Brain map dialogue strategy visualization method and device based on artificial intelligence

Abstract

The invention provides a brain graph dialogue strategy visualization method based on artificial intelligence, and the method comprises the steps: obtaining a standard dialogue process brain graph which comprises a plurality of nodes, a jump slot rule, a nuclear demand verbal skill rule and an activation verbal skill rule, analyzing the standard dialogue process brain graph, converting the standard dialogue process brain graph into an MVEL language expression, and storing the MVEL language expression in a database; recognizing a statement sent by a user based on a semantic recognition model to obtain a user intention and a keyword; performing traversal in the MVEL language expression based on the user intention and the keyword, and determining a hit strategy; and outputting a response statement based on the hit strategy. According to the scheme, by obtaining the standard dialogue process brain map, the original abstract dialogue strategy is clearly presented in the form of the graphical nodes and the hierarchical relationship, so that the logic association and the jump path among the rules are visual and visible, an operator can quickly understand the whole dialogue strategy architecture, visual operation is realized, and the operation efficiency is improved. And the visualization degree and usability of strategy management are effectively improved.

Description

Technical Field

[0001] This application belongs to the field of artificial intelligence, and in particular relates to a method and device for visualizing mind map dialogue strategies based on artificial intelligence. Background Technology

[0002] Driven by the digital wave, the service marketing industry is undergoing a profound transformation from being human-dominated to being driven by intelligence. As user interaction channels become more diversified across multiple platforms and customer needs become increasingly personalized, the industry's demand for standardized, automated, and intelligent dialogue processes is becoming increasingly urgent. The maturity of artificial intelligence technologies (such as natural language processing and large-scale models) provides the technological foundation for upgrading marketing services, prompting companies to seek solutions that deeply integrate AI capabilities with business processes to improve service efficiency, reduce labor costs, and achieve precise management of dialogue strategies.

[0003] The industry has already achieved initial implementation of cross-platform information integration and AI technology. On the one hand, intelligent customer service systems enable 24 / 7 automated response, significantly improving service coverage efficiency. On the other hand, through large-scale models in vertical domains, trained via private deployment, user intent recognition and tag extraction are achieved, and responses are generated in conjunction with preset SOP (Standard Operating Procedure) rules, replacing human intervention in standardized dialogues in some scenarios.

[0004] However, the current approach still has some problems. The main reason is that the rules are too abstract, making it difficult for business personnel to grasp the full picture. Customer service standard operating procedures (SOPs) are written in text, with scripts, strategies, and dialogue order mixed together. AI can only provide simple script polishing or strategy recommendations, but it cannot visualize the entire process, making it difficult for business personnel to understand the overall logic. While each rule seems independent, changing one part can affect the entire dialogue flow; for example, changing a redirect condition might disrupt the subsequent scripts. The dialogue strategy rules are all flat text descriptions, making it easy for customer service personnel to make mistakes when updating and maintaining them, resulting in low efficiency. Summary of the Invention

[0005] The purpose of this invention is to provide an artificial intelligence-based method for visualizing mind map dialogue strategies, aiming to solve the problems of easy errors and low efficiency in existing rule updates and maintenance. The artificial intelligence-based method for visualizing mind map dialogue strategies provided in this application includes:

[0006] The first aspect of this application provides a method for visualizing mind map dialogue strategies based on artificial intelligence, including:

[0007] Obtain a standard dialogue flow mind map, which is constructed through a hierarchical node architecture. The standard dialogue flow mind map includes multiple nodes, jump slot rules, core requirement dialogue rules, and activation dialogue rules. The multiple nodes can jump between each other through the jump slot rules, core requirement dialogue rules, and activation dialogue rules.

[0008] The standard dialogue flow mind map is parsed and converted into MVEL language expressions. The conversion process includes filtering invalid text, adding default rules, and replacing variable rules.

[0009] The semantic recognition model is used to identify the statements sent by the user and obtain the user's intent and keywords.

[0010] Based on the user intent and the keywords, the MVEL language expression is traversed to determine the hit strategy;

[0011] Output a response statement based on the hit strategy.

[0012] Based on the AI-based mind map dialogue strategy visualization method provided in the first aspect of the embodiments of this application, optionally,

[0013] The jump slot rules include rules for determining the next node. The jump slot rules prioritize the next node determined by the acquired keywords. If there are no acquired keywords, the next node is determined in order from left to right and from top to bottom.

[0014] Based on the AI-based mind map dialogue strategy visualization method provided in the first aspect of the embodiments of this application, optionally, the core requirement dialogue rule is the child node to which the jump slot rule is executed, and the core requirement dialogue rule is executed in order from left to right and from top to bottom.

[0015] Based on the AI-based mind map dialogue strategy visualization method provided in the first aspect of the embodiments of this application, optionally, the activation dialogue rule is the timeout sub-node to which the jump slot rule is executed, and the activation dialogue rule is executed in order from left to right and from top to bottom.

[0016] Based on the first aspect of the embodiments of this application, the AI-based mind map dialogue strategy visualization method can optionally include outputting response statements based on the hit strategy, which includes polishing the response statements before outputting them. The polishing effect is achieved by configuring multiple sets of semantically similar dialogue content for the same strategy and dynamically selecting the dialogue output according to the real-time dialogue scenario.

[0017] The real-time dialogue scenario includes at least one of the user's current intent, historical dialogue rounds, and slot filling status; the dynamic selection of dialogue includes sorting and selection based on preset weights or contextual similarity.

[0018] Based on the AI-based mind map dialogue strategy visualization method provided in the first aspect of the embodiments of this application, optionally, the step of parsing the standard dialogue flow mind map and converting it into an MVEL language expression includes:

[0019] The mind map is converted into a text path description. Invalid text and default rules are filtered out from the text path description, and pre-escape rules are generated. The pre-escape rules are then replaced with variable expressions to obtain MVEL language expressions.

[0020] Based on the AI-based mind map dialogue strategy visualization method provided in the first aspect of the embodiments of this application, optionally, the MVEL language expression is loaded through the easy rules rule engine framework.

[0021] A second aspect of this application provides an artificial intelligence-based mind map dialogue strategy visualization device, comprising:

[0022] The acquisition unit is used to acquire a standard dialogue flow mind map. The standard dialogue flow mind map is constructed through a hierarchical node architecture. The standard dialogue flow mind map includes multiple nodes, jump slot rules, core requirement dialogue rules, and activation dialogue rules. The multiple nodes can jump between each other through the jump slot rules, core requirement dialogue rules, and activation dialogue rules.

[0023] The parsing unit is used to parse the standard dialogue flow mind map and convert it into MVEL language expressions. The conversion process includes filtering invalid text, adding default rules, and replacing variable rules.

[0024] The recognition unit is used to recognize the statements sent by the user based on the semantic recognition model, and to obtain the user's intent and keywords;

[0025] The determining unit is used to traverse the MVEL language expression based on the user intent and the keywords to determine the hit strategy;

[0026] The output unit is used to output a response statement based on the hit strategy.

[0027] Optionally, the AI-based mind map dialogue strategy visualization device provided in the second aspect of the embodiments of this application may be configured as follows:

[0028] The jump slot rules include rules for determining the next node. The jump slot rules prioritize the next node determined by the acquired keywords. If there are no acquired keywords, the next node is determined in order from left to right and from top to bottom.

[0029] According to the AI-based mind map dialogue strategy visualization device provided in the second aspect of the present application, optionally, the core requirement dialogue rule is the child node to which the jump slot rule is executed, and the core requirement dialogue rule is executed in order from left to right and from top to bottom.

[0030] According to the AI-based mind map dialogue strategy visualization device provided in the second aspect of the present application, optionally, the activation dialogue rule is the timeout sub-node to which the jump slot rule is executed, and the activation dialogue rule is executed in order from left to right and from top to bottom.

[0031] According to the AI-based mind map dialogue strategy visualization device provided in the second aspect of the present application, the step of outputting a response statement based on the hit strategy includes polishing the response statement before outputting it. The polishing effect is achieved by configuring multiple sets of semantically similar dialogue content for the same strategy and dynamically selecting the dialogue output according to the real-time dialogue scenario.

[0032] The real-time dialogue scenario includes at least one of the user's current intent, historical dialogue rounds, and slot filling status; the dynamic selection of dialogue includes sorting and selection based on preset weights or contextual similarity.

[0033] According to the AI-based mind map dialogue strategy visualization device provided in the second aspect of the embodiments of this application, optionally, the step of parsing the standard dialogue flow mind map and converting it into an MVEL language expression includes:

[0034] The mind map is converted into a text path description. Invalid text and default rules are filtered out from the text path description, and pre-escape rules are generated. The pre-escape rules are then replaced with variable expressions to obtain MVEL language expressions.

[0035] According to the second aspect of the embodiments of this application, the AI-based mind map dialogue strategy visualization device may optionally load the MVEL language expression through the easy rules rule engine framework.

[0036] A third aspect of this application provides an artificial intelligence-based mind map dialogue strategy visualization device, comprising:

[0037] Central processing unit, memory, input / output interfaces, wired or wireless network interfaces, and power supply;

[0038] The memory is either a short-term storage memory or a persistent storage memory;

[0039] The central processing unit is configured to communicate with the memory and execute instructions in the memory on the device to perform the method described in any one of the first aspects of the embodiments of this application.

[0040] A fourth aspect of this application provides a computer-readable storage medium including instructions that, when executed on a computer, cause the computer to perform the method as described in any one of the first aspects of this application.

[0041] The fifth aspect of this application provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the method described in any one of the first aspects of this application.

[0042] As can be seen from the above technical solutions, the embodiments of this application have the following advantages: This application provides a mind map-based visualization method for dialogue strategies based on artificial intelligence, including: acquiring a standard dialogue flow mind map, wherein the standard dialogue flow mind map includes multiple nodes, jump slot rules, core requirement dialogue rules, and activation dialogue rules; parsing the standard dialogue flow mind map and converting it into an MVEL language expression, wherein the conversion process includes filtering invalid text, adding default rules, and replacing variable rules; recognizing the statements sent by the user based on a semantic recognition model to obtain the user intent and keywords; traversing the MVEL language expression based on the user intent and the keywords to determine the hit strategy; and outputting a response statement based on the hit strategy. This solution, by acquiring a standard dialogue flow mind map, clearly presents the originally abstract dialogue strategy with graphical nodes and hierarchical relationships, making the logical connections and jump paths between rules intuitively visible, facilitating operators to quickly understand the overall dialogue strategy architecture, achieving intuitive operation, and effectively improving the visualization and ease of use of strategy management. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort. It is understood that the drawings provided in this section are only for better understanding of this solution and do not constitute a limitation on this application.

[0044] Figure 1 A flowchart illustrating an embodiment of the AI-based mind map dialogue strategy visualization method provided in this application;

[0045] Figure 2 A schematic diagram of the mind map provided in this application;

[0046] Figure 3 A partial schematic diagram of the mind map provided in this application;

[0047] Figure 4 A timing flowchart of an embodiment of the AI-based mind map dialogue strategy visualization method provided in this application;

[0048] Figure 5 A schematic diagram of a structure of an embodiment of the AI-based mind map dialogue strategy visualization device provided in this application;

[0049] Figure 6 This is another structural schematic diagram of an embodiment of the AI-based mind map dialogue strategy visualization device provided in this application. Detailed Implementation

[0050] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application are clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application. At the same time, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0051] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0052] Driven by the digital wave, the service marketing industry is undergoing a profound transformation from being human-dominated to being driven by intelligence. As user interaction channels become more diversified across multiple platforms and customer needs become increasingly personalized, the industry's demand for standardized, automated, and intelligent dialogue processes is becoming increasingly urgent. The maturity of artificial intelligence technologies (such as natural language processing and large-scale models) provides the technological foundation for upgrading marketing services, prompting companies to seek solutions that deeply integrate AI capabilities with business processes to improve service efficiency, reduce labor costs, and achieve precise management of dialogue strategies.

[0053] The industry has already achieved initial implementation of cross-platform information integration and AI technology. On the one hand, intelligent customer service systems enable 24 / 7 automated response, significantly improving service coverage efficiency. On the other hand, through large-scale models in vertical domains, trained via private deployment, user intent recognition and tag extraction are achieved, and responses are generated in conjunction with preset SOP (Standard Operating Procedure) rules, replacing human intervention in standardized dialogues in some scenarios.

[0054] However, the current approach still has some problems. The main reason is that the rules are too abstract, making it difficult for business personnel to grasp the full picture. Customer service standard operating procedures (SOPs) are written in text, with scripts, strategies, and dialogue order mixed together. AI can only provide simple script polishing or strategy recommendations, but it cannot visualize the entire process, making it difficult for business personnel to understand the overall logic. While each rule seems independent, changing one part can affect the entire dialogue flow; for example, changing a redirect condition might disrupt the subsequent scripts. The dialogue strategy rules are all flat text descriptions, making it easy for customer service personnel to make mistakes when updating and maintaining them, resulting in low efficiency.

[0055] For solutions to the above problems, please refer to [link / reference]. Figure 1 An embodiment of the AI-based mind map dialogue strategy visualization method provided in this application includes steps 101-105.

[0056] 101. Obtain a mind map of the standard dialogue flow.

[0057] Obtain a standard dialogue flow mind map, which is constructed through a hierarchical node architecture. The standard dialogue flow mind map includes multiple nodes, jump slot rules, core requirement dialogue rules, and activation dialogue rules. The multiple nodes can jump between each other through the jump slot rules, core requirement dialogue rules, and activation dialogue rules.

[0058] This step involves constructing a dialogue strategy model for service marketing scenarios using a visual mind map. The mind map adopts a hierarchical structure, with second-level nodes representing strategy categories and sub-nodes representing specific execution conditions. For details, a mind map illustration provided in this application can be found here. Figure 2 The rules within the visualized mind map are divided into three categories: jump slot rules, core requirement dialogue rules, and activation dialogue rules. Different nodes are interconnected through jump slot rules, which control the jumps in the dialogue flow, including the brain... Figure 2The mind map consists of hierarchical rules for primary nodes and "reply" sub-node rules; core requirement dialogue rules for obtaining core user needs; and activation dialogue rules for handling situations where users do not reply within a timeout period. Nodes are linked by rule conditions, forming a hierarchical logic of "strategy goal → triggering condition → action." For example, when "slot - renovation purpose in (rental, owner-occupied)" is selected, the "next slot" jump strategy is triggered. Each node in the mind map contains clear jump rules, the corresponding slot information, and the dialogue to be triggered upon entering that node. This allows strategy designers to intuitively understand and modify the dialogue flow. This visualization method avoids the problems of abstract and difficult-to-maintain rules in traditional text descriptions, transforming complex SOP rules into graphical nodes, making the logical connections and jump paths between rules clearly visible. Simultaneously, the mind map supports parallel processing of multiple nodes, allowing different types of rules to work collaboratively within the same mind map to form a complete dialogue strategy system. In this way, the mind map provides a unified visual modeling platform for dialogue strategies, enabling strategy designers to fully control the dialogue flow and improve the efficiency and accuracy of strategy design. When modifying this mind map, users can clearly see the relationships between the various child nodes. In actual display, different colors or symbols can be used to display nodes at different levels. At the same time, when users select a specific node to modify, it can indicate which other related nodes are being modified, thereby improving usability. Users can modify them all at once.

[0059] 102. The standard dialogue flow mind map is parsed and converted into MVEL language expressions. The conversion process includes filtering invalid text, adding default rules, and replacing variable rules.

[0060] This step maps graphical rules to executable code. First, node path parsing is performed, extracting the complete path from the "Reply" node to the "Jump Strategy". For example, starting from the "Current Inquiry Slot = Decoration Purpose" node, it passes through nodes such as "Condition 2: Slot - Decoration Purpose: Not Empty" and "Condition 1: Multiple Intents in (QA Q&A)" to form a complete rule path. Then, invalid text filtering and rule appending are performed, removing notes and automatically adding default rules such as "Skip Slot Judgment". It's understood that default rules can be determined based on actual circumstances, and no restrictions are imposed here. Pre-escaping rules are then generated. For example, the rule containing notes, "Current Inquiry Slot = Decoration Purpose → Condition 2: Slot - Decoration Purpose: Not Empty → Condition 1: Multiple Intents in (QA Q&A) [This Slot + Non-Relevant Information]", is filtered to "(Skip Slot notin Decoration Purpose) and (Current Inquiry Slot = Decoration Purpose) and (Slot - Decoration Purpose: Not Empty) and (Multiple Intents in (QA Q&A))". Finally, variable substitution and expression generation are performed to convert natural language conditions into logical expressions supported by MVEL. For example, "slot-renovation purpose in (rental, owner-occupied)" is converted to "DECORATION_USE_VALUE.contains('rental')||DECORATION_USE_VALUE.contains('owner-occupied')", and "multiple intents in (QA / answer)" is converted to "currentIntentionList.contains('QA / answer')". The conversion process is automated based on a pre-edited program. For business personnel, they only need to send the edited mind map to the program, and the mind map rules will be automatically parsed and loaded. In actual implementation, users can also use other methods or means to convert the mind map, which are not limited here. This conversion process ensures that the visual rules in the mind map can be accurately converted into executable code, enabling the dialogue strategy to be executed automatically in the system. The converted MVEL expressions have a clear logical structure, can efficiently process user input, and perform corresponding jumps and responses according to preset rules. At the same time, this conversion method preserves the hierarchical structure and logical relationships of the mind map, enabling strategy designers to intuitively understand and modify the generated code, thus improving code maintainability.

[0061] 103. Based on the semantic recognition model, identify the statements sent by the user to obtain the user's intent and keywords.

[0062] This step achieves intelligent parsing of dialogue text through a private AI model. A large-scale model from a vertical domain is deployed privately, performing joint analysis based on customer service question text and user responses. When a user inputs text, the model first performs intent recognition, outputting a set of intent tags. For example, for a user inputting "120 square meters, haven't received the keys yet," the model might output the intent tag set "expressing no handover" or "replying with housing information," assigning these tags to the variable `currentIntentionList`. Simultaneously, the model extracts entity tags using NER technology, such as extracting "house area = 120 square meters" from the input and assigning it to the variable `SLOT_HOUSE_AREA`. The model also normalizes the extracted results, converting the user's natural language expression into standard tags that the system can recognize. For example, mapping "living alone" to "self-occupied" ensures consistency with mind map rules. This intelligent parsing method enables the system to accurately understand the user's intent and the information provided, providing a foundation for subsequent strategy matching and response generation. By combining the semantic understanding capabilities of a large model with entity recognition technology, the system can handle various complex user inputs, improving the accuracy and fluency of the dialogue. Meanwhile, private deployment ensures that the model can be optimized for specific domains, improving the accuracy of intent recognition and keyword extraction. The form and method of the semantic recognition model used in actual implementation can be determined according to the actual situation, and are not limited here.

[0063] 104. Based on the user intent and the keywords, traverse the MVEL language expression to determine the hit strategy.

[0064] This step executes MVEL expressions through a rules engine, following specific matching logic. Jump slot rules have priorities: rules that match child nodes are executed first, meaning rules that extract NERs and match the phrasing under the corresponding slot rule node are executed. If no match is found, rules at the second-level node are executed, all following a left-to-right and top-to-bottom order. When user input triggers the corresponding condition, the system iterates through the MVEL expressions to find a matching strategy. For example, when a user's reply triggers "Slot - Renovation Purpose: Empty" and "Multiple Intentions in (QA / Answer)," the corresponding MVEL expression "((!skipSlotList.contains('Renovation Purpose'))&&(DECORATION_USE_VALUE==”)&&(currentIntentionList.contains('QA / Answer')))" will be hit, thus triggering the "Renovation Purpose - Follow-up Question 1" strategy. The system also avoids logical contradictions through rule order; multiple conditions under the same node are checked one by one according to the mind map's order, ensuring a unique strategy is hit. This traversal and matching process ensures that the system can accurately select the appropriate strategy based on the user's intent and the provided information, allowing the dialogue to proceed according to the preset flow. The efficient execution capability of the rule engine enables the system to process a large number of rules and conditions in a short time, improving the dialogue's response speed. At the same time, clear priority and order rules avoid strategy conflicts, ensuring the consistency and reliability of the dialogue.

[0065] 105. Output a response statement based on the hit strategy.

[0066] This step enables strategy execution and dynamic script generation. Based on the hit strategy, the corresponding script template is invoked. Each strategy typically involves multiple sets of alternative scripts. For example, the "Renovation Purpose - Follow-up Question 1" strategy might include scripts such as "Are you considering living in it yourself? I'll find some examples for your reference" and "Are you planning to renovate to live in it yourself or rent it out? The designers you've assigned are different." The system will select the appropriate script to output based on the specific situation. The system also controls the sending interval according to the "delay time" parameter in the mind map node, simulating the rhythm of a human conversation. For example, when the node is set to "delay time = 5 seconds," the system will send the corresponding script at an appropriate time, making the conversation more natural and fluent. The system dynamically replaces variables using MVEL expressions and combines this with user history information to generate personalized responses. For example, replacing "${SLOT_HOUSE_AREA}" in the script with the actual house area provided by the user generates a personalized response like "A 120-square-meter house, are you considering living in it yourself or renting it out?" Finally, the system sends the final script through the client message interface to complete the conversation. This method of generating and outputting dialogue allows the system to provide diverse and personalized responses based on different strategies and user situations, improving the user experience. Dynamic variable substitution and personalized generation ensure that each dialogue is unique and better meets user needs. At the same time, delay control that simulates the rhythm of human conversation makes the dialogue more natural and reduces the robotic feel.

[0067] To facilitate the use of this method in practical implementation, this application also provides a more detailed embodiment that can be implemented optionally. Specifically, an example of the mind map used in this solution can be found here. Figure 3 The rules contained in the mind map will be explained in detail below.

[0068] In this solution, the mind map serves as a SOP (Standard Operating Procedure) description tool for consultant-user interactions. Its core value lies in visually representing complex dialogue strategies, analogous to the role of high-level programming languages ​​in software development, enabling precise definition of the dialogue flow. The strategy rules within the mind map are systematically divided into three categories: jump slot rules, core requirement dialogue rules, and activation dialogue rules. Each category has its own function, collectively constructing a complete dialogue strategy system.

[0069] Jump slot rules are the core of dialogue flow navigation; their main function is to determine the path for the dialogue to jump from the current node to the next node. These rules include two types: the first type corresponds to the brain... Figure 2The first type of rule forms the basic framework of the dialogue flow, providing macro-level guidance for the direction of the conversation. The second type focuses on subsequent nodes in the jump strategy nodes whose child nodes are "replies." These nodes begin with "Current Inquiry Slot = xx" and perform micro-level processing on the user's specific response. In terms of execution priority, the system prioritizes the second type of rule because it is a direct response to the user's immediate reply and can quickly adjust the dialogue flow based on user input. If the second type of rule does not match a suitable jump condition, the first type of rule is executed to ensure that the dialogue continues to proceed according to the preset basic framework. Regardless of the type of jump slot rule, the specific search and matching process follows a left-to-right and top-to-bottom order. This strict order ensures the consistency and predictability of rule execution and avoids confusion in the dialogue flow. Furthermore, the jump slot rule prioritizes determining the next node based on the acquired keywords. Only when keywords are missing will the node be determined according to the predetermined left-to-right and top-to-bottom order, thereby achieving efficient capture and response to key user information in the dialogue flow.

[0070] The core needs dialogue rules focus on the stage of acquiring users' core needs. It corresponds to the child node pointed to by the jump slot rule; specifically, it's the subsequent node in the jump strategy node whose child node is "core needs dialogue". In actual dialogue, when the jump slot rule triggers a child node containing "core needs dialogue", the system will execute these core needs dialogues sequentially from left to right and from top to bottom. For example, in a home renovation consultation scenario, after the jump slot rule determines to enter the "renovation purpose inquiry" stage, the core needs dialogue rules will, according to a preset order, first ask, "Are you renovating your house to live in yourself?" If the user does not respond clearly, it will continue to ask, "Are you considering renting it out or using it for office purposes?" Through this orderly dialogue guidance, the system gradually and accurately acquires the user's core needs information.

[0071] Activation script rules are primarily used to handle timeouts during conversations, maintaining dialogue continuity. They correspond to the timeout sub-node pointed to by the jump slot rule, specifically the node following the "timeout script" sub-node in the jump strategy node. When a user does not respond within the specified time, the system executes the corresponding timeout script according to the activation script rule, proceeding from left to right and top to bottom. For example, when a user inquires about product information and receives no response after a certain time, the system first sends "Are you still there? Feel free to ask me any questions!" If there is still no response, it then sends "Our products have a recent promotional offer; would you like to learn more?" These scripts reactivate the user's participation in the conversation, preventing interruptions and improving the user experience.

[0072] These three types of rules work together: the jump slot rules guide the direction of the dialogue flow, the core message rules are responsible for collecting core information, and the activation message rules ensure the dialogue continues. Together, they form an organic whole, enabling the mind map-based dialogue strategy to accurately and systematically output appropriate responses based on user input and dialogue status, achieving an efficient intelligent customer service and user interaction process.

[0073] Regarding the process of interpreting a mind map into MVEL language, taking the section on "Jump Strategy: Renovation Purposes" as an example, the mind map portion of this section is as follows: Figure 3 :

[0074] Mind map node paths

[0075] First, the mind map is converted into a text description. Based on the connections and conditional judgments of each node, the system outlines the paths formed by each node, creating a preliminary textual logic diagram, as follows: 1.

[0077] => Condition: Current inquiry slot = renovation purpose

[0078] => Note: This slot has replies.

[0079] => Condition 2: Slot - Decoration Purpose: Not empty

[0080] => Note: This slot + Q&A

[0081] => Condition 1: Multiple intentions in (QA / Answer)

[0082] => Condition 1: Slot - Renovation Purpose (Rental, Owner-occupied) 2.

[0084] => Condition: Current inquiry slot = renovation purpose

[0085] => Note: This slot has replies.

[0086] => Condition 2: Slot - Decoration Purpose: Not empty

[0087] => Note: This slot + Q&A

[0088] => Condition 1: Multiple intentions in (QA / Answer)

[0089] => Condition 2: Slot - Renovation purpose not in (rental, owner-occupied) 3.

[0091] => Condition: Current inquiry slot = renovation purpose

[0092] => Note: This slot has replies.

[0093] => Condition 2: Slot - Decoration Purpose: Not Empty

[0094] => Note: This slot [+non-related information] 4.

[0096] => Condition: Current inquiry slot = renovation purpose

[0097] => Note: This slot has replies.

[0098] => Condition 2: Slot - Decoration Purpose: Not Empty

[0099] => Note: This slot [+non-related information]

[0100] => Condition 3: Slot - Renovation purpose not in (rental, owner-occupied) "5.

[0101] => Condition: Current inquiry slot = renovation purpose

[0102] => Note: No reply to this slot

[0103] => Condition 3: Slot - Decoration Purpose: Empty

[0104] => Note: Q&A

[0105] => Condition 1: Multiple intentions in (QA / Answer) 6.

[0107] => Condition: Current inquiry slot = renovation purpose

[0108] => Note: No reply to this slot

[0109] => Condition 3: Slot - Decoration Purpose: Empty

[0110] => Note: Unrelated intent

[0111] => Condition 2: Multiple intentions in (expressing that one is busy at the moment) 7.

[0113] => Condition: Current inquiry slot = renovation purpose

[0114] => Note: No reply to this slot

[0115] => Condition 3: Slot - Decoration Purpose: Empty

[0116] => Note: Unrelated intent

[0117] => Condition 3: Intention = No obvious intention expressed" 8.

[0119] => Condition: Current inquiry slot = renovation purpose

[0120] => Note: No reply to this slot

[0121] => Condition 3: Slot - Decoration Purpose: Empty

[0122] => Note: Unrelated intent

[0123] => Condition 4: Intention = Expressing that there is no rush. 9.

[0125] => Condition: Current inquiry slot = renovation purpose

[0126] => Note: No reply to this slot

[0127] => Condition 3: Slot - Decoration Purpose: Empty

[0128] => Note: Unrelated intent

[0129] => Condition 4: Intention = Expression are both acceptable. 10.

[0131] => Condition: Current inquiry slot = renovation purpose

[0132] => Note: No reply to this slot

[0133] => Condition 3: Slot - Decoration Purpose: Empty

[0134] => Note: Unrelated intent

[0135] => Condition 5: Multiple intentions in (greetings, expressions of gratitude) 11.

[0137] => Condition: Current inquiry slot = renovation purpose

[0138] => Note: No reply to this slot

[0139] => Condition 3: Slot - Decoration Purpose: Empty

[0140] => Note: Unrelated intent

[0141] => Condition 6: Multiple intentions (affirmative answer, negative answer, expressing uncertainty, refusing service, expressing being out of town) 12.

[0143] => Condition: Current inquiry slot = renovation purpose

[0144] => Note: No reply to this slot

[0145] => Condition 3: Slot - Decoration Purpose: Empty

[0146] => Note: Unrelated intent

[0147] => Condition 9: Slot - Decoration Purpose: Empty

[0148] -2. Filter out invalid text and append default rules, and generate pre-escaped rules.

[0149] This step filters invalid text: remove auxiliary text such as "notes" and "explanations" from the mind map (e.g., "note: QA Q&A", only keep the core conditions), and add default rules: complete system-level general conditions (e.g., add "(skip slot not in decoration purpose)" to all rules to ensure that the current slot is not skipped).

[0150] 1. (Skip slot not in renovation purpose) and (Currently queried slot = renovation purpose) and (Slot - renovation purpose: not empty) and (Multiple intents in (QA / Answer)) and (Slot - renovation purpose in (rental, owner-occupied))

[0151] 2. (Skip slot not in renovation purpose) and (Currently queried slot = renovation purpose) and (Slot - renovation purpose: not empty) and (Multiple intents in (Q&A)) and (Slot - renovation purpose not in (rental, owner-occupied))

[0152] 3. (Skip slot not in renovation purpose) and (Currently querying slot = renovation purpose) and (Slot - renovation purpose: not empty) and (Slot - renovation purpose in (rental, owner-occupied))

[0153] 4. (Skip slot not in renovation purpose) and (Currently querying slot = renovation purpose) and (Slot - renovation purpose: not empty) and (Slot - renovation purpose not in (rental, owner-occupied))

[0154] 5. (Skip slot not in decoration purpose) and (Currently querying slot = decoration purpose) and (Slot - decoration purpose: empty) and (Multiple intents in (QA / Answer))

[0155] 6. (Skip slot not in decoration purpose) and (Currently querying slot = decoration purpose) and (Slot - decoration purpose: empty) and (Multiple intents in (expressing currently busy))

[0156] 7. (Skip slot not in decoration purpose) and (Currently querying slot = decoration purpose) and (Slot - decoration purpose: empty) and (Intent = expressing no obvious intent)

[0157] 8. (Skip slot not in decoration purpose) and (Currently querying slot = decoration purpose) and (Slot - decoration purpose: empty) and (Intent = expressing no rush)

[0158] 9. (Skip slot not in decoration purpose) and (Currently querying slot = decoration purpose) and (Slot - decoration purpose: empty) and (Intent = expression are both acceptable)

[0159] 10. (Skip slot not in decoration purpose) and (Currently querying slot = decoration purpose) and (Slot - decoration purpose: empty) and (Multiple intents in (greeting, expressing gratitude))

[0160] 11. (Skip slot not in decoration purpose) and (Currently queried slot = decoration purpose) and (Slot - decoration purpose: empty) and (Multiple intents in (positive answer, negative answer, expressing uncertainty, refusing service, expressing being out of town))

[0161] 12. (Skip slot not in decoration purpose) and (Currently querying slot = decoration purpose) and (Slot - decoration purpose: empty) and (Slot - decoration purpose: empty)

[0162] -3. Replace the variables in the pre-escape rules and interpret the rules as MVEL language.

[0163] Rule conditions are essentially transforming generic rule expressions into MVEL code that the rule engine (EasyRules) can run.

[0164] 1. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE!=”)&&((currentIntentionList.contains('QA Q&A'))) &&((DECORATION_USE_VALUE.contains('Rental') ||DECORATION_USE_VALUE.contains('Ownership')))

[0165] 2. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE!=")&&((currentIntentionList.contains('QA Q&A'))) &&((!DECORATION_USE_VALUE.contains('Rental')&&!DECORATION_USE_VALUE.contains('Ownership')))

[0166] 3. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE!=")&&((DECORATION_USE_VALUE.contains('Rental')||DECORATION_USE_VALUE.contains('Ownership')))

[0167] 4. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE!=")&&((!DECORATION_USE_VALUE.contains('Rental')&&!DECORATION_USE_VALUE.contains('Ownership')))

[0168] 5. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE==”)&&((currentIntentionList.contains('QA Q&A')))

[0169] 6. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE==”)&&((currentIntentionList.contains('Expressing I'm Busy Right Now')))

[0170] 7. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE==")&&(currentIntention=='Expressing no obvious intention')

[0171] 8. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE==”)&&(currentIntention=='Expressing No Urgency')

[0172] 9. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE==”)&&(currentIntention=='Can be expressed as either')

[0173] 10. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE==”)&&((currentIntentionList.contains('Greeting')||currentIntentionList.contains('Expressing Gratitude')))

[0174] 11. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE==”)&&((currentIntentionList.contains('Affirmative Answer')||currentIntentionList.contains('Negative Answer')||currentIntentionList.contains('Uncertainty')||currentIntentionList.contains('Refusal of Service')||currentIntentionList.contains('Expressing Indecisiveness')||currentIntentionList.contains('Refusal of Service')||currentIntentionList.contains('Expressing Being Out of Town')))

[0175] 12. ((!skipSlotList.contains('Renovation Purpose')))&&(currentAskSlot=='Renovation Purpose')&&(DECORATION_USE_VALUE==”)&&(DECORATION_USE_VALUE==”)

[0176] The following describes the overall execution process of this plan. The execution flow of this plan can be found by referring to... Figure 4 , Figure 4 This is a flowchart illustrating the implementation process of this solution.

[0177] This includes: The mind map policy rule engine initializes and loads policy rules into memory.

[0178] When a micro-message receives a new message from a user, it is passed on to the intelligent response system.

[0179] The intelligent response system marks a message as a current message to be saved; if there is a message to be sent, it will be discarded.

[0180] The intelligent response system invokes a model, triggering the intent recognition / NER extraction / QA process.

[0181] The model completes the analysis and returns the extracted information to the intelligent response system.

[0182] The intelligent response system calls the mind map strategy rule engine to perform retrieval and slot filling.

[0183] The intelligent response system initiates custom assignment and strategy lookup to the mind map strategy rule engine.

[0184] The mind map strategy rule engine initializes variables, searches for and returns strategy text / materials according to rules.

[0185] The strategy is sent to the delay queue (containing text / materials) and enters the "waiting for the delay time to expire" stage.

[0186] The delayed queue triggers message consumption, executing the process of "checking managed status → checking event validity → other checks".

[0187] After the inspection is completed, a delayed message (text / material) is sent to the user, and the user receives the message.

[0188] Trigger the delay hook event and repeat the process of "waiting for the delay time to expire → consuming queue messages → performing status / event checks".

[0189] Calling the smart null value / retry function triggers the process of "calling the new strategy → searching for rules according to the timeout strategy → returning the latest strategy", which realizes the strategy looping or updating.

[0190] The above process is described in detail below. During the service startup phase, the system first initializes the policy rules. Based on the Easy Rules rule engine framework, the system reads and loads the MVEL expressions in the "Rule Conditions" column of the file "Current Inquiry Slot = Decoration Purpose.xlsx". These expressions store the logic of the dialogue strategy in a coded form, covering rules for redirecting to specific slots, rules for required dialogue phrases, etc., laying the foundation for the automated execution of subsequent dialogue processes. At this time, the rule engine acts like a command center on standby, waiting for user input to trigger corresponding instructions.

[0191] When a user initiates an interaction (such as through a chat, adding someone on WeChat, or clicking a link), the dialogue process officially begins. The system immediately invokes a private model, including a large model, a small model, and an agent, to perform semantic parsing of the user's input. The model takes the consultant's question text and the user's response as input, combining pre-defined intents and label ranges to identify the user's expressed intent and extract relevant labels. For example, after the consultant asks, "Are you planning to live in this renovation yourself or rent it out? (The quote is different from the designer's~)," the user replies, "120 square meters, haven't received the keys yet, what's the difference between these two?" After processing, the model outputs two key results: In terms of intent recognition, it returns two intents: "expressing no handover of the property" and "replying with property information," which the system assigns to the [Multi-intent] variable in the MVEL expression, i.e., "Multi-intent in (expressing no handover of the property, QA question and answer)"; in terms of named entity recognition (NER), it extracts the label "property area," corresponding to the value "120 square meters," normalizes it, and assigns it to the [slot-property area] variable, making its value "120." These results become the core basis for subsequent rule matching.

[0192] After obtaining the user's semantic parsing results, the system returns a rule system composed of a mind map (i.e., MVEL expressions). In addition to the intent recognition and NER results returned by the large model, it also incorporates relevant system parameters, such as "current query slot = decoration purpose", "slot - decoration purpose: empty", and "multiple intents in (QA question and answer)". These parameters, like puzzle pieces, together form the complete conditions for the rule engine to operate, providing data support for accurately matching dialogue strategies.

[0193] Subsequently, the rule engine initiates the execution process. It combines the loaded MVEL expression with the constructed parameters and executes the condition judgments sequentially according to a predetermined order. In this case, based on the rule logic in the file "Current Inquiry Slot = Renovation Purpose.xlsx", the system compares and calculates the parameters, ultimately hitting rule 5. The condition expression for this rule is "(Skip Slot notin Renovation Purpose) and (Current Inquiry Slot = Renovation Purpose) and (Slot - Renovation Purpose: Empty) and (Multiple Intents in (QA Question and Answer))". After hitting the rule, the jump strategy is determined to be "Specified Jump", and the next question slot is "Renovation Purpose - Follow-up Question 1". This process is like finding the only correct exit in a complex maze based on clues, ensuring that the dialogue process proceeds according to the preset strategy.

[0194] Finally, the system enters the strategy execution phase. Based on the hit strategy "Renovation Purpose - Follow-up Question 1", the system filters from a pre-configured similar dialogue library to achieve fluency and diversity in the dialogue. By loading the MVEL expression corresponding to the required dialogue and substituting the previously extracted parameters, the system accurately matches the appropriate dialogue content. For example, it might select a dialogue like, "A 120-square-meter house is very spacious! After the renovation, will you mainly live there yourself, or do you plan to rent it out?" After determining the dialogue, the system calls the client message interface to send the generated response to the user, completing a full dialogue interaction. In actual implementation, the system also includes refining the response statement before outputting it. The dialogue refinement effect is achieved through the following methods: configuring multiple sets of semantically similar dialogue content for the same strategy, and dynamically selecting the dialogue output according to the real-time dialogue scenario; the real-time dialogue scenario includes at least one of the user's current intent, historical dialogue rounds, and slot filling status; the dynamic selection of dialogue includes sorting and selection based on preset weights or contextual similarity.

[0195] The foregoing description describes the AI-based mind map dialogue strategy visualization method provided in this application. To support the implementation of the above embodiments, this application also provides an AI-based mind map dialogue strategy visualization device. Please refer to [link to relevant documentation]. Figure 5 One embodiment of the AI-based mind map dialogue strategy visualization device provided in this application includes:

[0196] The acquisition unit 501 is used to acquire a standard dialogue flow mind map. The standard dialogue flow mind map is constructed through a hierarchical node architecture. The standard dialogue flow mind map includes multiple nodes, jump slot rules, core requirement dialogue rules, and activation dialogue rules. The multiple nodes can jump between each other through the jump slot rules, core requirement dialogue rules, and activation dialogue rules.

[0197] Parsing unit 502 is used to parse the standard dialogue flow mind map and convert it into MVEL language expressions. The conversion process includes filtering invalid text, adding default rules, and replacing variable rules.

[0198] The recognition unit 503 is used to recognize the statements sent by the user based on the semantic recognition model to obtain the user's intent and keywords;

[0199] The determining unit 504 is used to traverse the MVEL language expression based on the user intent and the keywords to determine the hit strategy;

[0200] Output unit 505 is used to output a response statement based on the hit strategy.

[0201] Optionally, the jump slot rule includes a rule for determining the next node. The jump slot rule prioritizes the next node determined by the acquired keyword. If there is no acquired keyword, the next node is determined in order from left to right and from top to bottom.

[0202] Optionally, the core requirement dialogue rule is the child node to which the jump slot rule jumps, and the core requirement dialogue rule is executed in order from left to right and from top to bottom.

[0203] Optionally, the activation dialogue rule is the timeout sub-node to which the jump slot rule is executed, and the activation dialogue rule is executed in order from left to right and from top to bottom.

[0204] Optionally, the step of outputting the response statement based on the hit strategy includes refining the response statement before outputting it. The refinement effect is achieved by configuring multiple sets of semantically similar dialogue content for the same strategy and dynamically selecting the dialogue output according to the real-time dialogue scenario.

[0205] The real-time dialogue scenario includes at least one of the user's current intent, historical dialogue rounds, and slot filling status; the dynamic selection of dialogue includes sorting and selection based on preset weights or contextual similarity.

[0206] Optionally, parsing the standard dialogue flow mind map and converting it into MVEL language expressions includes:

[0207] The mind map is converted into a text path description. Invalid text and default rules are filtered out from the text path description, and pre-escape rules are generated. The pre-escape rules are then replaced with variable expressions to obtain MVEL language expressions.

[0208] Optionally, the MVEL language expression is loaded via the easy rules rule engine framework.

[0209] In this embodiment, the processes executed by each unit in the device are the same as those described above. Figure 1 The method flow described in the corresponding embodiments is similar and will not be repeated here.

[0210] Figure 6 This is a schematic diagram of the structure of a mind map dialogue strategy visualization device server based on artificial intelligence provided in an embodiment of this application. The server 600 may include one or more central processing units (CPUs) 601 and a memory 605, in which one or more applications or data are stored.

[0211] In this embodiment, the specific functional module division in the central processing unit 601 can be the same as described above. Figure 6 The functional module division of each unit described in the text is similar, so it will not be repeated here.

[0212] The memory 605 can be volatile or persistent storage. The program stored in the memory 605 can include one or more modules, each module including a series of instruction operations on the server. Furthermore, the central processing unit 601 can be configured to communicate with the memory 605 and execute the series of instruction operations stored in the memory 605 on the server 600.

[0213] Server 600 may also include one or more power supplies 602, one or more wired or wireless network interfaces 603, one or more input / output interfaces 604, and / or one or more operating systems, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, etc.

[0214] The central processing unit 401 can perform the aforementioned... Figure 1 The specific operations performed by the AI-based mind map dialogue strategy visualization method in the illustrated embodiment will not be described in detail here.

[0215] This application also provides a computer storage medium for storing computer software instructions used for the above-described AI-based mind map dialogue strategy visualization method, including programs designed for executing the AI-based mind map dialogue strategy visualization method.

[0216] This AI-based mind map dialogue strategy visualization method can be as described above. Figure 1 The method for visualizing mind map dialogue strategies based on artificial intelligence, as described in the paper.

[0217] This application also provides a computer program product, which includes computer software instructions that can be loaded by a processor to implement the above-described functionality. Figure 1 The process of any one of the following AI-based mind map dialogue strategy visualization methods.

[0218] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, equivalent circuit transformations and unit divisions are only logical functional divisions. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connections shown or discussed may be through some interfaces, or indirect coupling or communication connections between apparatuses or units, and may be electrical, mechanical, or other forms.

[0219] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0220] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0221] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions or improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for visualizing mind map dialogue strategies based on artificial intelligence, characterized in that, include: Obtain a standard dialogue flow mind map, which is constructed through a hierarchical node architecture. The standard dialogue flow mind map includes multiple nodes, jump slot rules, core requirement dialogue rules, and activation dialogue rules. The multiple nodes can jump between each other through the jump slot rules, core requirement dialogue rules, and activation dialogue rules. The standard dialogue flow mind map is parsed and converted into MVEL language expressions. The conversion process includes filtering invalid text, adding default rules, and replacing variable rules. The semantic recognition model is used to identify the statements sent by the user and obtain the user's intent and keywords. Based on the user intent and the keywords, the MVEL language expression is traversed to determine the hit strategy; Output a response statement based on the hit strategy.

2. The artificial intelligence-based mind map dialogue strategy visualization method according to claim 1, characterized in that, The jump slot rules include rules for determining the next node. The jump slot rules prioritize the next node determined by the acquired keywords. If there are no acquired keywords, the next node is determined in order from left to right and from top to bottom.

3. The artificial intelligence-based mind map dialogue strategy visualization method according to claim 1, characterized in that, The core requirement dialogue rules are the child nodes to which the jump slot rules are executed, and the core requirement dialogue rules are executed in order from left to right and from top to bottom.

4. The artificial intelligence-based mind map dialogue strategy visualization method according to claim 1, characterized in that, The activation script rule is the timeout sub-node to which the jump slot rule jumps, and the activation script rule is executed in order from left to right and from top to bottom.

5. The artificial intelligence-based mind map dialogue strategy visualization method according to claim 1, characterized in that, The step of outputting response statements based on the hit strategy includes: refining the response statements before outputting them, and the refinement effect includes: configuring multiple sets of semantically similar dialogue content for the same strategy, and dynamically selecting the dialogue output according to the real-time dialogue scenario; The real-time dialogue scenario includes at least one of the user's current intent, historical dialogue rounds, and slot filling status; the dynamic selection of dialogue includes sorting and selection based on preset weights or contextual similarity.

6. The artificial intelligence-based mind map dialogue strategy visualization method according to claim 1, characterized in that, The parsing of the standard dialogue flow mind map and its conversion into MVEL language expressions includes: The mind map is converted into a text path description. Invalid text and default rules are filtered out from the text path description, and pre-escape rules are generated. The pre-escape rules are then replaced with variable expressions to obtain MVEL language expressions.

7. The artificial intelligence-based mind map dialogue strategy visualization method according to claim 1, characterized in that, The MVEL language expression is loaded through the easy rules rule engine framework.

8. A mind map dialogue strategy visualization device based on artificial intelligence, characterized in that, include: The acquisition unit is used to acquire a standard dialogue flow mind map. The standard dialogue flow mind map is constructed through a hierarchical node architecture. The standard dialogue flow mind map includes multiple nodes, jump slot rules, core requirement dialogue rules, and activation dialogue rules. The multiple nodes can jump between each other through the jump slot rules, core requirement dialogue rules, and activation dialogue rules. The parsing unit is used to parse the standard dialogue flow mind map and convert it into MVEL language expressions. The conversion process includes filtering invalid text, adding default rules, and replacing variable rules. The recognition unit is used to recognize the statements sent by the user based on the semantic recognition model, and to obtain the user's intent and keywords; The determining unit is used to traverse the MVEL language expression based on the user intent and the keywords to determine the hit strategy; The output unit is used to output a response statement based on the hit strategy.

9. A mind map dialogue strategy visualization device based on artificial intelligence, characterized in that, include: Central processing unit, memory, input / output interfaces, wired or wireless network interfaces, and power supply; The memory is either a short-term storage memory or a persistent storage memory; The central processing unit is configured to communicate with the memory and execute instructions in the memory on the device to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Includes instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 7.

Images