Intelligent Agent system implementation method and system for skill adaptive parameter filling
By using large language models and parameter inspection modules in the intelligent Agent system for parameter extraction and clarification, the problem of inaccurate parameter extraction in complex user intention processing is solved, and more efficient and accurate task execution and better user experience is achieved.
Patent Information
- Application Number
- CN202510002650.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-02
- Publication Date
- 2025-05-06
AI Technical Summary
When dealing with complex user intentions and multi-step tasks, the intelligent agent system faces the problem of inaccurate parameter extraction and lack of effective parameter clarification mechanisms, resulting in a decline in user experience.
Large language model (LLM) is used to parse user input, generate initial parameters, and perform multi-dimensional inspection through the parameter inspection module. When errors or missing are found, use the predefined prompt template to generate parameters to fill in the request or clarify the problem, and guide the system or user to provide necessary information.
It improves the efficiency and accuracy of complex task processing, especially in the multi-parameter and multi-round interaction skill calling scenarios, which improves the user experience and reduces the possibility of task execution failure.
Smart Images
Figure CN119940538A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of artificial intelligence and natural language processing, and in particular to an intelligent Agent system implementation method and system for skill adaptive parameter filling. Background Art
[0002] With the development of artificial intelligence technology, especially the progress in the field of natural language processing (NLP), intelligent agent systems have become an important bridge between humans and the digital world. These systems usually include chatbots, virtual assistants and other forms of human-computer interaction interfaces, which play an increasingly important role in daily life, such as in customer service, smart home control, online shopping and other fields.
[0003] However, in practical applications, one of the main challenges facing intelligent agent systems is how to accurately understand and process user intentions, especially when complex operations or multi-step tasks are involved. To achieve this, the system must be able to correctly parse user input and convert it into executable instructions. This process involves parameter extraction, parameter validation, and necessary user interaction to clarify ambiguous or incomplete instructions.
[0004] Traditional parameter extraction methods usually rely on fixed pattern matching or simple keyword recognition. Although this method is simple and easy to use, it is incapable of dealing with diverse and complex user input. Incorrect parameter extraction will lead to skill execution failure and affect user experience. In addition, when the system cannot correctly interpret user intent, there is a lack of effective mechanisms for parameter clarification and filling, forcing users to try repeatedly to successfully perform tasks, which undoubtedly increases user frustration.
[0005] In recent years, with the advancement of deep learning technology, especially the development of large language models (LLMs), machines can better understand natural language input and generate more natural and accurate responses. However, even advanced LLMs may produce outputs that do not conform to the expected format or type, especially in parameter extraction. Therefore, how to ensure that the parameters generated by LLMs can both meet functional requirements and maintain a good user experience has become an urgent problem to be solved.
[0006] In order to solve the above problems, some attempts in the existing technology include introducing more complex natural language understanding algorithms, developing specialized parameter verification tools, and designing more humanized user interaction interfaces. Although these measures have improved system performance to a certain extent, there are still deficiencies in terms of parameter accuracy, flexibility, and naturalness of interaction. Summary of the invention
[0007] The technical task of the present invention is to address the above shortcomings and provide an intelligent Agent system implementation method and system for skill adaptive parameter filling, which provides a more efficient, accurate and user-friendly skill call processing method, enabling the system to efficiently perform tasks while improving user experience.
[0008] The technical solution adopted by the present invention to solve its technical problem is:
[0009] A method for implementing an intelligent agent system with adaptive skill parameter filling, the method comprising:
[0010] Parse user input through a large language model (LLM) to generate initial parameters for the downstream skills selected by the user;
[0011] The parameter checking module performs a comprehensive check on the initial parameters for various correctness, including multiple dimensions such as quantity, type, format and existence; and the parameter checking module can handle complex data structures including arrays and enumerations;
[0012] When parameter errors or missing parameters are found, targeted parameter filling requests or clarification questions are generated through predefined prompt templates and large language models to guide the system or user to provide the necessary missing or error-correcting information.
[0013] This method aims to improve the efficiency and accuracy of complex task processing, especially in skill invocation scenarios with multiple parameters and multiple rounds of interaction. It overcomes the defects of the existing technology through precise parameter checking mechanism, flexible skill parameter filling process and intelligent parameter clarification dialogue ability, and provides a more efficient, accurate and user-friendly skill invocation processing method. This enables the system to perform tasks efficiently while improving user experience.
[0014] Furthermore, predefined prompt templates are used to guide large language models to fill in parameters;
[0015] The prompt template clearly defines the task objectives, output format and precautions, and provides multiple examples to illustrate the correct output in different situations.
[0016] This method not only ensures the accuracy of parameter filling, but also improves the flexibility of the system and can adapt to various complex parameter filling scenarios.
[0017] Furthermore, the parameter checking module is implemented based on a regular rule algorithm, which can verify the number, type, format and existence of parameters, and check whether the parameters generated by the model meet the predefined parameter interface requirements in combination with the predefined skill parameter list; the parameter checking module not only checks basic data types, but also handles array element types and enumeration values. Through this multi-dimensional checking mechanism, task execution failures caused by parameter errors can be effectively reduced.
[0018] Furthermore, when the parameter checking mechanism finds a parameter problem, it generates targeted clarification questions based on the template and the parameter checking results, accurately points out the missing or wrong parameters, and expresses them in natural and professional language, so that users can better understand the parameter problem and complete the necessary information supplementation;
[0019] It supports multiple rounds of interactive parameter clarification and filling. In each round of interaction, it not only focuses on the parameters that need to be clarified at the moment, but also considers the context of the entire task and the previous dialogue history, enabling the system to solve multiple parameter problems in a single interaction, reducing unnecessary back-and-forth communication.
[0020] The present invention also claims protection for an intelligent agent system for skill adaptive parameter filling, comprising a user input parsing module, a parameter generation module, a parameter checking module, a parameter filling and clarification module, a multi-round interaction management module and a downstream skill calling module, wherein:
[0021] The parameter generation module preliminarily generates call parameters based on user input and downstream skill requirements. The parameter generation module is tightly integrated with the Large Language Model (LLM) to fill in parameter values based on the natural language description provided by the user.
[0022] The parameter checking module performs a comprehensive multi-dimensional check on the generated initial parameters to check whether the number of parameters meets the requirements and verifies the type, format and existence of the parameters based on regular expressions or rule bases. In addition, complex data structures (such as arrays or enumeration types) will also be processed by a special verification mechanism.
[0023] Parameter filling and clarification module. When the parameter checking module finds a problem, the parameter filling and clarification module cooperates with the large language model LLM through predefined prompt templates to generate targeted supplementary or clarification requests; these requests are transmitted to the user or system in the form of natural language, and the user provides the correct information based on the system feedback, or the system automatically infers and fills in the information.
[0024] Furthermore, the parameter generation module includes the following steps:
[0025] The system automatically fills in some explicit parameters by analyzing the user input;
[0026] For parameters that are not provided, the system marks the status as Required Clarification and starts the clarification process in the subsequent steps.
[0027] Furthermore, the parameter checking module includes the following steps:
[0028] The parameter checking module receives the generated initial parameters and checks each parameter in turn;
[0029] The inspection contents include: quantity, data type, format verification (such as date, email address), parameter legal value range, etc.
[0030] For parameters that do not meet the requirements, the system records the error type and passes this information to the parameter filling and clarification module.
[0031] Furthermore, the parameter filling and clarification module includes the following steps:
[0032] Based on the feedback provided by the parameter checking module, the system selects the appropriate prompt template and generates clarifying questions or supplementary requests;
[0033] The user or system responds to the clarification request by providing missing or corrected parameter information;
[0034] The parameter filling and clarification module passes the new information back to the parameter checking module for re-verification.
[0035] The present invention also claims a device for implementing an intelligent Agent system with adaptive skill parameter filling, comprising: at least one memory and at least one processor;
[0036] The at least one memory is used to store a machine-readable program;
[0037] The at least one processor is used to call the machine-readable program to implement the above method.
[0038] The present invention also claims protection for a computer-readable medium having computer instructions stored thereon, which implement the above method when executed by a processor.
[0039] Compared with the prior art, the intelligent agent system implementation method and system for skill adaptive parameter filling of the present invention have the following beneficial effects:
[0040] This invention proposes a new intelligent Agent system design, which overcomes the defects of the prior art through precise parameter checking mechanism, flexible skill parameter filling process and intelligent parameter clarification dialogue capability, and improves the efficiency and accuracy in complex task processing, especially in skill calling scenarios with multiple parameters and multiple rounds of interaction. The introduction of precise parameter checking mechanism, flexible parameter filling process and intelligent parameter clarification dialogue capability enables the system to improve user experience while efficiently executing tasks. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 This is a schematic diagram of the implementation principle of an intelligent Agent system for skill adaptive parameter filling provided by an embodiment of the present invention;
[0042] Figure 2 This is an example diagram of a parameter filling prompt word template provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0043] The present invention will be further described below in conjunction with specific embodiments.
[0044] The embodiment of the present invention provides a method for implementing an intelligent agent system with adaptive skill parameter filling.
[0045] The user input is parsed through a large language model (LLM) to generate initial parameters for the downstream skills selected by the user; these initial parameters are not the final results, but are verified in multiple dimensions in detail through a dedicated parameter checking module.
[0046] The parameter checking module performs a comprehensive check on the initial parameters for various correctness, including multiple dimensions such as quantity, type, format and existence; and the parameter checking module can handle complex data structures including arrays and enumerations;
[0047] When parameter errors or missing parameters are found, targeted parameter filling requests or clarification questions are generated through predefined prompt templates and large language models to guide the system or user to provide the necessary missing or error-correcting information.
[0048] The implementation of this method includes:
[0049] Template parameter filling process:
[0050] Predefined prompt templates are used to guide the large language model (LLM) to fill in parameters. Each template defines the task objectives, expected output format, and necessary precautions in detail, and provides multiple examples to show the correct output form in different situations. This design ensures the accuracy of parameter filling and the flexibility of the system, enabling it to cope with various complex parameter filling requirements. Figure 2The parameter filling prompt word template example shown is shown.
[0051] Multi-dimensional parameter checking mechanism:
[0052] The system has a built-in parameter checking module based on regular rules and other algorithms to verify the number, type, format and existence of parameters. The system will check whether the parameters generated by the model meet the predefined parameter interface requirements in combination with the predefined skill parameter list. In addition to basic data type verification, the module can also identify and verify more complex structures such as array element types and enumeration values. Through this multi-dimensional checking mechanism, task execution failures caused by parameter errors can be effectively reduced.
[0053] Context-aware multi-round interaction optimization:
[0054] When the parameter check mechanism detects a parameter problem, the system will automatically generate a series of targeted clarification questions based on the pre-set template and the results of the parameter check. These questions not only clearly indicate which parameters have problems, but also express them in easy-to-understand language to help users accurately identify and correct parameter errors and complete the necessary information supplement. In addition, the system supports multiple rounds of interactive clarification and filling, and can take into account the overall context of the task and the previous conversation history in each conversation, so as to solve as many parameter problems as possible in one conversation round. That is, the system can solve multiple parameter problems in a single interaction, reducing unnecessary communication between users and the system.
[0055] The specific implementation of this method is as follows:
[0056] 1. The parameter generation module generates the initial call parameters based on the user input and the requirements of downstream skills. The parameter generation module is closely integrated with the large language model (LLM) and can fill in the parameter values based on the natural language description provided by the user. The parameter generation module specifically includes:
[0057] 1.1. The system automatically fills in some explicit parameters by analyzing the user input;
[0058] 1.2. For parameters that are not provided, the system marks the status as requiring clarification and starts the clarification process in subsequent steps.
[0059] 2. Through the parameter checking module, the generated initial parameters are comprehensively checked in multiple dimensions to check whether the number of parameters meets the requirements, and the type, format and existence of the parameters are verified based on regular expressions or rule bases; in addition, complex data structures (such as arrays or enumeration types) will also be processed by a special verification mechanism. The parameter checking module specifically includes:
[0060] 2.1. The parameter checking module receives the generated initial parameters and checks each parameter in turn;
[0061] 2.2. Check contents include: quantity, data type, format verification (such as date, email address), parameter legal value range, etc.;
[0062] 2.3. For parameters that do not meet the requirements, the system records the error type and passes this information to the parameter filling and clarification module.
[0063] 3. Through the parameter filling and clarification module, when the parameter checking module finds a problem, it cooperates with the large language model LLM through the predefined prompt template to generate targeted supplementary or clarification requests. These requests are transmitted to the user or system in the form of natural language. The user provides the correct information based on the system feedback, or the system automatically guesses and fills in the information. The parameter filling and clarification module specifically includes:
[0064] 3.1. The system selects appropriate prompt templates and generates clarification questions or supplementary requests based on the feedback provided by the parameter check module;
[0065] 3.2. The user or system responds to the clarification request and provides missing or corrected parameter information;
[0066] 3.3. The parameter filling and clarification module passes the new information back to the parameter checking module for re-verification.
[0067] This method overcomes the defects of the existing technology through a precise parameter checking mechanism, a flexible skill parameter filling process and an intelligent parameter clarification dialogue capability, and provides a more efficient, accurate and user-friendly skill call processing method.
[0068] The embodiment of the present invention also provides an intelligent agent system for skill adaptive parameter filling. The system first parses the user input through a large language model (LLM) to generate initial parameters for the downstream skills selected by the user. Subsequently, a special parameter checking module performs a comprehensive check on the correctness of these parameters, including multiple dimensions such as quantity, type, format and existence. If a parameter problem is found, the system will generate targeted parameter filling requests or clarifying questions through LLM according to the predefined prompt template, and guide the system or user to provide missing or corrected information. This process may involve multiple rounds of interaction until all parameters meet the call requirements of downstream skills. In this way, an intelligent, precise and powerful dialogue-capable intelligent agent skill call processing system is formed. It can not only effectively improve the processing efficiency and accuracy of complex tasks, but also enhance the user experience through natural and professional dialogue methods, making the entire interaction process more smooth and friendly.
[0069] The system includes a user input parsing module, a parameter generation module, a parameter checking module, a parameter filling and clarification module, a multi-round interaction management module and a downstream skill calling module. Each module cooperates with each other to realize an intelligent and automated parameter filling and skill calling process.
[0070] 1. The parameter generation module is responsible for initially generating call parameters based on user input and downstream skill requirements. The parameter generation module is tightly integrated with the Large Language Model (LLM) and can fill in parameter values based on the natural language description provided by the user. The implementation steps include:
[0071] 1.1. The system automatically fills in some explicit parameters by analyzing the user input;
[0072] 1.2. For parameters that are not provided, the system marks the status as requiring clarification and starts the clarification process in subsequent steps.
[0073] 2. The parameter checking module is responsible for performing a comprehensive multi-dimensional check on the generated initial parameters, checking whether the number of parameters meets the requirements, and verifying the type, format and existence of the parameters based on regular expressions or rule bases; in addition, complex data structures (such as arrays or enumeration types) will also be processed by a special verification mechanism. The implementation steps include:
[0074] 2.1. The parameter checking module receives the generated initial parameters and checks each parameter in turn;
[0075] 2.2. Check contents include: quantity, data type, format verification (such as date, email address), parameter legal value range, etc.;
[0076] 2.3. For parameters that do not meet the requirements, the system records the error type and passes this information to the parameter filling and clarification module.
[0077] 3. Parameter filling and clarification module. When the parameter checking module finds a problem, the parameter filling and clarification module cooperates with the large language model LLM through predefined prompt templates to generate targeted supplementary or clarification requests. These requests are delivered to the user or system in the form of natural language. The user provides the correct information based on the system feedback, or the system automatically guesses and fills in the information. The implementation steps include:
[0078] 3.1. The system selects appropriate prompt templates and generates clarification questions or supplementary requests based on the feedback provided by the parameter check module;
[0079] 3.2. The user or system responds to the clarification request and provides missing or corrected parameter information;
[0080] 3.3. The parameter filling and clarification module passes the new information back to the parameter checking module for re-verification.
[0081] The embodiment of the present invention also provides an intelligent Agent system implementation device for skill adaptive parameter filling, comprising: at least one memory and at least one processor;
[0082] The at least one memory is used to store a machine-readable program;
[0083] The at least one processor is used to call the machine-readable program to implement the intelligent Agent system implementation method for skill adaptive parameter filling described in the above embodiment.
[0084] The embodiment of the present invention also provides a computer-readable medium, on which a computer instruction is stored, and when the computer instruction is executed by a processor, the processor executes the method for implementing the intelligent agent system for filling skill adaptive parameters described in the above embodiment. Specifically, a system or device equipped with a storage medium can be provided, on which a software program code for implementing the functions of any of the above embodiments is stored, and a computer (or CPU or MPU) of the system or device reads and executes the program code stored in the storage medium.
[0085] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.
[0086] The storage medium embodiments for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code can be downloaded from a server computer by a communication network.
[0087] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling an operating system operating on the computer to complete part or all of the actual operations based on instructions from the program code.
[0088] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or written to a memory provided in an expansion unit connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or the expansion unit is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.
[0089] The present invention is shown and described in detail above through the accompanying drawings and preferred embodiments. However, the present invention is not limited to these disclosed embodiments. Based on the above multiple embodiments, those skilled in the art can know that the code review methods in the above different embodiments can be combined to obtain more embodiments of the present invention, and these embodiments are also within the protection scope of the present invention.
Claims
1. A method for implementing an intelligent agent system with adaptive skill parameter filling, characterized in that: The implementation of this method includes: Parse user input through a large language model to generate initial parameters for the downstream skills selected by the user; The parameter checking module performs a comprehensive check on the initial parameters for various correctness, including the number, type, format and existence dimensions; and the parameter checking module is capable of processing complex data structures including arrays and enumerations; When parameter errors or missing parameters are found, targeted parameter filling requests or clarification requests are generated through predefined prompt templates and large language models to guide the system or user to provide the necessary missing or error-correcting information.
2. The method for implementing an intelligent agent system for skill adaptive parameter filling according to claim 1, characterized in that: Use predefined hint templates to guide large language models for parameter filling; The prompt template clearly defines the task objectives, output format and precautions, and provides multiple examples to illustrate the correct output in different situations.
3. The method for implementing an intelligent agent system for skill adaptive parameter filling according to claim 1, characterized in that: The parameter checking module is implemented based on a regular rule algorithm, and can verify the number, type, format and existence of parameters. Combined with the predefined skill parameter list, it checks whether the parameters generated by the model meet the predefined parameter interface requirements; and can handle array element types and enumeration values.
4. The method for implementing an intelligent agent system for skill adaptive parameter filling according to claim 1, characterized in that: When the parameter check mechanism finds a parameter problem, it generates targeted clarification questions based on the template and the parameter check results, accurately points out the missing or incorrect parameters, and expresses them in natural and professional language, so that users can better understand the parameter problem and complete the necessary information supplementation; It supports multi-round interactive parameter clarification and filling. In each round of interaction, it not only focuses on the parameters that need to be clarified at the moment, but also considers the context of the entire task and the previous dialogue history, enabling the system to solve multiple parameter problems in a single interaction.
5. An intelligent agent system with adaptive skill parameter filling, characterized in that: It includes user input parsing module, parameter generation module, parameter checking module, parameter filling and clarification module, multi-round interaction management module and downstream skill calling module, among which, The parameter generation module initially generates call parameters based on user input and downstream skill requirements. The parameter generation module is tightly integrated with the large language model and fills in parameter values based on the natural language description provided by the user. The parameter checking module performs a comprehensive multi-dimensional check on the generated initial parameters to check whether the number of parameters meets the requirements and verifies the type, format and existence of the parameters based on regular expressions or rule bases; Parameter filling and clarification module. When the parameter checking module finds a problem, the parameter filling and clarification module cooperates with the large language model through predefined prompt templates to generate targeted supplementary or clarification requests; these requests are transmitted to the user or system in the form of natural language, and the user provides the correct information based on the system feedback, or the system automatically infers and fills in the information.
6. The intelligent agent system for skill adaptive parameter filling according to claim 5, characterized in that: The parameter generation module includes the following steps: The system automatically fills in some explicit parameters by analyzing the user input; For parameters that are not provided, the system marks the status as Required Clarification and starts the clarification process in the subsequent steps.
7. The intelligent agent system for skill adaptive parameter filling according to claim 5, characterized in that: The parameter checking module includes the following steps: The parameter checking module receives the generated initial parameters and checks each parameter in turn; The inspection contents include: quantity, data type, format verification, and legal value range of parameters; For parameters that do not meet the requirements, the system records the error type and passes this information to the parameter filling and clarification module.
8. The intelligent agent system for skill adaptive parameter filling according to claim 5, characterized in that: The parameter filling and clarification module includes the following steps: Based on the feedback provided by the parameter checking module, the system selects the appropriate prompt template and generates clarifying questions or supplementary requests; The user or system responds to the clarification request by providing missing or corrected parameter information; The parameter filling and clarification module passes the new information back to the parameter checking module for re-verification.
9. An intelligent agent system implementation device for skill adaptive parameter filling, characterized in that: include: at least one memory and at least one processor; The at least one memory is used to store a machine-readable program; The at least one processor is used to call the machine-readable program to implement the method described in any one of claims 1 to 4.
10. A computer-readable medium, characterized in that The computer readable medium stores computer instructions, and when the computer instructions are executed by a processor, the method according to any one of claims 1 to 4 is implemented.