the efficiency is low, please refer to Figure 1 , the present invention provides a method for automatically developing an agent learning system by computing power in an intelligent computing center. This method can also be called a method for the intelligent computing center to develop an agent learning system for a to-be-developed agent by computing power, which is executed by the learning system development agent. The method includes:

[0073] Step S1: Receive task information provided by the user for constructing the learning system of the to-be-developed agent through an interaction interface;

[0074] The "learning system development agent" described in the present invention has the ability to develop a learning system for other agents by combining the method of large language models.

[0075] In the present invention, the learning system development agent can be an agent dedicated to developing the learning system of the to-be-developed agent, or a development agent for developing the entire to-be-developed agent.

[0076] The learning system development agent can be one agent or multiple agents. The multiple agents form a multi-agent cooperation system. Each agent has specific capabilities, and the multiple agents can cooperate to complete the task of developing a learning system for the to-be-developed agent.

[0077] In the present invention, the agent to be developed may be an agent for completing a specified task, such as a mathematical processing agent, a data analysis agent, etc.

[0078] In the present invention, the task information may include at least one of the following information: the development requirements of the learning system, the framework code of the agent to be developed or the storage path of the code, the storage path of the relevant documents of the framework of the agent to be developed, etc.

[0079] Step S2: The learning system development agent determines the development requirements of the learning system according to the task information, in combination with the large language model method and the user;

[0080] Optionally, the development requirements may be to develop a new agent and develop a learning system for the new agent, or to develop a learning system for an existing agent or upgrade the learning system of an existing agent.

[0081] Optionally, the development requirements of the learning system include at least one of the following: the input method of learning materials, whether to prompt the user to provide learning key points during learning, whether to require the user to provide knowledge tags during learning, the learning capabilities required by the learning system, the learning depth, whether to provide the ability to visually access the learned knowledge, the application link of the learning system in the agent to be developed, whether to provide practice cases, test cases, acceptance criteria, output requirements, technical requirements, performance requirements, security and privacy.

[0082] The above development requirements are explained separately below.

[0083] The input method of learning materials refers to the way of providing the materials to be learned to the learning system. For example, it can be provided in the form of a file, or provided in a dialogue way on the interaction interface.

[0084] Whether to prompt the user to provide learning key points during learning. For example, when providing the code to be learned to the learning system, the learning system can be prompted to mainly learn which functions in the code.

[0085] The learning capabilities may include at least one of the following: knowledge learning, knowledge archiving, knowledge extraction, knowledge matching, knowledge application, knowledge practice, knowledge display, knowledge coding, knowledge maintenance, knowledge version maintenance, multi-document knowledge learning, large-document knowledge learning, parallel learning.

[0086] Learning depth: refers to learning capabilities at different depths, which may include at least one of the following:

[0087] 1. Simple archiving

[0088] Definition: Preserve knowledge or information in its original form without processing, solely for storage and backup purposes.

[0089] Implementation method: The large language model method can be used to store knowledge in the form of vector similarity. In the future, knowledge can be matched and extracted based on the similarity of tasks or content.

[0090] 2. Summary

[0091] Definition: Simplify and compress the learning content, extract key information and reorganize it to form a more refined version.

[0092] Implementation method: Summarize the original content, and while saving a simple archive, also save the summary knowledge. The large language model method can be used to store the summary knowledge in the form of vector similarity and establish an association with the simple archive. In the future, priority matching can be performed according to the similarity of tasks or content. If a match is found and it is considered necessary to expand the detailed knowledge, the simple archive knowledge can be cited; if no match is found, knowledge matching can be carried out.

[0093] 3. Concept Summary

[0094] Definition: Extract the core concepts from the information and summarize them into concepts. For example, interpret the hierarchical structure of department labels from a document and record it in the form of a document. When matching knowledge, concept-level matching is preferred. If a match is found, the summary or simple archive knowledge can be extracted as needed; if no concept summary knowledge is matched, a lower learning depth can be used to attempt knowledge extraction.

[0095] 4. Methodology Summary

[0096] Definition: Extract generalizable solution strategies from specific experiences to form reusable operation processes or decision-making principles.

[0097] 5. Draw inferences about other cases from one instance

[0098] Definition: When learning knowledge, it is possible to discover whether there is similar knowledge content in the provided learning materials and learn and archive it together. For example, when the user requests to learn the department coding structure and discovers the product coding structure in the code, it can be learned together and labeled as learning by drawing inferences about other cases from one instance. Ultimately, the user needs to confirm the knowledge.

[0099] The application link of the learning system in the to-be-developed intelligent agent refers to the position of the learning system in the entire workflow of the to-be-developed intelligent agent.

[0100] Whether to provide practice cases means whether to provide practice cases for the learning system to verify whether the learning system can learn knowledge.

[0101] A test case refers to an example used to test whether a learning system meets the acceptance criteria.

[0102] The acceptance criteria are the basis for evaluating whether a learning system meets the development requirements, and may include at least one of the following: whether the function implementation meets expectations (e.g., whether it can apply the learned knowledge in the process of handling new tasks), whether the performance indicators (such as speed, accuracy, resource consumption) meet the standards, whether the technical implementation meets the compatibility and environmental requirements, whether the security and privacy protection comply with regulations, whether the output is clear and intuitive, whether the application effect meets the scenario requirements, and whether necessary documentation and technical support are provided.

[0103] The output requirements refer to the requirements for the content or format of the output of the learning system, which are convenient for understanding and analysis.

[0104] The technical requirements refer to the specific requirements for the technical implementation of the learning system, such as algorithm support, computing power requirements, compatible development environments, etc. For example, whether the learning system needs to support multi-modal data processing (such as text, images, audio, etc.), whether it needs to be integrated with existing systems, etc. The clarification of technical requirements helps to ensure the feasibility and applicability of the learning system.

[0105] The performance requirements refer to the specific requirements for the performance of the learning system, including inference speed, accuracy, resource consumption, scalability, etc.

[0106] Security and privacy refer to the ability of the learning system to protect the security and privacy of data during use. For example, whether the learning system can prevent the leakage of sensitive data, whether it supports data encryption, and whether it complies with relevant laws and regulations (such as GDPR, CCPA, etc.). Security and privacy protection are important requirements that cannot be ignored in the actual application of the framework, especially in scenarios involving user data or sensitive information.

[0107] The above learning capabilities are described below.

[0108] Knowledge learning refers to the use of large language model methods to perform learning at different depths on the knowledge provided by the user in the conversation and / or document knowledge. The documents include ordinary text, PPT, CSV, Excel, code files, PDF, HTML, mind map documents, e-book documents, and even multi-modal images, videos, audio and other documents.

[0109] Knowledge archiving refers to storing knowledge using different storage engines according to different depths of learning. The storage engines include at least one of the following: vector database, graph database, relational database, document storage, etc.

[0110] Knowledge extraction refers to the way of extracting knowledge when solving problems. It supports knowledge extraction by keywords and / or knowledge extraction based on vector similarity.

[0111] Knowledge matching refers to matching the extracted knowledge with the current problem or task, screening and confirming whether the knowledge meets the steps of solving the current problem or task. Optionally, it supports a knowledge extraction mode with decreasing depth, deleting unnecessary knowledge, retaining available knowledge, and identifying whether lower-level knowledge needs to be extracted. If so, perform knowledge extraction and knowledge matching until the knowledge for the current task is obtained.

[0112] Knowledge application refers to embedding the ability to apply knowledge in some modules of the intelligent agent to be developed.

[0113] Knowledge practice means that learning cases can be provided to verify whether the learning ability meets the standard. Optionally, when practicing knowledge, check whether newly learned knowledge is used when solving problems. If the newly learned knowledge is not used in the output when solving new problems, the requirement is not met, and learning, archiving, and then practicing need to be done again until knowledge can be extracted from the practice cases. If the user does not provide practice cases, there is no need to practice.

[0114] Knowledge display means providing different display methods according to the depth of learning for the user to confirm.

[0115] Among them, the display methods can include at least one of the following:

[0116] Simple archiving: Display tags, display archived files, display partial documents

[0117] Summary: On the basis of the display of simple archiving, display summary knowledge

[0118] Concept summary: On the basis of the summary, display the summarized conceptual knowledge

[0119] Methodology summary: On the basis of the concept summary, display the methodology summary knowledge

[0120] Draw inferences about other cases from one instance: On the basis of the methodology summary knowledge, display the knowledge learned by drawing inferences about other cases

[0121] Knowledge coding provides a unique code for each learned knowledge for easy management.

[0122] Knowledge maintenance refers to providing capabilities such as knowledge viewing, knowledge editing, knowledge conflict resolution, knowledge deletion, and knowledge modification.

[0123] Among them:

[0124] 1. Knowledge viewing means that the user can ask the to-be-developed intelligent agent to consult knowledge in a certain aspect.

[0125] 2. Knowledge editing means that the user can, by means of description, ask the to-be-developed intelligent agent to modify the learned knowledge according to the knowledge coding.

[0126] 3. Knowledge deletion means that the user can ask the to-be-developed intelligent agent to delete knowledge in a certain aspect according to the knowledge coding.

[0127] 4. Knowledge conflict resolution means that when the to-be-developed intelligent agent saves new knowledge, it needs to verify whether the existing knowledge conflicts with the new knowledge. If there is a conflict, the new knowledge needs to be shown to the user, and the user can directly modify the knowledge or inform the conflict resolution method.

[0128] Knowledge version means that the historical versions of knowledge can be saved, and the user can view the historical versions of knowledge and restore the knowledge of historical versions to avoid the deletion of knowledge caused by misoperation instructions.

[0129] Multi-document knowledge learning means supporting the learning of knowledge from multiple documents and learning document by document through the large language model method. Optionally, it supports screening necessary learning materials for vectorized archival storage. When the total size of the saved documents exceeds the threshold, screening of documents for storage is triggered. When screening and saving knowledge, the original documents do not need to be vectorized for storage. The archived knowledge records the archival paths of the original document materials.

[0130] Large-document knowledge learning means supporting the segmented learning of knowledge through the large language model method. Optionally, it supports screening necessary fragment contents and reorganizing them into knowledge documents for archival. When the total size of the saved documents exceeds the threshold, screening of fragments for storage is triggered. When screening and saving fragments, the original documents do not need to be vectorized for storage. The archived knowledge records the archival paths of the original document materials.

[0131] Parallel learning means that during multi-document knowledge learning and large-document learning, the computing power of the computing center is used to perform parallel document-by-document and segment-by-segment knowledge learning, and then the knowledge is merged to improve the overall learning efficiency.

[0132] In the present invention, the combination of the large language model method and the user to determine the development requirements of the learning system may be: the learning system development intelligent agent constructs prompt words and invokes the large language model to generate replies or questions, etc., to complete the task of the learning system development intelligent agent having a conversation with the user, and obtains the development requirements according to the conversation content.

[0133] Step S3: The learning system development intelligent agent determines the design scheme of the learning system according to the development requirements, in combination with the large language model method.

[0134] The design solution in the present invention may include at least one of the following:

[0135] Modify the task processing module of the to-be-developed intelligent agent so that the task processing module can identify whether the user task is a learning task;

[0136] Modify the upload module of the to-be-developed intelligent agent to support the upload and archiving of learning materials;

[0137] Add a learning system;

[0138] Modify the task processing module in the to-be-developed intelligent agent so that the task processing module can call the knowledge extraction module of the learning system and integrate the knowledge learned by the learning system into the prompt words of the task processing module;

[0139] Add a practice module;

[0140] General design document, detailed design document, interface design document, and / or database design document.

[0141] Among them, the detailed design document refers to a document that describes in detail the functional implementation of each module in the learning system, including the logical process of the module, algorithm design, data processing method, exception handling mechanism, and other contents.

[0142] The interface design document refers to a document that describes the interfaces between each module in the learning system or the interfaces for interacting with external systems, including the functional definition of the interface, input and output parameters, call method, data format, communication protocol, and error code description, etc.

[0143] The database design document refers to a document that designs the data storage structure involved in the learning system, including the table structure of the database, field definition, data type, primary key and foreign key relationship, index design, stored procedure, and trigger, etc.

[0144] In the present invention, the learning system development intelligent agent determines the design solution of the learning system according to the development requirements and in combination with the large language model method. Specifically, the learning system development intelligent agent constructs prompt words according to the development requirements and calls the large language model to complete the task of generating the design solution of the learning system.

[0145] Step S4: Generate the code of the learning system according to the development requirements and design solution, in combination with the large language model method;

[0146] In the present invention, the learning system development agent generates the code of the learning system according to the development requirements and design scheme, specifically: the learning system development agent constructs prompt words and invokes the large language model according to the development requirements and design scheme to complete the task of generating the code of the learning system.

[0147] Step S5: Deploy the operating environment of the learning system and debug the code of the learning system in combination with the large language model method; when the debugging result indicates that the learning system does not meet the development requirements, check the design scheme and code of the learning system in combination with the large language model method. When it is necessary to modify the design scheme, return to step S3. When it is necessary to modify the code, return to step S4 until the debugging result indicates that the learning system meets the development requirements or the number of debugging or modification times exceeds the preset threshold.

[0148] The deployment of the operating environment of the learning system and the debugging of the code of the learning system in combination with the large language model method can be specifically: the learning system development agent constructs prompt words and invokes the large language model according to the development requirements, design scheme and the code to complete the tasks of operating environment deployment and code debugging.

[0149] The learning system development agent checks the design scheme and code of the learning system in combination with the large language model method, specifically: the learning system development agent constructs prompt words and invokes the large language model according to the design scheme, code and debugging result of the learning system to complete the task of checking the design scheme and code. That is, the learning system development agent inputs the prompt words into the large language model and obtains the inspection result output by the large language model. The inspection result can indicate errors or defects in the design scheme or code, thus supporting the optimization and improvement of the learning system.

[0150] It should be noted that in the case of returning to step S3 if the design scheme needs to be modified, in the new step S3, the learning system development agent needs to re-determine the design scheme of the learning system in combination with the large language model method according to the development requirements, the design scheme and the inspection result.

[0151] It should be noted that in the case of returning to step S4 if the code needs to be modified, in the new step S4, the learning system development agent regenerates the code of the learning system in combination with the large language model method according to the development requirements, the code and the inspection result.

[0152] In the present invention, the test cases, acceptance criteria, performance requirements, etc. in the above development requirements can be used to determine whether the learning system meets the development requirements.

[0153] Step S6: Feedback result information to the user through the interaction interface.

[0154] Among them, the feedback result information may include at least one of the following: information about the code of the learning system, debugging information of the learning system, analysis of the applicable scenarios of the learning system, optimization suggestions for the learning system, and conclusions on the success or failure of the development of the learning system.

[0155] Among them, the conclusion of development failure may be a conclusion after multiple debuggings and the number of debuggings exceeds the failure threshold.

[0156] The information about the code of the learning system may be the code itself or the storage location of the code, etc.

[0157] The debugging information may include at least one of the following: running logs, intermediate data, debugging results, etc.

[0158] In the present invention, the learning system agent relies on the computing power of the intelligent computing center and can simultaneously develop learning systems for a large number of agents to be developed.

[0159] In the present invention, by developing an agent through the learning system running in the intelligent computing center, it can communicate with the user to confirm clear and detailed development requirements, automatically complete the development of the learning system of the agent to be developed, can automatically develop learning systems for different types of agents, has universality, and replaces the process of manually designing and developing the code of the learning system, thereby effectively saving development costs, and the intelligent computing center can provide sufficient computing power resources, greatly improving the development efficiency.

[0160] In the present invention, optionally, the step S2 includes:

[0161] Step S21: Based on the task information, communicate with the user through the interaction interface at least once based on the large language model;

[0162] After the learning system development agent receives the task information sent by the user through the interaction interface, it can construct a prompt word according to the task information. The prompt word is used to prompt the large language model to analyze whether the task information contains development requirements, whether the development requirements are complete, and whether it is necessary to confirm more detailed development requirements with the user. The learning system development agent inputs the prompt word into the large language model and obtains the analysis result output by the large language model. The analysis result includes the content of the development requirements to be determined. The learning system development agent displays the content of the development requirements to be determined to the user based on the interaction interface.

[0163] When the user provides new development requirements through the interaction interface, the learning system development agent can construct a prompt according to the new development requirements. The prompt is used to prompt the large language model to analyze whether the development requirements are complete and whether it is necessary to confirm more detailed development requirements with the user. The learning system development agent inputs the prompt into the large language model and obtains the analysis result output by the large language model. The analysis result includes the content of the development requirements to be determined. The learning system development agent displays the content of the development requirements to be determined to the user based on the interaction interface.

[0164] The process of determining development requirements by combining the large language model method above can be carried out in multiple rounds to obtain more detailed development requirements.

[0165] Step S22: Determine the development requirements of the learning system according to the task information and the communication content with the user, by combining the large language model method.

[0166] In the present invention, the development agent communicates with the user for at least one round to confirm the detailed development requirements of the learning system, ensuring a comprehensive and accurate understanding of the functions, performance, compatibility, etc. of the learning system expected by the user, and laying a solid foundation for subsequent design and development work.

[0167] In the present invention, optionally, the step S3 includes:

[0168] Step S31: According to the development requirements of the learning system, adopt at least one of the following learning methods by combining the large language model method to obtain learning results: search for relevant materials through a search engine, view and learn open source code, learn the existing code and documents of the agent to be developed, and consult data;

[0169] Step S32: Determine the design scheme of the learning system according to the learning results.

[0170] During the process of sorting out and understanding the development requirements, the learning system development agent sorts out unclear or unknown concepts and knowledge points, and conducts targeted searches for these contents. Obtain relevant information through a search engine, screen and classify the search results, and extract key contents to ensure a comprehensive understanding of the requirements, providing necessary knowledge support for the development of the learning system of the agent.

[0171] The learning method is indicated to the learning system development agent by the large language model according to the development requirements, and the learning system development agent obtains the learning results.

[0172] Specifically, the learning system development agent can generate a prompt word according to the development requirements and the learning results. The prompt word is used to prompt the large language model to determine the design scheme of the learning system according to the development requirements and the learning results. The learning system development agent inputs the prompt word into the large language model and obtains the design scheme output by the large language model.

[0173] Through the above learning method, the learning system can be deeply understood, and the learning results will be used as important references to provide diverse ideas and methods for the design of the learning system, so as to design a learning system that better meets the requirements and is more complete.

[0174] If the learning system development agent finds during the debugging process that the current design of the learning system cannot meet the development requirements and it belongs to a design defect after analysis, the existing design can be overturned or optimized.

[0175] In the present invention, optionally, the step S3 includes:

[0176] Step S33: If the design of the learning system does not meet the development requirements, at least one of the following adjustment methods is adopted in combination with the large language model method to adjust the design scheme of the learning system:

[0177] The first adjustment method is to optimize the design scheme of the learning system to better meet the development requirements;

[0178] The second adjustment method is to adopt other design schemes, that is, to explore and try a completely new design of the learning system to cope with special requirements or complex scenarios.

[0179] In the present invention, optionally, the acceptance criteria include: whether the learning system can apply the learned knowledge in the process of processing new tasks; the step S5 includes:

[0180] Step S51: If there is a practice case, check whether the learned knowledge points are output after the to-be-developed agent runs the practice case. If the learned knowledge points are output, the debugging result indicates that the learning system meets the development requirements; if the learned knowledge points are not output, the debugging result indicates that the learning system does not meet the development requirements.

[0181] In the present invention, optionally, the step S5 includes at least one of the following sub-steps:

[0182] Step S52: When the number of debugging attempts reaches the first preset threshold, or the number of modifications to the design scheme reaches the second preset threshold, or the number of modifications to the code of the learning system reaches the third preset threshold, and the code of the learning system still fails to meet the development requirements, terminate the development task of the learning system, and feedback the termination reason and related situations to the user through the interaction interface to help the user re-evaluate the requirements and adjust the development strategy to more efficiently promote the subsequent work.

[0183] This step can avoid excessive consumption of computing power resources in the intelligent computing center.

[0184] Step S53: During the debugging process, record the running logs, intermediate data, and debugging results in combination with the large language model method, and archive and save them.

[0185] During the development and debugging of the learning system, key information such as running logs, intermediate data, and test results can be recorded and properly archived and saved. For error information and related code, key records and sorting will be carried out to facilitate in-depth analysis of the root cause of problems, summarize improvement experience in the future, and provide important reference for the optimization and development of the learning system.

[0186] As mentioned above, the learning system development agent can include multiple agents. In the present invention, optionally, the multiple agents include at least one of the following agents: requirement confirmation agent, framework design agent, search agent, learning agent, coding agent, debugging agent, archiving agent, and deployment agent, etc.

[0187] The requirement confirmation agent is responsible for interacting with the user, accurately understanding and clarifying the development requirements, and converting them into executable design documents;

[0188] The framework design agent is responsible for designing the overall architecture and functional modules of the learning system of the agent to be developed;

[0189] The search agent is responsible for finding relevant technical materials and best practices to provide reference for the design;

[0190] The learning agent summarizes the materials found and extracts key technologies and implementation methods;

[0191] The coding agent converts the design scheme into code implementation and develops the core functional modules;

[0192] The debugging agent is responsible for testing and fixing problems and recording error information during the debugging process;

[0193] The archiving agent archives and manages the documents, code, and logs during the development process to ensure the traceability of the materials;

[0194] The described deployment agent is responsible for deploying the developed agent to the target running environment and monitoring the running status.

[0195] These agents work together to ensure the completion of the design, implementation, and optimization of the learning system development agent for the learning system, in order to meet the requirements of the user for the development of the agent to be developed in the learning system.

[0196] The following is an example of "determining the development requirements of the learning system in combination with the large language model method and the user according to the task information" in step S2 below in combination with a specific application scenario.

[0197] In this embodiment, the user needs the learning system development agent to develop a learning system for a data processing agent (i.e., the agent to be developed).

[0198] On the interaction interface, the conversation between the user and the learning system development agent can be as follows:

[0199] **User**: I want to add learning capabilities to the data processing agent so that I can provide it with the code developed by ourselves for it to learn, master the data processing methods developed by ourselves, and apply them to future data processing problems.

[0200] **Learning System Development Agent**: Wait a moment. Let me first check the project documents and code of the data processing agent. <Check the existing documents and code of the data processing agent project through the large language model method>

[0201] **Learning System Development Agent**: The current data processing agent is an AI agent based on the large language model, communicates with the user about tasks through a browser page, and supports file uploads. Its workflow is: receive the user's data processing instructions, explore the data, generate data processing code, execute the data processing code, output the results and feedback to the user. It itself does not have learning capabilities and needs to add a learning system and apply the extracted knowledge to the original data processing process.

[0202] The following points need your confirmation:

[0203] Input method of learning materials: Whether the code to be learned is provided in the form of a file or in the form of a dialogue input.

[0204] Learning key points prompt: Whether it is mandatory to require the user to give learning key points when learning, such as telling the data processing agent which function in the code to mainly learn while providing the code.

[0205] Knowledge label prompt: Whether it is required to provide knowledge labels together when learning, and refer to the necessary knowledge labels during knowledge archiving and extraction.

[0206] Learning Depth: Which of the following learning depths needs to be selected:

[0207] 1) Simple Archiving: Save knowledge or information in its original form without processing, only for storage and backup.

[0208] 2) Summary: Simplify and compress the learning content, extract key information and reorganize it to form a more refined version.

[0209] 3) Concept Summary: Extract core concepts from the information and summarize them into abstract expressions, and prefer to match knowledge through the concept layer.

[0210] 4) Methodology Summary: Extract general problem-solving strategies from specific experiences to form reusable operation processes or decision-making principles.

[0211] 5) Draw inferences from one instance: Actively discover similar content related to the current knowledge during the learning process, expand learning synchronously and mark the relevance.

[0212] Output of Learning Results: Whether it is necessary to provide the ability to visually view the learning results so that users can understand the reliability of the learned knowledge.

[0213] Application of Learned Knowledge: Whether it is mainly used in the processes of exploring data and generating data processing code.

[0214] Practice Mechanism: Whether practice cases can be provided, including learning materials and cases to be practiced (Case).

[0215] **User**: My specific expectations are as follows:

[0216] Input Method of Learning Materials: Priority is given to supporting file upload (Python / Jupyter Notebook) and specifying the learning file path, and it is also allowed to input code snippets through the dialogue interface.

[0217] Learning Key Hints: The agent needs to automatically analyze the code structure, but the user is allowed to manually specify key functions (such as #LEARNING_CORE) through comments or additional description files.

[0218] Knowledge Tag Hints: Automatically extract technical tags (such as data cleaning / feature engineering) and allow users to supplement business tags (such as financial risk control / medical imaging).

[0219] Learning Depth: Adopt concept summary.

[0220] Visualization Requirement: Yes, the learned knowledge should be viewable.

[0221] Application Scenario: Focus on enhancing the processes of exploring data and generating data processing code.

[0222] Practice mechanism: It needs to be provided.

[0223] **Learning system to develop an agent**: The requirements have been recorded, and the preliminary design plan is summarized as follows:

[0224] Based on the existing code structure, modify and add the following modules:

[0225] 1. Modify the task recognition module to be able to recognize whether it is a learning task or a data processing task.

[0226] 2. Modify the upload module to support the upload and archiving of learning materials.

[0227] 3. Add a learning system that supports material loading, knowledge learning, knowledge archiving, knowledge display, and knowledge extraction.

[0228] 4. Modify the generated data processing code module to call the knowledge extraction module of the learning system and integrate the matched knowledge into the prompt words generated by the data processing code.

[0229] 5. Add a practice module that supports case running, case analysis, etc.

[0230] Knowledge storage architecture support:

[0231] Vector database, supporting RAG retrieval augmented generation.

[0232] Graph database, supporting knowledge associated storage and extraction.

[0233] Document storage, supporting the storage of original documents.

[0234] Are there any other requirements?

[0235] **User**: None.

[0236] **Learning system to develop an agent**: Can you provide the code file of the practice case, learning key points, knowledge tags, learning depth, and the specific content of the practice case?

[0237] **User**:

[0238] The code file of the practice case is at / mnt / share / codes / data_process / Department.py.

[0239] The learning key point is to master the department coding format therein. When encountering similar coding in the future, it is necessary to know that this is the department coding and be able to parse the department coding at different levels according to the analysis requirements.

[0240] The knowledge tags are department coding and organizational structure.

[0241] Learning Depth: To concept summary.

[0242] Practice Case: Please count the cost expenditures for different departments in the first quarter of 2025 for the classification purpose according to the first-level departments in the dataset / mnt / share / datas / 2025 / finance / cost_detail.csv.

[0243] **Learning System Development Agent**:

[0244] Learning Code File: Confirm

[0245] Learning Key Points: Confirm

[0246] Knowledge Tags: Confirm

[0247] Learning Depth: Confirm

[0248] Practice Case: Confirm

[0249] There is no problem, and the development will start later.

[0250] In the present invention, the agent to be developed before the learning system development does not have learning ability, and the agent to be developed for processing after the learning system development has learning ability. Before executing a task, it can first execute a learning task, learn knowledge points, and can verify whether the learning goal is achieved according to the practice case provided by the user.

[0251] The following is an example to illustrate the working processes of the data processing agent before the learning system development and the data processing agent after the learning system development.

[0252] Before the data processing agent has learning ability:

[0253] **User**: Please count the revenues of different departments in the first quarter of 2025 for the classification purpose according to the first-level departments in the dataset / mnt / share / datas / 2025 / finance / income_detail.csv.

[0254] **Data Processing Agent**:

[0255] <The data processing agent receives the user's data processing instruction, explores the data, generates data processing code, and executes the data processing code>

[0256] The cost expenditure data has been processed and is placed in / mnt / share / dataprocess / income_department_lev_1_2025_q1.csv.

[0257] **Explanation**:

[0258] The result does not meet the user's requirements. There is no original data in the dataset classified by the first-level department field. There is only the department code, and the statistical result is based on the department code.

[0259] After the data processing agent has learning ability:

[0260] I. Learning stage

[0261] **User**: Please learn the / mnt / share / codes / data_process / Department.py file to understand the department code format. In the future, when encountering similar codes, it should be able to recognize that this is a department code and parse out different levels of department codes according to the analysis requirements.

[0262] **Data processing agent**:

[0263] <Identified as a learning task using the large language model method>

[0264] Received. Please provide the following information:

[0265] Learning file: / mnt / share / codes / data_process / Department.py

[0266] Learning focus: Understand the department code format. In the future, when encountering similar codes, be able to recognize them as department codes and parse out different levels of department codes according to the analysis requirements.

[0267] Knowledge tags: <To be provided>

[0268] Practice cases: <To be provided>

[0269] **User**:

[0270] The knowledge tags are department code, first-level department, second-level department, and organizational structure. They may be used when such keywords are mentioned.

[0271] Practice cases: For the dataset / mnt / share / datas / 2025 / finance / cost_detail.csv, count the cost expenditures for different departments in the first quarter of 2025 classified by the first-level department.

[0272] **Data processing agent**:

[0273] Received. Details are as follows:

[0274] Learning file: / mnt / share / codes / data_process / Department.py.

[0275] Learning Focus: Understand the department coding format. When encountering similar codes in the future, be able to recognize them as department codes and parse department codes at different levels according to the analysis requirements.

[0276] Knowledge Tags: The knowledge tags are "department coding", "department coding parsing rules", "primary department", "secondary department", "organization structure", which may be used when such keywords are mentioned.

[0277] Practice Case: For the dataset / mnt / share / datas / 2025 / finance / cost_detail.csv, count the cost expenditures for different departments in the first quarter of 2025 by primary department for the purpose of classification.

[0278] Please confirm.

[0279] **User**: Confirm.

[0280] **Data Processing Agent**:

[0281] <Using large language model method for data loading, knowledge learning, and knowledge archiving>

[0282] Knowledge learning and knowledge archiving have been completed. There is no similar knowledge in the historical knowledge, so it belongs to new knowledge and there is no knowledge conflict.

[0283] <Execute the practice case

[0284] Simulated Input: For the dataset / mnt / share / datas / 2025 / finance / cost_detail.csv, count the cost expenditures for different departments in the first quarter of 2025 by primary department for the purpose of classification.

[0285] Use the large language model method to extract knowledge & explore data, extract knowledge & generate data processing code, execute the data processing code, and print the processing results.

[0286] Iteratively check and learn to ensure whether the processing results contain new reference knowledge.

[0287] >

[0288] The practice has been completed. The cost expenditure data has been processed and is placed in / mnt / share / dataprocess / cost_department_lev_1_2025_q1.csv.

[0289] Among them, the reference knowledge when exploring data and generating code:

[0290] {

[0291] "k_name": "department coding parsing rules",

[0292] "k_labels": ["Department Code", "Department Code Parsing Rule", "First-level Department", "Second-level Department", "Organization Structure"],

[0293] "k_id": "concept_c5f3a7d0-8e1a-4f7b-9c2d-1b6e8f3a5d4c",

[0294] "k_depth_level": "Concept",

[0295] "k_associated": ["simple_655d9c2c-497a-4d07-8c29-865f4999c558", "summary_936c7f9e-8c2b-4732-825a-796c3e2b483f"]

[0296] }

[0297] **Data Processing Agent**:

[0298] The learning task has been completed. The learned knowledge is as follows:

[0299] {

[0300] "k_name": "Department Code Parsing Rule",

[0301] "k_labels": ["Department Code", "Department Code Parsing Rule", "First-level Department", "Second-level Department", "Organization Structure"],

[0302] "k_id": "concept_c5f3a7d0-8e1a-4f7b-9c2d-1b6e8f3a5d4c",

[0303] "k_depth_level": "Concept",

[0304] "k_associated": ["simple_655d9c2c-497a-4d07-8c29-865f4999c558", "summary_936c7f9e-8c2b-4732-825a-796c3e2b483f"]

[0305] }

[0306] 2) New Task Phase

[0307] **User**: For the dataset / mnt / share / datas / 2025 / finance / income_detail.csv, count the revenue by classification purpose for different departments in the first quarter of 2025 according to the first-level department.

[0308] **Data Processing Agent**:

[0309] <The data processing agent uses the large language model method to receive the user's data processing instructions, extract knowledge & explore data, extract knowledge & generate data processing code, and execute the data processing code>

[0310] **Data Processing Agent**:

[0311] The cost expenditure data has been processed and is placed in / mnt / share / dataprocess / income_department_lev_1_2025_q2.csv

[0312] Among them, the knowledge referred to during data exploration and code generation:

[0313] {

[0314] "k_name": "Department Code Parsing Rule",

[0315] "k_labels": ["Department Code", "Department Code Parsing Rule", "First-Level Department", "Second-Level Department", "Organization Structure"],

[0316] "k_id": "concept_c5f3a7d0-8e1a-4f7b-9c2d-1b6e8f3a5d4c",

[0317] "k_depth_level": "Concept",

[0318] "k_associated": ["simple_655d9c2c-497a-4d07-8c29-865f4999c558", "summary_936c7f9e-8c2b-4732-825a-796c3e2b483f"]

[0319] }

[0320] Note: The original dataset did not have the first-level department field. The data analysis agent enhanced the data exploration by extracting the learned knowledge. Only the department code was used, and the department code processing in the generated data processing code was used to break down the first-level department and perform statistical analysis, and the results met the user's requirements.

[0321] Please refer to Figure 2, the present invention also provides an automatic development device 10 for an intelligent agent learning system through computing power in an intelligent computing center, including:

[0322] A receiving module 11, configured to receive task information for constructing a to-be-developed intelligent agent learning system provided by a user through an interaction interface;

[0323] A requirement confirmation module 12, configured to determine the development requirements of the learning system with the user according to the task information by combining the large language model method;

[0324] A design module 13, configured to determine the design scheme of the learning system according to the development requirements by combining the large language model method;

[0325] A coding module 14, configured to generate the code of the learning system according to the development requirements and design scheme by combining the large language model method;

[0326] A debugging module 15, configured to deploy the running environment of the learning system and debug the code of the learning system by combining the large language model method; when the debugging result indicates that the learning system does not meet the development requirements, check the design scheme and code of the learning system by combining the large language model method, trigger the design module to continue working when the design scheme needs to be modified, and trigger the coding module to continue working when the code needs to be modified, until the debugging result indicates that the learning system meets the development requirements or the number of debugging or modification times exceeds a preset threshold;

[0327] A feedback module 16, configured to feedback result information to the user through the interaction interface.

[0328] In the present invention, by developing an intelligent agent through a learning system running in an intelligent computing center, it can communicate with the user and confirm clear and detailed development requirements, automatically complete the development of the learning system of the to-be-developed intelligent agent, can automatically develop learning systems for different types of intelligent agents, has universality, and replaces the process of manually designing and developing the code of the learning system, thereby effectively saving development costs, and the intelligent computing center can provide sufficient computing power resources, greatly improving the development efficiency.

[0329] Optionally, the development requirements of the learning system include at least one of the following: the input method of learning materials, whether to prompt the user to provide learning key points during learning, whether to require the user to provide knowledge tags during learning, the learning ability required by the learning system, the learning depth, whether to provide the ability to visually consult the learned knowledge, the application link of the learning system in the to-be-developed intelligent agent, whether to provide practice cases, test cases, acceptance criteria, output requirements, technical requirements, performance requirements, security and privacy.

[0330] Optionally, the requirement confirmation module 12 is configured to communicate with the user through the interaction interface for at least one round according to the task information, in combination with the large language model method; determine the development requirements of the learning system according to the task information and the communication content with the user, in combination with the large language model method.

[0331] Optionally, the design module 13 is configured to obtain learning results by adopting at least one of the following learning methods according to the development requirements of the learning system, in combination with the large language model method: search the learning system knowledge base built in the learning system development intelligent body, search for relevant materials through a search engine, view and learn open source code, learn the existing code and documents of the intelligent body to be developed, and consult data;

[0332] The design module 13 is further configured to determine the design scheme of the learning system according to the learning results.

[0333] Optionally, the design scheme includes at least one of the following:

[0334] Modify the task processing module of the intelligent body to be developed so that the task processing module can identify whether the user task is a learning task;

[0335] Modify the upload module of the intelligent body to be developed to support the upload and archiving of learning materials;

[0336] Add a learning system;

[0337] The application link of the learning system in the intelligent body to be developed;

[0338] Modify the task processing module in the intelligent body to be developed so that the task processing module can call the knowledge extraction module of the learning system and integrate the knowledge learned by the learning system into the prompt words of the task processing module;

[0339] Add a practice module;

[0340] General design document, detailed design document, interface design document, and / or database design document.

[0341] Optionally, if the design of the learning system does not meet the development requirements, the design module 13 is configured to adjust the design scheme of the learning system by adopting at least one of the following adjustment methods in combination with the large language model method:

[0342] The first adjustment method is to optimize the design scheme of the learning system;

[0343] The second adjustment method is to adopt other design schemes.

[0344] Optionally, the debugging module 15 includes at least one of the following sub-modules:

[0345] The first debugging sub-module is used to terminate the development task of the learning system and feedback the termination reason and related situations to the user through the interaction interface when the number of debugging times reaches the first preset threshold, or the number of modifications to the design scheme reaches the second preset threshold, or the number of modifications to the code of the learning system reaches the third preset threshold, and the code of the learning system still cannot meet the development requirements;

[0346] The second debugging sub-module is used to record the running logs, intermediate data, and debugging results in combination with the large language model method during the debugging process and archive and save them.

[0347] Please refer to Figure 3 , the present invention also provides an electronic device 20, including a processor 21, a memory 22, and a computer program stored on the memory 22 and executable on the processor 21. When the computer program is executed by the processor 21, it implements each process of the above-mentioned method embodiment for automatically developing an intelligent agent learning system through computing power in the intelligent computing center, and can achieve the same technical effect. To avoid repetition, it will not be elaborated here.

[0348] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements each process of the above-mentioned method embodiment for automatically developing an intelligent agent learning system through computing power in the intelligent computing center, and can achieve the same technical effect. To avoid repetition, it will not be elaborated here. Among them, the computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, etc.

[0349] The present invention also provides a computer program product, including computer instructions, which implement each process of the above-mentioned Figure 1 shown method embodiment for automatically developing an intelligent agent learning system through computing power in the intelligent computing center when the computer instructions are executed by a processor, and can achieve the same technical effect. To avoid repetition, it will not be elaborated here.

[0350] It should be noted that in this text, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the phrase "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising that element.

[0351] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions for causing a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of the present invention.

[0352] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the purpose of the present invention and the scope protected by the claims, and all of them belong to the protection scope of the present invention.

Claims

1. A method for automatically developing an intelligent agent learning system through computing power in an intelligent computing center, characterized in that Executed by the intelligent agent developed by the learning system, including: Step S1: Receive, through the interaction interface, the task information provided by the user for constructing the learning system of the to-be-developed intelligent agent; Step S2: According to the task information, determine the development requirements of the learning system in combination with the large language model method and the user; Step S3: According to the development requirements, determine the design scheme of the learning system in combination with the large language model method; Step S4: Generate the code of the learning system according to the development requirements and design scheme in combination with the large language model method; Step S5: Deploy the operating environment of the learning system and debug the code of the learning system in combination with the large language model method; when the debugging result indicates that the learning system does not meet the development requirements, check the design scheme and code of the learning system in combination with the large language model method. When the design scheme needs to be modified, return to Step S3. When the code needs to be modified, return to Step S4 until the debugging result indicates that the learning system meets the development requirements or the number of debugging or modification times exceeds the preset threshold; Step S6: Feedback the result information to the user through the interaction interface.

2. The method according to claim 1, wherein The development requirements of the learning system include at least one of the following: the input method of learning materials, whether to prompt the user to provide learning key points during learning, whether to require the user to provide knowledge tags during learning, the learning capabilities that the learning system needs to possess, the learning depth, whether to provide the ability to visually view the learned knowledge, the application link of the learning system in the to-be-developed intelligent agent, whether to provide practice cases, test cases, acceptance criteria, output requirements, technical requirements, performance requirements, security and privacy.

3. The method according to claim 1, wherein The said Step S2 includes: Step S21: According to the task information, communicate with the user through the interaction interface for at least one round in combination with the large language model method; Step S22: Determine the development requirements of the learning system according to the task information and the communication content with the user in combination with the large language model method.

4. The method according to claim 1, wherein The said Step S3 includes: Step S31: According to the development requirements of the learning system, obtain learning results by using at least one of the following learning methods in combination with the large language model method: search relevant materials through a search engine, view and learn open source code, learn the existing code and documents of the to-be-developed intelligent agent; Step S32: Determine the design scheme of the learning system according to the learning results.

5. The method according to claim 4, characterized in that, The design scheme includes at least one of the following: Modify the task processing module of the to-be-developed intelligent agent so that the task processing module can identify whether the user task is a learning task; Modify the upload module of the to-be-developed intelligent agent to support the upload and archiving of learning materials; Add a learning system; The application link of the learning system in the to-be-developed intelligent agent; Modify the task processing module in the to-be-developed intelligent agent so that the task processing module can call the knowledge extraction module of the learning system and integrate the knowledge learned by the learning system into the prompt words of the task processing module; Add a practice module; General design document, detailed design document, interface design document and / or database design document.

6. The method according to claim 2 or 5, characterized in that, The learning system has at least one of the following learning capabilities: knowledge learning, knowledge archiving, knowledge extraction, knowledge matching, knowledge application, knowledge practice, knowledge display, knowledge encoding, knowledge maintenance, knowledge version maintenance, multi-document knowledge learning, large-document knowledge learning, parallel learning.

7. The method according to claim 2, wherein The acceptance criteria include: whether the learning system can apply the learned knowledge in the process of handling new tasks; Step S5 includes: Step S51: If there are practice cases, check whether the learned knowledge points are output after the to-be-developed intelligent agent runs the practice cases. If the learned knowledge points are output, the debugging result indicates that the learning system meets the development requirements; if the learned knowledge points are not output, the debugging result indicates that the learning system does not meet the development requirements.

8. An intelligent computing center realizes an automatic development device for an intelligent agent learning system through computing power, characterized in that Includes: A receiving module, configured to receive, through an interaction interface, task information provided by a user for constructing a learning system of a to-be-developed intelligent agent; A requirements confirmation module, configured to determine the development requirements of the learning system in combination with a large language model method according to the task information and the user; A design module, configured to determine the design scheme of the learning system in combination with a large language model method according to the development requirements; A coding module, configured to generate the code of the learning system in combination with a large language model method according to the development requirements and the design scheme; A debugging module, configured to deploy the running environment of the learning system and debug the code of the learning system in combination with a large language model method; When the debugging result indicates that the learning system does not meet the development requirements, check the design scheme and code of the learning system in combination with a large language model method. When it is necessary to modify the design scheme, trigger the design module to continue working. When it is necessary to modify the code, trigger the coding module to continue working until the debugging result indicates that the learning system meets the development requirements or the number of debugging or modification times exceeds a preset threshold; A feedback module, configured to feedback result information to the user through the interaction interface.

9. An electronic device, characterized in that, Includes: A processor, a memory, and a program stored on the memory and executable on the processor. When the program is executed by the processor, it implements the steps of the method for automatically developing an intelligent agent learning system by computing power of the intelligent computing center as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, it implements the steps of the method for automatically developing an intelligent agent learning system by computing power of the intelligent computing center as described in any one of claims 1 to 7.

11. A computer program product, characterized in that, Includes computer instructions. When the computer instructions are executed by a processor, it implements the steps of the method for automatically developing an intelligent agent learning system by computing power of the intelligent computing center as described in any one of claims 1 to 7.

Images