Water conservancy industry large model intelligent agent application system

Abstract

The invention provides a water conservancy industry large model agent application system, and belongs to the field of intelligent water conservancy, and the system comprises a large language model base unit which is used for providing a standard interface to interact with an open source or non-open source LLM, and carrying out the response operation of a user demand; a training data set of the business knowledge base unit is used for accumulating knowledge training LLM, and a knowledge base data set is used for supporting customized demands and expressing customized businesses as knowledge for intelligent agents to call; the water conservancy assembly library unit special for the large model is used for providing a pipe accommodating platform for all assemblies special for the large model; the water conservancy large model service decision support system unit is used for assembling business components into an application system and providing an operation tool to adjust the assembling process to form an applicable system; and the knowledge accumulation unit is used for extracting the system knowledge generated by the water conservancy large model service decision support system unit. The water conservancy application system architecture provided by the invention has complete functions, can be flexible and changeable along with businesses, and is short and long in service cycle.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent water conservancy, and in particular relates to a large-scale intelligent body application system for the water conservancy industry. Background Art

[0002] In recent years, the application of generative AI technology, multimodal data processing and preprocessing, pre-training of large models, and reinforcement learning has become the mainstream trend in the field of artificial intelligence applications. Today, LLM (Large Language Model) has had a significant impact on the AI ​​community, and the emergence of ChatGPT and GPT-4 has prompted people to rethink the possibility of AGI (general artificial intelligence). The adoption of these technologies has enabled artificial intelligence application products to have diversified interaction methods, the ability to process multiple tasks in parallel, and intelligent information perception, and has become the development direction of global artificial intelligence technology applications.

[0003] Currently, the entire industry is trying to use LLM to empower water conservancy business and create intelligent software. So far, there is no intelligent application that can meet actual business needs. Although there are some simple applications, they are not deep enough in the industry business, which makes these applications perform poorly. There is no mature system framework to go deep into the industry and combine new technologies to produce integrated applications.

[0004] Under the interactive mode of traditional information system design, the existing water conservancy industry application system has two major defects: first, due to its architectural limitations, the existing business application system is finalized after development and delivery. It is complex to operate and lacks business flexibility, making it difficult to adapt to the changing needs of the industry; second, the current system that combines LLM with the water conservancy industry is still mainly at the level of intelligent question and answer and text interaction, and in-depth industry business applications have not yet appeared. Summary of the invention

[0005] In order to solve the problems of complex operation, insufficient business flexibility and low industry deep integration of existing water conservancy industry application systems, the present invention provides a large model intelligent application system for the water conservancy industry.

[0006] In order to achieve the above object, the present invention provides the following technical solutions: A large model intelligent agent application system for water conservancy industry, comprising: Large language model base unit, used to provide a standard interface to interact with open source or non-open source large language model LLM, and respond to user needs; The business knowledge base unit includes a training data set for LLM model training and a knowledge base data set for LLM plug-in. The training data set is used to accumulate knowledge for LLM training, and the knowledge base data set is used to support customized needs and express customized business as knowledge for intelligent agents to call; The large-model dedicated water conservancy component library unit is used to provide a management platform for all large-model dedicated components, provide component service functions, and respond to system requirements; The water conservancy big model service decision support system unit is used to assemble business components into an application system and provide operation tools to adjust the assembly process to form an applicable system. According to the dynamic changes in demand, the system can be modified by adjusting components to adapt to the changes in demand; The knowledge accumulation unit is used to extract the system knowledge generated by the water conservancy large model service decision support system unit, and then organize the knowledge and classify it into a data set.

[0007] Preferably, the large language model base unit performs operations in response to user needs, specifically including: When the large language model base unit receives a user demand, it identifies the intention of the demand, first vectorizes the demand to obtain a demand vector, and then classifies the demand vector to obtain a classification result; The classification results are returned to LLM to determine whether the business knowledge base unit needs to be involved. If necessary, the business knowledge base unit is called up, and the results are reconstructed based on the result content output by the business knowledge base unit. The results are processed and output by LLM. If the business knowledge base unit does not need to be involved, LLM is used for direct output.

[0008] Preferably, the standard interface is an interface that takes a string as input and returns a string.

[0009] Preferably, the operation of responding to user needs specifically includes: The business knowledge base is divided into subject domains, and user needs are matched through intent recognition, and then LLM gives a recommended subject domain; under this subject domain, all business descriptions are vectorized to obtain business description vectors, and then a round of matching is performed; after matching the corresponding business, the business process description standard and the corresponding business are input into LLM as prompts, and then LLM generates the business process content to form a business description chain.

[0010] Preferably, the large-scale model-specific hydraulic component library unit is specifically used for: Provide a unified access address for component registration; Implant component service strategies and own component service governance strategies; Access component library units through authentication authorization; All connected components are managed in a unified group and divided into blacklist and whitelist groups; Generate a log of requested access after calling the component; A built-in authentication system is used for authorization, auditing, and access control when accessing the component library in other ways.

[0011] Preferably, the component access of the large-model dedicated water conservancy component library unit includes: component purpose and scope of application, component input parameters, component output parameters, and component protocol authentication.

[0012] Preferably, the large model-specific water conservancy component library unit responds to the LLM's component retrieval requirements, disassembles the generated business description chain, feeds it back to the LLM, gradually searches for corresponding components to match them, feeds the upstream and downstream relationships of the components back to the LLM, and generates corresponding component retrieval relationships; integrates the business description chain to generate business component relationships, and based on the business component relationships, passes them to the water conservancy large model service decision support system unit to generate a business application system.

[0013] Preferably, the water conservancy large model service decision support system unit includes: BS architecture belongs to the browser client architecture and runs in the browser; Generate a knowledge graph strategy, which is used to generate a logical topology knowledge graph based on the business description chain, and provide a modifiable interface on each generated topology node; The knowledge graph operation tool set is used to modify the nodes of the knowledge graph, and has the functions of adding, deleting and modifying nodes; Generate standard knowledge strategies so that the business system generated based on the knowledge graph can be output as text data that meets the business process description standards; Logging, used to generate system process logs.

[0014] Preferably, the operation of the water conservancy large model service decision support system unit to access the components is: The generated business expression chain is input into the unit, and the unit system parses the relationship and then generates a business process topology framework diagram. The business process topology framework diagram is delivered to the front-end system interface as a sketch, and the initial version of the business process topology framework diagram is determined, which is run to generate a business application system, and then the system is tested. The unit supports system rollback and correction until it meets business requirements, at which time the correction is terminated to obtain a business application system.

[0015] Preferably, the operations of the knowledge accumulation unit on knowledge accumulation are: extracting knowledge from the generated business application system, and entering the business knowledge pool through data mining and cleaning of text knowledge, and managing the knowledge in the business knowledge pool; configuring the knowledge in the knowledge pool as a knowledge base dedicated to a specific region to provide plug-in services to LLM, and serving as a training data set for LLM to guide LLM to emerge business optimization methods for the industry.

[0016] The beneficial effects of the water conservancy industry large model intelligent agent application system proposed by the present invention are as follows: The present invention responds to industry needs, generates business practices that meet the needs by combining LLM with a series of tools, transmits business components through the component library, assembles components from the water conservancy big model service decision support system unit to generate a business system, and then transfers knowledge from the business system to LLM, forming a universal solution paradigm for the water conservancy industry with a full-process closed loop. Its advantages are: First of all, it can deeply integrate water conservancy knowledge with LLM, so that LLM can use this knowledge to exert its own reasoning and emergence capabilities when facing water conservancy professional business. Combined with water conservancy professional knowledge, it can generate business practices within the entire industry. In theory, if the training data is sufficient, LLM will even give business understanding practices that surpass industry experts, thanks to its learning of a large amount of basic knowledge that is difficult for humans to achieve.

[0017] Then, the business logic is described in a standardized manner through business knowledge base units, which is beneficial to the long-term development of the industry. The unified description method will enable the large model to better understand the intention of personnel needs.

[0018] With the help of the large-model dedicated water conservancy component library unit, all functional modules are componentized to facilitate the decoupling of all systems, making the generated business application system more flexible and able to respond to changes in demand in a timely manner. On the other hand, the componentized modules also eliminate the need for repeated development work, and each component can be reused.

[0019] The water conservancy large model service decision support system unit can flexibly assemble all components. At this stage, industry professionals are involved in system generation to make the generated system more in line with the needs of actual applications. Since it is based on a component-based assembly system, the system development cycle is short and timeliness is strong. All generated business application systems can quickly respond to changes in demand, making the application system more flexible and easy to use.

[0020] With the help of the knowledge accumulation unit, the implicit industry expertise in the generated business application system can be fed back to the LLM for learning. Its data set can be used to train the LLM, forming a scientific closed-loop application model in which ideas guide practice, and practice in turn provides feedback on guiding ideas. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiment of the present invention and its design scheme, the following briefly introduces the drawings required for this embodiment. The drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 This is a block diagram of the large model intelligent agent application system for the water conservancy industry according to Embodiment 1 of the present invention; DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the technical solution of the present invention and implement it, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the scope of protection of the present invention.

[0024] In order to solve the problem of the existing water conservancy application system with a single architecture, inflexible and short service life, a more efficient, professional, accurate and secure intelligent application system based on a large language model in the water conservancy industry and a system development framework are needed to meet the comprehensive needs of the water conservancy field. This new technology needs to overcome the defects and problems of existing technologies, and also requires cross-industry professional knowledge for guidance. At the same time, it also requires the system to have a dynamic response mechanism and closed-loop learning logic to cope with the new needs and challenges of the water conservancy industry.

[0025] In summary, with the continuous development of artificial intelligence technology and water conservancy professional technology, smart water conservancy has become the development direction of the industry. The application of intelligent bodies based on large language models in the water conservancy industry will challenge traditional applications and define a new paradigm for water conservancy business software applications, improving the intelligence level, application efficiency, information-based decision-making level and knowledge accumulation of the business. In the future, after accumulating enough professional and useful knowledge, this technology will have emergent capabilities and the ability to generate industry problems. These capabilities can be combined with existing industry intelligent equipment to produce more valuable applications and realize all-round, efficient and safe intelligent applications for the water conservancy industry.

[0026] Example 1 Based on this, the present invention provides a large model intelligent agent application system for the water conservancy industry, specifically a system that can grasp the core of water conservancy business analysis and decision-making, support full-process information development and have dynamic adaptability, specifically Figure 1 As shown, the system uses the large language model base unit to identify and process water conservancy business needs, combines the business knowledge base unit to generate business application steps for the needs, combines the large model dedicated water conservancy component library unit to output the components required for the business steps, and then combines the water conservancy large model service decision support system unit to assemble the components, and then fine-tune the industry application personnel to generate an application system, and then combines the knowledge accumulation unit to accumulate and reuse business knowledge. The overall process is closed-loop, and each application is a feedback learning process, which makes the intelligent body more in line with the needs of the user, forming a virtuous circle, and realizing the definition of a new application paradigm for water conservancy industry applications, safety, efficiency, intelligence and knowledge accumulation, providing the water conservancy industry with a large model intelligent body software development and application solution that is available, usable, easy to use, effective, necessary, commonly used, and enjoyable.

[0027] Specifically, the present invention provides a water conservancy industry large model intelligent agent application system based on a large language model, including a large language model base unit, a business knowledge base unit, a large model dedicated water conservancy component library unit, a water conservancy large model service decision support system unit and a knowledge accumulation unit. The functions of each unit are further described below.

[0028] Large Language Model Base Unit: The large language model base unit is a collection used to support LLMs (all public large language models). In this unit, there are method classes for interface interaction with most mainstream LLMs. Its functions are common to all LLM types, providing a standard interface to interact with open source or non-open source LLMs. This interface takes a string as input and returns a string, and provides services in accordance with the REST API standard. Therefore, the LLM base can be flexibly selected according to the specific application scenario.

[0029] Since the source code of open source LLM is open, local deployment or cloud resource deployment can be selected according to the actual situation. Its advantage is that the base capability is fully controllable, and the base can be flexibly trained with all parameters or fine-tuned with some parameters, such as GLM, Llama2, Gemini and other bases. Closed source (i.e. non-open source) LLMs are often models with a huge number of parameters, and they perform well in basic capabilities. At present, these LLMs can provide services through standard interfaces and charge a certain amount of money. Their capabilities depend on the depth of model development by the owner. For example, GPT-4, Qwen, ERNIE Bot, etc. There are method classes that support two types of LLMs in the large language model base unit.

[0030] Furthermore, the large language model base unit performs a response operation to the user's needs, and the specific operation process of the response operation is: When the large language model base unit receives user needs, it recognizes the intention of the needs, first embeds (vectorizes) the needs to obtain the need vector, classifies the vector to obtain the classification result, and returns the result to the LLM to determine whether the business knowledge base unit needs to participate. If necessary, the business knowledge base unit is called, and the result is reconstructed according to the result content output by the business knowledge base unit, processed and output by LLM. If the business knowledge base unit does not need to participate, LLM processing is performed and output directly. For example: No knowledge base unit participation is required: (Q: "Hello", A: "Hello, I am XX"). Knowledge base unit participation is required: (Q: "Precipitation in XX area, how does the flood caused by this rainfall affect XX section?" A: "① Get the rainfall in XX area, call the rainfall query component, and get the rainfall sequence; ② Get the position relationship between XX section and XX area, call the position relationship component, and get the geographical location topology relationship; ③ Get the runoff model construction relationship, call the runoff model component, and get the model calculation result; ④ Get the result display method, call the software application display component, and get the software application interface").

[0031] Furthermore, when the large language model base unit needs to call the business knowledge base unit, it will combine the demand with the business knowledge base and input it into the LLM. The business knowledge base unit is further introduced below.

[0032] Business Knowledge Base Unit: The business knowledge base unit is divided into a training data set for LLM model training and a knowledge base data set for LLM plug-in. The two have different uses. The former is used to accumulate knowledge to train LLM, making LLM more in line with industry application requirements. The latter supports some customized requirements, mainly used to express customized business as knowledge for intelligent agents to call.

[0033] Since the intelligent body is driven by a large language model, the storage type of the business knowledge base data is text; the storage format of business knowledge conforms to the business process description standard. Under the present invention, the business process description standard is defined as: the business needs an overall description, including but not limited to: (business theme, business description set, business type, business scope of application, etc.) Business knowledge is described in steps, and each step must have three elements (① definition: this step describes how to do things ② function: what type of component needs to be called in this step ③ positioning: the relationship between this step and the upper and lower steps), and the components involved in the steps must have three elements (① definition: the purpose and scope of application of the component ② input parameters: what data the component requires ③ output parameters: what the component outputs).

[0034] Furthermore, the business knowledge base unit performs a response operation to the user's needs. The specific operation process of the response operation is as follows: First, the business knowledge base is divided into subject domains, such as (flood and drought disaster domain, irrigation area water distribution domain, ecological compensation domain, prediction domain, water rights trading domain, configuration and regulation domain, etc.). Each subject domain has its own description text, keywords, etc. The description is embedded to obtain the subject domain vector, and then the user needs are matched through intent recognition. Then LLM gives a recommended subject domain. Under this subject domain, all business descriptions are embedded to obtain the business description vector, and then a round of matching is performed. After matching the corresponding business, the business process description standard and the corresponding business are input into LLM as prompt (prompt word engineering), and then LLM generates the business process content, that is, the business expression chain.

[0035] Furthermore, the business knowledge base unit and the large language model base unit collaborate to generate a business expression chain, and extract corresponding components from the large model component library unit based on the business expression chain.

[0036] Large model dedicated hydraulic component library unit: The large-model dedicated water conservancy component library unit provides a management platform for all large-model dedicated components. It integrates all component services together and deploys them independently as a separate layer to solve some service governance issues. This unit is used to add, delete, modify and query components, which is equivalent to a component library. New components will be connected to this unit. It mainly includes six aspects: 1. Provide a unified access address for component registration This unit supports the access of all external components with standard interface protocols. Due to different component development methods, some provide HTTP services, some provide RPC services, and some provide Web Service services. In this unit, all components will be dynamically routed together, and different service protocols will be converted into REST API interface specifications.

[0037] 2. Have a component service governance strategy The service strategies of components can be implanted in this unit, such as component service fusing, diversion, degradation, etc.

[0038] 3. Authentication and authorization When accessing some component interfaces that require authentication and authorization, such as Oauth protocol authentication, Cookie authentication, Token authentication, etc. This unit provides a unified processing entry, and you only need to configure it in this unit to access the component library unit.

[0039] 4. Black and White Lists This unit can manage all connected components in a unified group, such as blacklist and whitelist groups for dimensions such as IP.

[0040] 5. Logging All components of this unit will have access request logs after they are called, which facilitates the maintenance of the unit.

[0041] 6. Unified authentication exposure The unit has a built-in authentication system for authorization, auditing, access control, etc. when accessing the component library in other ways.

[0042] Furthermore, the specific operation process of the large model dedicated water conservancy component library unit to access the component is as follows: Clarify the purpose and scope of application of the component. This item is a required option. This step is a text-based description. This description will be embedded in the unit to obtain a vector of the component introduction so that LLM can better understand the function of the component.

[0043] Clarify what data the component needs. This item is mandatory. This step mainly clarifies the input parameters of the component. Similarly, the vector of the component input parameters is obtained through embedding based on the text description. For example, {time series data, integer}.

[0044] Specify what the component outputs. This single option is mandatory. This step mainly specifies the output parameters of the component. Similarly, the vector of the component output parameters is obtained through embedding based on the text description. For example, {GIS data, grid data}.

[0045] Specify how the component is connected to the component library unit. This item is a required option. Select the type of component to be connected from the three mainstream service types, such as access via RPC service.

[0046] Specify what kind of protocol authentication the component requires. This item is optional. If the component does not require authentication, it can be directly connected. If authentication is required, select the corresponding access method. After configuring the rules, you can access the unit.

[0047] Furthermore, the large model dedicated water conservancy component library unit responds to the LLM component call demand, and the specific operation process of the response operation is: The business expression chain generated in the previous step is disassembled into steps and fed back to the LLM disassembly step method with examples. The disassembled steps are gradually matched with corresponding components. The upstream and downstream relationships of the components need to be fed back to the LLM with examples, and then the LLM generates the corresponding component call relationship.

[0048] Furthermore, the large model dedicated water conservancy component library unit integrates the business expression chain to generate business component relationships, and based on the business component relationships, transmits them to the water conservancy large model service decision support system unit to generate a business application system.

[0049] Water conservancy large model service decision support system unit: The water conservancy large model service decision support system unit is an integrated application system designed and developed specifically for large model services. Its function is to assemble business components into an application system and provide operating tools to adjust the assembly process. After the adjustment, an applicable system is directly formed, which can be dynamically changed according to demand. Only components need to be adjusted to quickly change the system to adapt to changes in demand. Finally, the assembled system knowledge is extracted and further passed to the knowledge accumulation unit.

[0050] Furthermore, the main components of the water conservancy large model service decision support system unit are the following five aspects: 1.BS architecture This unit belongs to the browser client architecture and runs in the browser.

[0051] 2. Generate knowledge graph strategy This unit can generate a logical topology knowledge graph based on the business description chain, and provide a modifiable interface on each generated topology node.

[0052] 3. Knowledge graph operation toolset The toolset can modify the nodes of the knowledge graph and has the functions of adding, deleting, and modifying nodes.

[0053] 4. Generate standard knowledge strategy The business system generated by this unit based on the knowledge graph can be output as text data that meets the business process description standards.

[0054] 5. Logging All system generation processes of this unit will generate logs to facilitate the maintenance of the unit.

[0055] Furthermore, the specific operation process of the water conservancy big model service decision support system unit to access the components is as follows: Input the business component relationship generated in the previous step (the large model-specific water conservancy component library unit integrates the business expression chain) into the unit, and the unit system parses the relationship, then generates a business process topology framework diagram, and delivers the business process topology framework diagram as a sketch to the front-end system interface.

[0056] Furthermore, experts from the water conservancy industry participated in revising the business process topology framework diagram, and the revised content included: overall topology structure, node attributes, node data, etc.

[0057] The water conservancy industry experts then determine the initial version of the business process topology framework diagram, run it, generate the business application system, and then test the system. This unit supports system rollback and correction until it meets business requirements, then terminate the correction and obtain the business application system.

[0058] Furthermore, when the demand changes, water conservancy industry experts can quickly roll back the system, modify the business logic in the business process topology knowledge graph, and regenerate the business application system.

[0059] Furthermore, the water conservancy big model service decision support system unit can translate the generated business application system into text knowledge that meets the business process description standards, and further pass it to the knowledge accumulation unit.

[0060] Knowledge accumulation unit: The function of the knowledge accumulation unit is to extract the system knowledge generated by the water conservancy large model service decision support system unit, and then organize the knowledge and classify it into the data set.

[0061] Furthermore, the specific operation process of the knowledge accumulation unit for knowledge accumulation is: First, the business application system generated in the previous step is subjected to knowledge extraction, and after data mining and cleaning of text knowledge, it enters the business knowledge pool. Knowledge can be managed in the business knowledge pool, such as knowledge sharing, knowledge reuse, knowledge privatization, etc. The knowledge in the knowledge pool can be further configured as a knowledge base dedicated to a specific region to provide plug-in services to LLM. On the other hand, it can be used as a training data set for LLM to guide LLM to emerge business optimization methods for the industry.

[0062] Each module in the above-mentioned water conservancy industry large model intelligent agent application system can be implemented in whole or in part through software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of the above modules.

[0063] The advantages of the large-scale intelligent agent application system for the water conservancy industry proposed in the present invention are: First of all, it can deeply integrate water conservancy knowledge with LLM, so that LLM can use this knowledge to exert its own reasoning and emergence capabilities when facing water conservancy professional business. Combined with water conservancy professional knowledge, it can generate business practices within the entire industry. In theory, if the training data is sufficient, LLM will even give business understanding practices that surpass industry experts, thanks to its learning of a large amount of basic knowledge that is difficult for humans to achieve.

[0064] Then, the business logic is described in a standardized manner through business knowledge base units, which is beneficial to the long-term development of the industry. The unified description method will enable the large model to better understand the intention of personnel needs.

[0065] With the help of the large-model dedicated water conservancy component library unit, all functional modules are componentized to facilitate the decoupling of all systems, making the generated business application system more flexible and able to respond to changes in demand in a timely manner. On the other hand, the componentized modules also eliminate the need for repeated development work, and each component can be reused.

[0066] The water conservancy large model service decision support system unit can flexibly assemble all components. At this stage, industry professionals are involved in system generation to make the generated system more in line with the needs of actual applications. Since it is based on a component-based assembly system, the system development cycle is short and timeliness is strong. All generated business application systems can quickly respond to changes in demand, making the application system more flexible and easy to use.

[0067] With the help of the knowledge accumulation unit, the implicit industry expertise in the generated business application system can be fed back to the LLM for learning. Its data set can be used to train the LLM, forming a scientific closed-loop application model in which ideas guide practice, and practice in turn provides feedback on guiding ideas.

[0068] It should be understood by those skilled in the art that embodiments of the present invention may provide methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0069] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0070] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0071] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0072] It should be pointed out that the specific implementation methods described above can enable those skilled in the art to understand the invention more comprehensively, but do not limit the invention in any way. Therefore, although the invention has been described in detail in this specification and embodiments, those skilled in the art should understand that the invention can still be modified or replaced by equivalents; and all technical solutions and improvements that do not deviate from the spirit and scope of the invention are included in the protection scope of the patent for the invention. Any figure mark in the claims should not be regarded as limiting the claims involved. Any simple change or equivalent replacement of the technical solution that can be obviously obtained by any technician familiar with the field within the technical scope disclosed in the present invention belongs to the protection scope of the present invention.

Claims

1. A large model intelligent agent application system for water conservancy industry, characterized in that: include: Large language model base unit, used to provide a standard interface to interact with open source or non-open source large language model LLM, and respond to user needs; The business knowledge base unit includes a training data set for LLM model training and a knowledge base data set for LLM plug-in. The training data set is used to accumulate knowledge for LLM training, and the knowledge base data set is used to support customized needs and express customized business as knowledge for intelligent agents to call; The large-model dedicated water conservancy component library unit is used to provide a management platform for all large-model dedicated components, provide component service functions, and respond to system requirements; The water conservancy big model service decision support system unit is used to assemble business components into an application system and provide operation tools to adjust the assembly process to form an applicable system. According to the dynamic changes in demand, the system can be modified by adjusting components to adapt to the changes in demand; The knowledge accumulation unit is used to extract the system knowledge generated by the water conservancy large model service decision support system unit, and then organize the knowledge and classify it into a data set.

2. The large-scale intelligent agent application system for water conservancy industry according to claim 1 is characterized in that: The large language model base unit responds to user needs, including: When the large language model base unit receives a user demand, it identifies the intention of the demand, first vectorizes the demand to obtain a demand vector, and then classifies the demand vector to obtain a classification result; The classification results are returned to LLM to determine whether the business knowledge base unit needs to be involved. If necessary, the business knowledge base unit is called up, and the results are reconstructed based on the result content output by the business knowledge base unit. The results are processed and output by LLM. If the business knowledge base unit does not need to be involved, LLM is used for direct output.

3. The large-scale intelligent agent application system for water conservancy industry according to claim 1 is characterized in that: The standard interface is an interface that takes a string as input and returns a string.

4. The large-scale intelligent agent application system for water conservancy industry according to claim 2 is characterized in that: The operation of responding to user needs specifically includes: The business knowledge base is divided into subject domains, and user needs are matched through intent recognition, and then LLM gives a recommended subject domain; under this subject domain, all business descriptions are vectorized to obtain business description vectors, and then a round of matching is performed; after matching the corresponding business, the business process description standard and the corresponding business are input into LLM as prompts, and then LLM generates the business process content to form a business description chain.

5. The large-scale intelligent agent application system for water conservancy industry according to claim 4 is characterized in that: The large model dedicated water conservancy component library unit is specifically used for: Provide a unified access address for component registration; Implant component service strategies and own component service governance strategies; Access component library units through authentication authorization; All connected components are managed in a unified group and divided into blacklist and whitelist groups; Generate a log of requested access after calling the component; A built-in authentication system is used for authorization, auditing, and access control when accessing the component library in other ways.

6. The water conservancy industry large model intelligent agent application system according to claim 4 is characterized in that: The large-model dedicated water conservancy component library unit performs unified grouping management on the connected components, and the management content includes: component purpose and scope of application, component input parameters, component output parameters, and component protocol authentication.

7. The water conservancy industry large model intelligent agent application system according to claim 5 is characterized in that: The large model-specific water conservancy component library unit responds to the LLM's component retrieval requirements, disassembles the generated business description chain, feeds it back to the LLM, gradually searches for corresponding components to match them, feeds the upstream and downstream relationships of the components back to the LLM, and generates corresponding component retrieval relationships; integrates the business description chain to generate business component relationships, and passes them to the water conservancy large model service decision support system unit based on the business component relationships to generate a business application system.

8. The water conservancy industry large model intelligent agent application system according to claim 1 is characterized in that: The water conservancy large model service decision support system unit includes: BS architecture belongs to the browser client architecture and runs in the browser; Generate a knowledge graph strategy, which is used to generate a logical topology knowledge graph based on the business description chain, and provide a modifiable interface on each generated topology node; The knowledge graph operation tool set is used to modify the nodes of the knowledge graph, and has the functions of adding, deleting and modifying nodes; Generate standard knowledge strategies so that the business system generated based on the knowledge graph can be output as text data that meets the business process description standards; Logging, used to generate system process logs.

9. The water conservancy industry large model intelligent agent application system according to claim 8 is characterized in that: The operation of the water conservancy large model service decision support system unit also includes: The generated business expression chain is input into the unit, and the unit system parses the relationship, then generates a business process topology framework diagram, and transmits the business process topology framework diagram as a sketch to the front-end system interface. Determine the initial version of the business process topology framework diagram, run it, generate the business application system, and then test the system. This unit supports system rollback and correction until it meets business requirements, then terminate the correction and obtain the business application system.

10. The water conservancy industry large model intelligent agent application system according to claim 9 is characterized in that: The operation of the knowledge accumulation unit on knowledge accumulation is: extracting knowledge from the generated business application system, and entering the business knowledge pool through data mining and cleaning of text knowledge, and managing the knowledge in the business knowledge pool; The knowledge in the knowledge pool can be configured as a knowledge base dedicated to a specific region to provide plug-in services to LLM, and serve as a training data set for LLM to guide LLM to emerge business optimization methods for the industry.

Images