Multi-system capability intelligent execution method and system based on human-computer interaction
Through multi-system capability intelligent execution methods and systems based on human-computer interaction, management and execution problems in multi-system environments are solved, user-friendly intelligent operation and cross-system collaboration are achieved, and work efficiency and system intelligence are improved.
Patent Information
- Application Number
- CN202411020560.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-12-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In a multi-system environment, it is difficult for users to effectively manage and execute system capabilities, resulting in inefficiency and increased possibility of errors. The lack of a unified management and scheduling mechanism makes it difficult to achieve collaborative operation and data sharing across systems.
Using multi-system capability intelligent execution methods and systems based on human-computer interaction, the user's intentions are matched through background configuration and intention algorithms, and slots and execution solutions are automatically extracted to achieve unified management and efficient execution of multiple systems.
It improves user's work efficiency and experience, realizes the intelligent and personalized service of the system, simplifies user operation processes, enhances the scalability, robustness and reusability of the system, and supports cross-system collaborative operation and data sharing.
Smart Images

Figure CN119065703A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence technology, and more specifically to a multi-system capability intelligent execution method and system based on human-computer interaction. Background Art
[0002] At present, with the rapid development of information technology, modern society has entered a new era of digitization, networking and intelligence. Various complex systems emerge in an endless stream and are widely used in various fields, such as enterprise management, production and manufacturing, financial services, medical health, etc. However, with the increase in the number of systems and the increasing complexity of functions, users face many challenges when using these systems. Among them, one of the most prominent issues is how to effectively manage and execute the capabilities of these systems.
[0003] However, when there are a large number of systems, it is often difficult for users to accurately remember the address and access method of each system, let alone have an in-depth understanding of the functions and operation methods of each system. This not only reduces the user's work efficiency, but also increases the possibility of errors. In addition, even if the user can access a system, the functions and data between systems are often independent of each other and lack a unified management and scheduling mechanism. It is difficult for users to achieve cross-system collaborative operations and data sharing.
[0004] Therefore, how to achieve unified management and efficient execution of multiple systems to improve user work efficiency and experience is an urgent problem that technical personnel in this field need to solve. Summary of the invention
[0005] In view of this, the present invention provides a multi-system capability intelligent execution method and system based on human-computer interaction, which can not only solve the management and execution problems faced by users when using multiple systems and improve work efficiency and accuracy, but also provide users with a more intelligent and personalized service experience.
[0006] In order to achieve the above object, the present invention adopts the following technical solution:
[0007] A multi-system capability intelligent execution method based on human-computer interaction includes the following steps:
[0008] S1. Perform dialogue configuration, system configuration, function configuration, intent configuration, slot configuration, execution plan configuration and knowledge base configuration through the background according to the changes in customer needs;
[0009] S2. After the configuration is completed, the user inputs the action to be executed through human-computer interaction and converts it into text;
[0010] S3. Call the intent algorithm to match the most similar intent and the slot associated with the intent according to the intent configured by the system until all slots of the intent have been extracted;
[0011] S4. According to the matching intention and the relevant slot information, the matching intention matches the corresponding execution plan, matches the corresponding execution plan, and executes the nodes configured in the plan in sequence according to the steps of the execution plan;
[0012] S5. Display the results after the plan is executed.
[0013] Preferably, the specific content of step S1 is:
[0014] The dialog configuration is used to define the parameters during the interaction between the user and the system;
[0015] System configuration is to complete the configuration of the system to be accessed, including the system name, code, URL, login script and system parameters;
[0016] Function configuration is to complete the page corresponding to the system and the button information under the page, including the system to which the function belongs, the function attribution, the function type, the function execution script, the execution end, the function input parameter and the function output parameter;
[0017] Intent configuration is to complete the definition of the interaction requirements that users need to perform, including the intent name, associated system, associated function, intent type, intent parsing method, and intent-associated slot;
[0018] The slot configuration is to complete the keyword extraction definition of the user input information, including the slot name and the slot extraction method;
[0019] The execution scheme configuration is to complete the execution link configuration for the intent. The link contains each execution step and the execution action configuration of each step, including the intent ID associated with the execution scheme, the function node of the execution scheme, and the link of the function node;
[0020] The knowledge base is configured to store the data required in the execution plan by importing, and vectorizes and stores the corresponding fields of the imported data as needed.
[0021] Preferably, the intent types include global intent and dynamic intent; global intent is an intent type that is used by multiple systems, and dynamic intent is an intent type for a specific system.
[0022] Preferably, the method of parsing intent and the method of slot extraction both include regular expression parsing, machine learning training model parsing and large model prompt word parsing.
[0023] Preferably, the solution node includes node CODE, node serial number, function ID, input parameters, and output parameters; the node link includes the previous node ID, the next node ID, and a conditional expression, and the node link is executed if the conditional expression is met.
[0024] Preferably, step S4 further includes: if the configuration scheme is complex, in the process of interaction, the user is asked multiple times to complete the parameters required for the intention, and steps S2-S4 are repeated for multiple interactions to complete the action that the user needs to perform.
[0025] A multi-system capability intelligent execution system based on human-computer interaction, based on the multi-system capability intelligent execution method based on human-computer interaction, including a background and a user end;
[0026] The backend includes the backend configuration terminal, intent recognition module, slot extraction module, knowledge base and solution planning module, and function execution module;
[0027] The user end includes a human-computer interaction interface, specifically an input module and a display module;
[0028] The backend configuration end is used to configure the dialogue, system, function, intent, slot, execution plan, and knowledge base according to the changing needs of customers;
[0029] The input module is used to allow users to input actions to be executed through human-computer interaction and convert them into text after the configuration is completed;
[0030] The intent recognition module is used to call the intent algorithm to match the most similar intent according to the intent configured by the system;
[0031] The slot extraction module is used to match the slots associated with the intent until all slots of the intent have been extracted;
[0032] The knowledge base is used to store the data required for the execution plan in an imported manner, and vectorize the corresponding fields of the imported data as needed;
[0033] A solution planning module is used to match the corresponding execution plan according to the matching intention and related slot information;
[0034] The function execution module is used to execute the nodes configured in the scheme in sequence according to the steps of the execution scheme;
[0035] The display module is used to display the results after the plan is executed.
[0036] Preferably, the background configuration end includes a dialogue configuration module, a system configuration module, a function configuration module, an intention configuration module, a slot configuration module, an execution plan configuration module and a knowledge base configuration module;
[0037] The dialogue configuration module is used to define the parameters during the interaction between the user and the system;
[0038] The system configuration module is used to define the configuration of the system to be accessed, including the system name, code, URL, login script and system parameters;
[0039] Function configuration module, used to define the page corresponding to the system and the button information under the page, including the system to which the function belongs, the function attribution, the function type, the function execution script, the execution end, the function input parameter and the function output parameter;
[0040] The intent configuration module is used to define the interaction requirements that users need to perform, including the intent name, associated system, associated function, intent type, intent parsing method, and intent-associated slot;
[0041] The slot configuration module is used to define the keyword extraction definition of the user input information, including the slot name and slot extraction method;
[0042] The execution scheme configuration module is used to define the execution link configuration for the intent. The link contains each execution step and the execution action configuration of each step, including the intent ID associated with the execution scheme, the function node of the execution scheme, and the link of the function node;
[0043] The knowledge base configuration module is used to define the data required for the execution plan by importing it, and to vectorize the corresponding fields of the imported data as needed.
[0044] Preferably, the intent recognition module and the slot extraction module both include a regular expression parsing unit, a machine learning training model parsing unit and a large model prompt word parsing unit.
[0045] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a multi-system capability intelligent execution method based on human-computer interaction.
[0046] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a multi-system capability intelligent execution method and system based on human-computer interaction, which has the following advantages:
[0047] System level: The present invention decouples the modules of the system so that each module can complete its own function and provide module capabilities to the outside world. This can not only ensure the independent operation of the modules, but also enhance the operational efficiency of the entire system. The change of a certain module will not cause changes to the entire system, and it has good support for the scalability, robustness, reusability and portability of the system.
[0048] Intelligent level: Use intelligent algorithms and machine learning technologies to learn and analyze system functions and user intentions, identify user intentions and execute corresponding solutions more quickly and accurately, thereby improving the system's intelligence level and user experience;
[0049] Business level: It enables users to access and operate multiple systems on a single platform through natural language interaction. Users do not need to remember the address and access method of each system. They can quickly locate and access the system through simple interactive operations, which greatly simplifies the user's operation process and improves work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.
[0051] Figure 1 A flow chart of a multi-system capability intelligent execution method based on human-computer interaction provided by the present invention;
[0052] Figure 2 A schematic diagram of a multi-system capability intelligent execution system based on human-computer interaction provided by the present invention. DETAILED DESCRIPTION
[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0054] To clearly describe the embodiments of the present invention, some of the terms involved in the embodiments of the present invention are explained as follows:
[0055] Human-computer interaction: Users and the system operate through voice or text input.
[0056] System: various existing information systems, such as OA, CRM, etc.
[0057] Function: Pages and buttons in the system, such as Zhang San’s details page in the OA system, the process application button in the OA system, etc.
[0058] Intent: Users express what they want to do in natural language. For example, if a user wants to ask if there is a person named Zhang San in the OA system, the backend system needs to use the intent algorithm to understand which of the user's words and the configured intent content in the system best matches.
[0059] Slot: It is regarded as a variable and is used to accurately express the key information of user intention. Each slot is associated with one or more entity libraries to help the machine understand human intention from a semantic perspective. For example, if a user wants to ask if there is a person named Zhang San in the OA system, there are two slots in this intention, OA system and Zhang San.
[0060] Solution: It is defined as a pipeline consisting of a series of interrelated execution actions, which are organized in a specific order and logical relationship to achieve a specific goal or result. In this pipeline, each execution action plays a specific role and is responsible for completing a specific task. Importantly, these execution actions are not isolated, but interdependent. Specifically, the output of the previous action is usually used as the input of the next action, so that data and information can flow and pass throughout the pipeline. This input-output relationship ensures the consistency of information and the correctness of the process, so that the entire execution plan can proceed as expected.
[0061] Knowledge base: The data required for the execution plan is stored by importing, and the corresponding fields of the imported data are vectorized and stored as needed, which helps to achieve more efficient data processing and query. Through vectorized storage, the knowledge base can support queries based on semantics or contextual meanings, improving the accuracy and efficiency of data retrieval.
[0062] The embodiment of the present invention discloses a multi-system capability intelligent execution method based on human-computer interaction, comprising the following steps:
[0063] S1. Perform dialogue configuration, system configuration, function configuration, intent configuration, slot configuration, execution plan configuration and knowledge base configuration through the background according to the changes in customer needs;
[0064] S2. After the configuration is completed, the user inputs the action to be executed through human-computer interaction and converts it into text;
[0065] S3. Call the intent algorithm to match the most similar intent and the slot associated with the intent according to the intent configured by the system until all slots of the intent have been extracted;
[0066] S4. According to the matching intention and the relevant slot information, the matching intention matches the corresponding execution plan, matches the corresponding execution plan, and executes the nodes configured in the plan in sequence according to the steps of the execution plan;
[0067] S5. Display the results after the plan is executed.
[0068] In order to further implement the above technical solution, the specific content of step S1 is:
[0069] The dialog configuration is used to define the parameters during the interaction between the user and the system, such as the initial greeting, waiting time, number of rounds of dialog, etc.
[0070] System configuration is to complete the configuration of the system to be accessed, including the system name, code, URL, login script and system parameters;
[0071] Function configuration is to complete the page corresponding to the system and the button information under the page, including the system to which the function belongs, the function ownership, the function type (that is, the execution type of the function, such as SQL type, which needs to be executed by the backend SQL executor, JavaScript type, which needs to be executed by the front-end executor, API type, which needs to be executed by the back-end interface executor), function execution script, execution end (need to be executed by the backend or the front-end), function input parameters and function output parameters;
[0072] For example, define "Video System", configure the system URL, and no other parameters. Define the "Video Home Page" function, which belongs to the "Video System" system. The parent function of this function does not exist. Define the function type as javascript type, the execution end is the front end, and the function input parameter is the video playback URL. Then define a button "Video Search Button" function under this function, which belongs to the "Video System" system. The parent function of this function is "Video Home Page", define the function type as sql, the function execution script is an sql script, the execution end is the back end, the function input parameter is the video name, and the function output parameter is the video playback URL address.
[0073] Intent configuration is to complete the definition of the interaction requirements that users need to perform, including the intent name, associated system, associated function, intent type, intent parsing method, and intent-associated slot;
[0074] For example, define an intent named "Open Video" and associate it with the "Video Retrieval Open Button" function of the "Video System". The intent type is global intent, and the intent parsing method is divided into three scenarios. When the system is initialized, the machine learning training model, i.e., the small model, will be obtained through machine learning training of the intent model according to the configured intent data. If the intent may not be in the trained model when it is entered into the system later, it is necessary to use regular expressions or large model prompts to identify the intent. For example, regular expressions are used to obtain intent. For example, if a user mentions "I want to open the video of the west gate of Guangzhou Gymnasium", and the configured regular expression is ".*Open.*(video|surveillance).*", then the sentence "I want to open the video of the west gate of Guangzhou Gymnasium" can be matched with this regular expression to get True, which is the intent of "Open Video". If the result is False after matching with the regular expression, it is not this intent. If the large model prompt is used to match the intent "I want to open the video of Guangzhou Gymnasium", The prompt word can be written as "Please analyze the following text: "I want to open the video of the west gate of Guangzhou Gymnasium", and determine which of the following categories it belongs to. The categories are as follows: "Open the video to query the weather". If there is a classification, only output the classification. If not, please output -1, and do not output the explanation." After calculation by the large model, output "open the video" to get this intent; when configuring the intent, dynamically adjust the intent parsing method according to the needs. The global intent will be trained when the system is initialized, so you only need to enter the intent model code here. The backend calls the intent algorithm according to the intent code to obtain the intent, and finally configure the slot information of the intent. Here you need to configure the video name slot of the video. You can configure multiple slot information according to the needs of the intent. For example, when configuring the venue slot, you need to locate the specific video according to the venue and video name. The slot definition here is to configure the slot information according to the needs of specific data. Each docked system is not fixed;
[0075] The slot configuration is to complete the keyword extraction definition of the user input information, including the slot name and the slot extraction method;
[0076] For example, when defining the "video name" slot, the slot extraction method is similar to the intention. The model for extracting the relevant slot can be trained when initializing the system. When configuring the slot extraction method, you only need to configure the extraction model.
[0077] The execution scheme configuration is to complete the execution link configuration for the intent. The link contains each execution step and the execution action configuration of each step, including the intent ID associated with the execution scheme, the function node of the execution scheme, and the link of the function node;
[0078] For example, define an execution plan, associate it with the "open video" intention, configure the "video search button" function in the plan node, the sequence number is 1, the node CODE is unique, the input parameter is: video name, the output parameter is: video playback URL address, the plan input parameter value will be assigned to the configured function input parameter, and the function output parameter will be assigned to the plan output parameter. The first node is to find the URL address of the video playback. To complete the user's visual display, the second node needs to be configured to complete the video playback display. The second node configures the "video homepage" function, the sequence number is 2, the node CODE is unique, the input parameter is: video playback URL address, and the output parameter is none;
[0079] The knowledge base is configured to store the data required in the execution plan by importing, and vectorizes and stores the corresponding fields of the imported data as needed.
[0080] In order to further implement the above technical solution, the intent types include global intent and dynamic intent; the global intent is an intent type that is used by multiple systems, and the dynamic intent is an intent type for a specific system.
[0081] In order to further implement the above technical solutions, the methods of parsing intent and extracting slots include regular expression parsing, machine learning training model parsing and large model prompt word parsing.
[0082] In order to further implement the above technical solution, the solution node includes node CODE, node sequence number, function ID, input parameter and output parameter; the node link includes the previous node ID, the next node ID and the conditional expression, and the node link is executed if the conditional expression is met.
[0083] In order to further implement the above technical solution, in step S4, it also includes: if the configured solution is complex, in the process of interaction, the user is asked multiple times to complete the parameters required for the intention, and steps S2-S4 are repeated for multiple interactions to complete the action that the user needs to perform.
[0084] In another embodiment, taking the user initiating "open the monitoring of the west gate of Guangzhou Gymnasium" as an example, the front end converts the voice into text, and then passes the text to the back end. The back end calls the intention algorithm, first matches the intention according to the configured regular expression, if no match is found, matches the intention according to the small model, if still no match is found, matches according to the large model prompt word, if none of them matches, it prompts "Unable to recognize the user's intention!" If the "open video" intention is recognized, the slot associated with the intention continues to be extracted. The slot extraction is also extracted according to the regular expression, small model, and large model method. In this example, it is necessary to extract the "video name" slot, that is, "Guangzhou Gymnasium West Gate", and retrieve the video ID from the vector library according to the extracted slot, and then find the video playback address according to the video ID; then execute the second node of this scheme, and pass the video playback address found by the first node to the input parameter of the second node. The second node is the front-end execution script, and the back end sends a message to the front end to execute the script. The front end executes this script and plays this video. At this point, the user has completed the voice interaction operation of "opening the monitoring of the west gate of Guangzhou Gymnasium", and the operation of manually entering the video name to open the video is realized without jumping to the video system.
[0085] A multi-system capability intelligent execution system based on human-computer interaction, based on a multi-system capability intelligent execution method based on human-computer interaction, including a backend and a user end;
[0086] The backend includes the backend configuration terminal, intent recognition module, slot extraction module, knowledge base and solution planning module, and function execution module;
[0087] The user end includes a human-computer interaction interface, specifically an input module and a display module;
[0088] The backend configuration end is used to configure the dialogue, system, function, intent, slot, execution plan, and knowledge base according to the changing needs of customers;
[0089] The input module is used to allow users to input actions to be executed through human-computer interaction and convert them into text after the configuration is completed;
[0090] The intent recognition module is used to call the intent algorithm to match the most similar intent according to the intent configured by the system;
[0091] The slot extraction module is used to match the slots associated with the intent until all slots of the intent have been extracted;
[0092] The knowledge base is used to store the data required for the execution plan in an imported manner, and vectorize the corresponding fields of the imported data as needed;
[0093] A solution planning module is used to match the corresponding execution plan according to the matching intention and related slot information;
[0094] The function execution module is used to execute the nodes configured in the scheme in sequence according to the steps of the execution scheme;
[0095] The display module is used to display the results after the plan is executed.
[0096] In order to further implement the above technical solution, the background configuration end includes a dialogue configuration module, a system configuration module, a function configuration module, an intention configuration module, a slot configuration module, an execution plan configuration module and a knowledge base configuration module;
[0097] The dialogue configuration module is used to define the parameters during the interaction between the user and the system;
[0098] The system configuration module is used to define the configuration of the system to be accessed, including the system name, code, URL, login script and system parameters;
[0099] Function configuration module, used to define the page corresponding to the system and the button information under the page, including the system to which the function belongs, the function attribution, the function type, the function execution script, the execution end, the function input parameter and the function output parameter;
[0100] The intent configuration module is used to define the interaction requirements that users need to perform, including the intent name, associated system, associated function, intent type, intent parsing method, and intent-associated slot;
[0101] The slot configuration module is used to define the keyword extraction definition of the user input information, including the slot name and slot extraction method;
[0102] The execution scheme configuration module is used to define the execution link configuration for the intent. The link contains each execution step and the execution action configuration of each step, including the intent ID associated with the execution scheme, the function node of the execution scheme, and the link of the function node;
[0103] The knowledge base configuration module is used to define the data required for the execution plan by importing it, and to vectorize the corresponding fields of the imported data as needed.
[0104] In order to further implement the above technical solution, the intention recognition module and the slot extraction module both include a regular expression parsing unit, a small model parsing unit and a large model prompt word parsing unit.
[0105] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a multi-system capability intelligent execution method based on human-computer interaction.
[0106] In summary, the present invention proposes a multi-system capability intelligent execution method and system based on human-computer interaction, which aims to achieve unified management and efficient execution of multiple systems through intelligent means, thereby improving user work efficiency and experience. Specifically:
[0107] First, by building a unified interactive interface, users can access and operate multiple systems on one platform. Users do not need to remember the address and access method of each system. They can quickly locate and access the system through simple interactive operations, which greatly simplifies the user's operation process and improves work efficiency.
[0108] Secondly, by using intelligent algorithms and machine learning technology to learn and analyze user operational requirements, the system can accurately locate the systems and functions required by users by continuously accumulating and optimizing data models, thereby improving the system's intelligence level and user experience;
[0109] In addition, the present invention also has the ability of cross-system collaborative operation and data sharing. By building a unified data exchange and collaboration mechanism, the system can achieve data interoperability and collaborative work between different systems, which enables users to complete operations and data processing of multiple systems on one platform, avoiding the tedious process of frequent switching and repeated input between different systems.
[0110] To sum up, the present invention has important practical significance and application value. It can not only solve the management and execution problems faced by users when using multiple systems and improve work efficiency and accuracy, but also provide users with a more intelligent and personalized service experience. In the current era of informatization, networking, and intelligence, the application prospects are broad, and it will provide strong support for the digital transformation and intelligent upgrading of various fields.
[0111] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0112] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A multi-system capability intelligent execution method based on human-computer interaction, characterized in that: The following steps are involved: S1. Perform dialogue configuration, system configuration, function configuration, intent configuration, slot configuration, execution plan configuration and knowledge base configuration through the background according to the changes in customer needs; S2. After the configuration is completed, the user inputs the action to be executed through human-computer interaction and converts it into text; S3. Call the intent algorithm to match the most similar intent and the slot associated with the intent according to the intent configured by the system until all slots of the intent have been extracted; S4. According to the matching intention and the relevant slot information, the matching intention matches the corresponding execution plan, matches the corresponding execution plan, and executes the nodes configured in the plan in sequence according to the steps of the execution plan; S5. Display the results after the plan is executed.
2. According to claim 1, a multi-system capability intelligent execution method based on human-computer interaction is characterized in that: The specific content of step S1 is: The dialog configuration is used to define the parameters during the interaction between the user and the system; System configuration is to complete the configuration of the system to be accessed, including the system name, code, URL, login script and system parameters; Function configuration is to complete the page corresponding to the system and the button information under the page, including the system to which the function belongs, the function attribution, the function type, the function execution script, the execution end, the function input parameter and the function output parameter; Intent configuration is to complete the definition of the interaction requirements that users need to perform, including the intent name, associated system, associated function, intent type, intent parsing method, and intent-associated slot; The slot configuration is to complete the keyword extraction definition of the user input information, including the slot name and the slot extraction method; The execution scheme configuration is to complete the execution link configuration for the intent. The link contains each execution step and the execution action configuration of each step, including the intent ID associated with the execution scheme, the function node of the execution scheme, and the link of the function node; The knowledge base is configured to store the data required in the execution plan by importing, and vectorizes and stores the corresponding fields of the imported data as needed.
3. According to the method of multi-system capability intelligent execution based on human-computer interaction according to claim 2, it is characterized in that: Intent types include global intent and dynamic intent; global intent is an intent type that is used by multiple systems, and dynamic intent is an intent type for a specific system.
4. According to claim 2, a multi-system capability intelligent execution method based on human-computer interaction is characterized in that: The methods of parsing intent and extracting slots include regular expression parsing, machine learning training model parsing, and large model prompt word parsing.
5. According to the method of multi-system capability intelligent execution based on human-computer interaction in claim 2, it is characterized in that: The solution node includes node CODE, node sequence number, function ID, input parameters, and output parameters; the node link includes the previous node ID, the next node ID, and the conditional expression. The node link is executed if the conditional expression is met.
6. The method for intelligent execution of multi-system capabilities based on human-computer interaction according to claim 1, characterized in that: In step S4, it also includes: if the configuration scheme is complex, in the process of interaction, the user is asked multiple times to complete the parameters required for the intention, and steps S2-S4 are repeated for multiple interactions to complete the action that the user needs to perform.
7. A multi-system capability intelligent execution system based on human-computer interaction, characterized in that: A multi-system capability intelligent execution method based on human-computer interaction according to any one of claims 1 to 6, comprising a backend and a user end; The backend includes the backend configuration terminal, intent recognition module, slot extraction module, knowledge base and solution planning module, and function execution module; The user end includes a human-computer interaction interface, specifically an input module and a display module; The backend configuration end is used to configure the dialogue, system, function, intent, slot, execution plan, and knowledge base according to the changing needs of customers; The input module is used to allow users to input actions to be executed through human-computer interaction and convert them into text after the configuration is completed; The intent recognition module is used to call the intent algorithm to match the most similar intent according to the intent configured by the system; The slot extraction module is used to match the slots associated with the intent until all slots of the intent have been extracted; The knowledge base is used to store the data required for the execution plan in an imported manner, and vectorize the corresponding fields of the imported data as needed; A solution planning module is used to match the corresponding execution plan according to the matching intention and related slot information; The function execution module is used to execute the nodes configured in the scheme in sequence according to the steps of the execution scheme; The display module is used to display the results after the plan is executed.
8. The method for intelligent execution of multi-system capabilities based on human-computer interaction according to claim 7, characterized in that: The backend configuration end includes a dialogue configuration module, a system configuration module, a function configuration module, an intention configuration module, a slot configuration module, an execution plan configuration module, and a knowledge base configuration module; The dialogue configuration module is used to define the parameters during the interaction between the user and the system; The system configuration module is used to define the configuration of the system to be accessed, including the system name, code, URL, login script and system parameters; Function configuration module, used to define the page corresponding to the system and the button information under the page, including the system to which the function belongs, the function attribution, the function type, the function execution script, the execution end, the function input parameter and the function output parameter; The intent configuration module is used to define the interaction requirements that users need to perform, including the intent name, associated system, associated function, intent type, intent parsing method, and intent-associated slot; The slot configuration module is used to define the keyword extraction definition of the user input information, including the slot name and slot extraction method; The execution scheme configuration module is used to define the execution link configuration for the intent. The link contains each execution step and the execution action configuration of each step, including the intent ID associated with the execution scheme, the function node of the execution scheme, and the link of the function node; The knowledge base configuration module is used to define the data required for the execution plan to be stored by importing, and to vectorize the corresponding fields of the imported data as needed.
9. The method for intelligent execution of multi-system capabilities based on human-computer interaction according to claim 7, characterized in that: Both the intent recognition module and the slot extraction module include a regular expression parsing unit, a machine learning training model parsing unit and a large model prompt word parsing unit.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements a multi-system capability intelligent execution method based on human-computer interaction as described in any one of claims 1 to 6.
