Intelligent Response Method, System and Storage Medium Based on Large Model

Abstract

An embodiment of this specification discloses an intelligent response method, system, and storage medium based on a large model. Among them, the system includes a user input module for obtaining user input information; a keyword extraction module for extracting key information corresponding to the input information through the large model; a historical record joint analysis module for determining several dimensional features of the user according to the key information and historical record information; a permission judgment module for determining the knowledge base to which the user has access rights according to the permission level; a comprehensive portrait establishment module for generating a dynamic user portrait according to several dimensional features of the user; and a knowledge recommendation module for outputting a personalized response for the user according to the dynamic user portrait and the knowledge base to which the user has access rights. The embodiment of this specification constructs an accurate user portrait through multi-dimensional analysis of user behavior and historical data, accurately matches user needs, provides personalized knowledge recommendations, and significantly improves the accuracy of knowledge services.

Description

Technical Field

[0001] One or more embodiments of this specification relate to the technical field of intelligent response, and specifically relate to an intelligent response method, system, and storage medium based on a large model. Background Art

[0002] With the development of Internet e-commerce, conventional manual customer service can no longer meet the service requirements. Existing customer service systems based on keyword replies mainly achieve fully automated interaction through a preset rule engine. That is, after the user inputs information, the system extracts keywords, matches the standard answers in the knowledge base, and returns the results in real time. However, the current customer service systems still have limitations in many aspects. For example, the semantic generalization ability is insufficient: it only supports literal keyword matching and cannot recognize synonyms or composite intentions, resulting in misjudgment of requests. Another example is the problem of context fragmentation: historical interactions cannot be associated in multi-round conversations, and keywords need to be triggered repeatedly. Another example is the high maintenance cost: the keyword library needs to be updated manually regularly, and the rule complexity increases exponentially with business expansion. Another example is the limitation of the user experience: mechanical replies are likely to cause dissatisfaction among users, and complex problems require multiple keyword probes and cannot provide personalized guidance. Therefore, there is an urgent need for an intelligent response system based on a large model that can provide personalized guidance to users. Summary of the Invention

[0003] Embodiments of this specification provide an intelligent response method, system, and storage medium based on a large model, and the technical solutions are as follows:

[0004] In a first aspect, an embodiment of this specification provides an intelligent response system based on a large model, including: a user input module for obtaining the input information of the user; a keyword extraction module for extracting the key information corresponding to the input information through a large model; a historical record joint analysis module for obtaining the historical record information of the user and determining several dimensional features of the user according to the key information and the historical record information; the several dimensional features include position information, basic attributes, interaction behaviors, skill tags, knowledge shortboards, and decision-making styles; an authority judgment module for determining the authority level of the user according to the position information of the user and determining the knowledge base to which the user has access rights according to the authority level; a comprehensive portrait establishment module for generating a dynamic user portrait according to the several dimensional features of the user; a knowledge recommendation module for outputting a personalized reply for the user according to the dynamic user portrait and the knowledge base to which the user has access rights.

[0005] Second aspect, the embodiments of this specification provide an intelligent response method based on a large model, including: obtaining the input information of a user; extracting key information corresponding to the input information through the large model; obtaining the historical record information of the user, and determining several dimensional features of the user according to the key information and the historical record information; the several dimensional features include position information, basic attributes, interaction behaviors, skill tags, knowledge gaps, and decision-making styles; determining the permission level of the user according to the position information of the user, and determining the knowledge base to which the user has access rights according to the permission level; generating a dynamic user profile according to the several dimensional features of the user; and outputting a personalized response for the user according to the dynamic user profile and the knowledge base to which the user has access rights.

[0006] Third aspect, the embodiments of this specification provide an electronic device, including a processor and a memory; the processor is connected to the memory; the memory is used for storing executable program code; the processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the steps of the intelligent response method based on the large model in the second aspect of the above embodiments.

[0007] Fourth aspect, the embodiments of this specification provide a computer storage medium, which stores multiple instructions, and the instructions are suitable for being loaded and executed by a processor to execute the steps of the intelligent response method based on the large model in the second aspect of the above embodiments.

[0008] The beneficial effects brought by the technical solutions provided by some embodiments of this specification at least include:

[0009] The embodiments of this specification can analyze user behaviors and historical data according to multi-dimensional information such as the historical records, interaction behaviors, and scenario tags of users, construct accurate user profiles, accurately match user needs, provide personalized knowledge recommendations, and significantly improve the accuracy of knowledge services; the embodiments of this specification can also provide decision-making support according to the comprehensive user profile. By analyzing the decision-making style and context correlation of users, it is possible to predict the possible decision-making needs of users, prepare relevant support information in advance, thereby shortening the decision-making time and effectively improving the decision-making efficiency and quality; the embodiments of this specification can also identify the knowledge gaps and learning preferences of users based on the historical learning records of users, recommend relevant courses or documents, and generate ability improvement suggestions in combination with job requirements to help employees improve their skills targeted, form a personalized learning path, and assist in the career development of employees; the embodiments of this specification can also screen and recommend knowledge by analyzing information such as permission levels and content preferences to ensure that users can access the knowledge base content that matches their responsibilities and interests. While avoiding information overload, it enables users to quickly locate the information they need and improves the utilization efficiency of the knowledge base. Description of the Drawings

[0010] To more clearly illustrate the technical solutions in the embodiments of this specification, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of this specification. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0011] Figure 1 It is a schematic diagram of an application scenario of an intelligent response system based on a large model provided by this specification.

[0012] Figure 2 It is a schematic diagram of the structure of an intelligent response system based on a large model provided by this specification.

[0013] Figure 3 It is a schematic diagram of the structure of the historical record joint analysis module provided by this specification.

[0014] Figure 4 It is a schematic diagram of the process for generating scenario-based recommendations provided by this specification.

[0015] Figure 5 It is a schematic diagram of the process of an intelligent response method based on a large model provided by this specification.

[0016] Figure 6 It is a schematic diagram of the structure of an electronic device provided by this specification. Detailed implementation manners

[0017] The following will clearly and completely describe the technical solutions in the embodiments of this specification in conjunction with the drawings in the embodiments of this specification.

[0018] Terms such as "first", "second", etc. in the specification, claims, and the above drawings of this specification are used to distinguish different objects, rather than to describe a specific order. In addition, the term "including" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0019] The intelligent response system based on a large model provided by multiple embodiments of this specification, and the execution subject of this intelligent response system based on a large model can be the intelligent response system based on a large model provided by the embodiments of the present invention.

[0020] Before elaborating on the intelligent response system based on a large model in detail in combination with one or more embodiments, this specification first introduces the application scenarios of this intelligent response system based on a large model.

[0021] Please refer to Figure 1 , Figure 1 which is a schematic diagram of an application scenario of an intelligent answering system based on a large model provided by an embodiment of the present invention. In this embodiment, the intelligent answering system 100 based on a large model may include a server 110 and a plurality of terminals 120, and the plurality of terminals 120 are respectively communicatively connected to the server 110.

[0022] In this embodiment of the specification, the terminal 120 may be a device such as a mobile phone, a tablet computer, an intelligent Bluetooth device, a notebook computer, or a personal computer (PC). The terminal 120 may send the current input information of the user and the historical input information of the user, etc. to the server 110. The terminal 120 includes a central processing unit (CPU), a graphics processing unit (GPU), a memory, a storage device, a network communication module, sensors, a display screen, a battery, and a power management module, etc. The central processing unit may execute logical calculations, resource scheduling, and rendering optimization algorithms of the front-end application, etc.; the graphics processing unit may be used to accelerate the rendering of the front-end page, especially the rendering of complex graphics, animations, and multimedia content; the memory may be used to store the runtime data, resource files, user behavior data, and environment perception data of the front-end application, etc.; the storage device may be used to store the code, resource files, and cache data of the front-end application, etc.; the network communication module may be used to communicate with the server to obtain dynamic resources and real-time environment information, etc.; the sensors may be used to sense the environmental state of the user device in real time. For example, the network sensor is used to obtain network bandwidth, latency, and connection status, etc. Another example is that the performance sensor is used to obtain the CPU, GPU, and memory usage of the terminal 120, etc. Another example is that the screen sensor can obtain the screen resolution, brightness, and refresh rate, etc.; the display screen may be used to present the content of the front-end application; the battery and power management module may provide power support for the operation of the device and optimize energy consumption.

[0023] The server 110 in this embodiment of the specification may be a single server or a server cluster composed of multiple servers, and the intelligent answering system based on a large model of the present application is implemented by multiple servers.

[0024] The server 110 in the embodiments of this specification may include a user input module, a keyword extraction module, a historical record joint analysis module, a permission judgment module, a comprehensive profile building module, a knowledge recommendation module, etc. The user input module is used to obtain the input information of the user; the keyword extraction module is used to extract the key information corresponding to the input information through a large model; the historical record joint analysis module is used to obtain the historical record information of the user, and determine several dimensional features of the user according to the key information and the historical record information; the several dimensional features include position information, basic attributes, interaction behaviors, skill tags, knowledge deficiencies, and decision-making styles; the permission judgment module is used to determine the permission level of the user according to the position information of the user, and determine the knowledge base to which the user has access rights according to the permission level; the comprehensive profile building module is used to generate a dynamic user profile according to the several dimensional features of the user; the knowledge recommendation module is used to output a personalized response for the user according to the dynamic user profile and the knowledge base to which the user has access rights.

[0025] It should be noted that Figure 1 The scenario schematic diagram of the intelligent answering system 100 based on a large model shown is only an example. The intelligent answering system and scenario described in the embodiments of the present invention are for more clearly illustrating the technical solutions of the embodiments of the present invention, and do not constitute a limitation on the technical solutions provided by the embodiments of the present invention. Those of ordinary skill in the art know that with the evolution of the intelligent answering system based on a large model and the emergence of new scenarios, the technical solutions provided by the embodiments of the present invention are equally applicable to similar technical problems.

[0026] Please refer to Figure 2 , Figure 2 is the structural schematic diagram of the intelligent answering system based on a large model provided by the embodiments of the present invention. This intelligent answering system based on a large model can be composed of Figure 1The server 110 shown performs. The large model-based intelligent response system can at least include a user input module 1100, a keyword extraction module 1110, a historical record joint analysis module 1120, a permission judgment module 1130, a comprehensive portrait establishment module 1140, a knowledge recommendation module 1150, etc.: The user input module 1100 is used to obtain the input information of the user; the keyword extraction module 1110 is used to extract the key information corresponding to the input information through the large model; the historical record joint analysis module 1120 is used to obtain the historical record information of the user and determine several dimensional features of the user according to the key information and the historical record information; the several dimensional features include position information, basic attributes, interaction behaviors, skill tags, knowledge shortboards, and decision-making styles; the permission judgment module 1130 is used to determine the permission level of the user according to the position information of the user and determine the knowledge base to which the user has access rights according to the permission level; the comprehensive portrait establishment module 1140 is used to generate a dynamic user portrait according to the several dimensional features of the user; the knowledge recommendation module 1150 is used to output a personalized reply for the user according to the dynamic user portrait and the knowledge base to which the user has access rights.

[0027] The system of the embodiments of this specification can analyze user behaviors and historical data according to multi-dimensional information such as the user's historical records, interaction behaviors, and scenario tags, construct an accurate user portrait, accurately match the user's needs, provide personalized knowledge recommendations, and significantly improve the accuracy of knowledge services. The dynamic user portrait of the embodiments of this specification can be a feature expression that is dynamically updated through a machine learning model and reflects the user's immediate state and potential intention based on several dimensional features of the user continuously and real-time collected. The system of the embodiments of this specification can associate the dynamic user portrait with the knowledge base to which the user has access rights to provide accurate knowledge services.

[0028] In some embodiments, please refer to Figure 3 , Figure 3It is a schematic structural diagram of the historical record joint analysis module 1120 provided by an embodiment of the present invention. The historical record joint analysis module 1120 further includes an attribute acquisition module 1122, a behavior analysis module 1124, a professional field portrait module 1126, a decision-making style classification module 1128, etc. The attribute acquisition module 1122 is used to obtain the basic attributes of the user according to the user's historical record information, and the basic attributes include years of employment, educational background, and skill certifications; the behavior analysis module 1124 is used to obtain the interaction behaviors of the user according to the user's historical record information, and the interaction behaviors include high-frequency operation modes, knowledge base usage time periods, and knowledge base usage frequencies; the professional field portrait module 1126 is used to determine the professional field portrait of the user according to the user's historical record information and the user's interaction behaviors, and the professional field portrait includes skill tags and knowledge shortboards; the decision-making style classification module 1128 is used to obtain the historical decision-making behaviors of the user according to the user's historical record information, and determine the decision-making style of the user according to the user's historical decision-making behaviors.

[0029] The system of the embodiments of this specification can provide decision-making support according to the comprehensive portrait of the user. By analyzing the decision-making style and context association of the user, it can predict the possible decision-making needs of the user, prepare relevant support information in advance, thereby shortening the decision-making time and effectively improving the decision-making efficiency and quality. The system of the embodiments of this specification classifies the decision-making style of the user according to the user's historical decision-making behaviors, such as whether to rely on data reports or historical solutions, so as to provide personalized decision-making support.

[0030] In some embodiments, the intelligent response system based on the large model further includes a learning path prediction module, etc. The learning path prediction module is used to obtain the historical learning records of the user and the job requirement information corresponding to the user's job position, predict associated courses or associated documents according to the historical learning records, and generate ability improvement suggestions according to the job requirement information.

[0031] The embodiments of this specification can also identify the knowledge shortboards and learning preferences of the user based on the user's historical learning records, recommend associated courses or documents, and generate ability improvement suggestions in combination with the job requirements, helping employees to improve skills targeted, form a personalized learning path, and contribute to the career development of employees.

[0032] In some embodiments, the intelligent response system based on the large model further includes a scenario label module, etc. The scenario label module is used to obtain multi-round context conversations, identify the current task stage of the user and the device type in real time according to the multi-round context conversations, and generate scenario-based recommendations for the user according to the current task stage of the user and the device type.

[0033] In some embodiments, please refer to Figure 4 , Figure 4It is a schematic flowchart of generating scenario-based recommendations provided by an embodiment of the present invention. The scenario tag module generates scenario-based recommendations in the following ways:

[0034] 400. When the user's current task stage is the project startup stage, recommend process guidance documents to the user;

[0035] 410. When the user's current task stage is the project delivery stage, recommend cross-departmental collaboration data to the user;

[0036] 420. Adjust the depth of the response content according to the device type.

[0037] Embodiments of this specification can identify the user's current task stage (such as the project startup stage or the project delivery stage, etc.) and device type (such as the mobile end or the PC end, etc.) in real time, and provide scenario-based knowledge support, such as differentiated services for in-depth reading requirements on the PC end and quick query on the mobile end.

[0038] For example, when the user's current task stage is the project startup period, the system can recommend process guidance, basic knowledge, and documents to help the user quickly understand the project background and process. When the user's current task stage is the project delivery period, the system may recommend decision support and cross-departmental collaboration data to help the user make informed decisions. When the user's current task stage is the new product release period, the system may automatically push competitor analysis documents to help the user understand the market dynamics and competitor situations. When the user's current task stage is the employee promotion period, the system may recommend corresponding leadership or management courses according to the user's promotion situation to help the user improve management capabilities. When conducting daily Q&A, the system can perform refined analysis and reply according to the questions raised by the user.

[0039] The system of the embodiments of this specification can dynamically optimize the information density, format, and interaction logic of the response content by identifying characteristics such as the device type, screen size, and input method of the user's device to adapt to the user experience requirements of different terminals. Embodiments of this specification can perceive the specific situation where the user is located through the scenario tag module, thereby adjusting the presentation method and content depth of the knowledge service, and optimizing the user experience.

[0040] In some embodiments, the intelligent answering system based on the large model further includes a context association module, a hierarchical strategy module, etc. The context association module is used to obtain the user's recent collaborative personnel and corporate strategic dynamics, and provide context-related knowledge recommendations to the user according to the user's recent collaborative personnel and corporate strategic dynamics; the hierarchical strategy module is used to determine the user's position information and permission level, and implement a hierarchical profiling strategy according to the user's position information and permission level to provide different knowledge services to the user.

[0041] In the embodiments of this specification, the system provides context - related knowledge recommendations by identifying the user's recent collaborative personnel and corporate strategic dynamics. For example, it can recommend relevant team knowledge or contacts, helping users find the required information and resources more quickly, improving team collaboration efficiency, and promoting team collaboration and knowledge sharing. The system in the embodiments of this specification can automatically push relevant knowledge and documents by identifying corporate strategic dynamics, helping employees understand corporate strategies and implement them in daily work. For example, when the enterprise is about to launch a new product, the system can automatically push a competitor analysis document to the user. The system in the embodiments of this specification can keep in sync with corporate strategic dynamics, timely push relevant knowledge and documents, enabling employees to better understand corporate strategies and translate them into practical actions.

[0042] In the embodiments of this specification, the system further includes a hierarchical policy module. The hierarchical policy module can implement a hierarchical profiling policy based on the user's job information and permission level. For example, the system can provide process guidance and basic knowledge recommendations to users with the job information of junior employees, and can provide cross - departmental collaboration data and decision - making support to users with the job information of managers.

[0043] In some embodiments, the permission judgment module further includes performing the following operations through a dynamic update mechanism: periodically reviewing the user's access permissions; when detecting abnormal behavior of the user, triggering permission adjustment or updating of knowledge recommendation content.

[0044] The embodiments of this specification can not only screen and recommend knowledge by analyzing information such as permission levels and content preferences, ensuring that users access knowledge base content that matches their responsibilities and interests, avoiding information overload, enabling users to quickly locate the required information, and improving the utilization efficiency of the knowledge base. Moreover, the embodiments of this specification also set up a dynamic update mechanism. The system can ensure the accuracy and timeliness of user information by periodically refreshing the user profile. At the same time, the system detects abnormal behavior of the user, such as suddenly frequently accessing documents outside their job field, which can trigger permission review or adjustment of knowledge recommendation content, etc. The embodiments of this specification enable the system to respond to changes in user needs in real - time through the dynamic update mechanism and provide personalized knowledge services.

[0045] The embodiments of this specification ensure that users can only access the knowledge base content within their permissions through the permission judgment module and the dynamic update mechanism, preventing information leakage. The system performs access control on knowledge according to the user's job information and permission level, and timely adjusts the permission settings by dynamically monitoring abnormal behavior to ensure the security of the knowledge base.

[0046] In some embodiments, the knowledge recommendation module supports multi-round context conversations, including: a conversation history module for storing user interaction data and constructing a conversation context graph with temporal associations based on the user interaction data; the user interaction data includes natural language input, operation behaviors, and system responses; a composite intent association engine for identifying explicit and implicit intents in the user's input information through a multi-level intent parsing model and mapping them into a structured intent tree; a knowledge node dynamic completion module for extracting unexposed nodes strongly associated with the structured intent tree from the conversation context graph with temporal associations when detecting a user's follow-up behavior, and generating a differentiated response based on the user device type.

[0047] In this embodiment, the conversation context graph with temporal associations can be a dynamically semantic network organized in chronological order, representing the context evolution relationship of multi-round conversations through nodes (entities, intents, or actions in the conversation) and edges (semantic or logical associations). Explicit intents can be intents directly conveyed by the user through clear language expressions, and explicit intents can be quickly identified through keywords or fixed sentence patterns. Implicit intents can be potential intents that the user does not directly state and need to be inferred through context, common sense, or behavior patterns. The structured intent tree can be a predefined, hierarchical intent logic framework that organizes possible user intent paths in a tree structure for standardizing the system response process.

[0048] In this embodiment, the conversation history module can store user interaction data and further construct a conversation context graph with temporal associations. The conversation context graph can include the semantic dependency relationship weights between nodes, and the weight values are dynamically updated according to the interaction frequency. Then, the composite intent association engine can identify explicit and implicit intents in the user input through a multi-level intent parsing model (such as a BERT-GRU joint network, etc.), thereby mapping discrete intents into a structured intent tree. Additionally, when detecting a user's follow-up behavior (such as the intent similarity > similarity threshold in consecutive conversation turns), the following can be executed: extracting unexposed nodes strongly associated with the structured intent tree (cosine similarity ≥ preset threshold) from the conversation context graph with temporal associations; generating a differentiated response based on the user's device type (PC side, mobile side, etc.).

[0049] For example, the conversation history module can store user interaction data in real time and construct a time-series conversation graph with a decay factor (i.e., a conversation context graph with time-series associations) through a graph neural network. At the same time, the system can dynamically adjust the weights between nodes according to interaction freshness; the composite intent association engine can adopt a three-level model of BERT-GRU-CRF to map the recognized intent to a predefined structured intent tree and generate an intent confidence score at the same time; when the knowledge node dynamic completion module detects a user's follow-up behavior (such as the intent inheritance score > 0.8 and the semantic focus matching degree > 75%), it preferentially extracts unshown nodes strongly associated with the structured intent tree from the time-series conversation graph and generates a differentiated response based on the user's device type. For example, it adaptively generates a response form according to the device API capabilities (such as the GPU memory of the mobile device). For the PC side, the system can push associated knowledge cards and a visualization roadmap in the form of a sidebar; for the mobile side, the system generates a voice summary and provides a quick jump deep link. Among them, the unshown nodes strongly associated with the structured intent tree meet the conditions: there is a reachable path with the current intent tree node, the association weight ranks in the top 20%, and they have not been shown in the last three rounds of conversations.

[0050] The embodiments of this specification can make the system easy to expand and adapt to new business requirements through modular design, such as adding a knowledge base, adjusting permission rules, etc. The system of the embodiments of this specification adopts modular design, and each module is independent and works together, enabling the system to flexibly expand and adapt to new business requirements and reducing the later maintenance cost.

[0051] In this embodiment, after the user inputs a question through the terminal, the system first extracts keywords, and then combines the user's historical records to obtain data such as the user's job information and basic attributes. Then, the system can judge the user's permission level and determine the access scope of the knowledge base. The system then analyzes the user's interaction behavior, constructs a professional field portrait, and predicts the learning path. At the same time, the system can real-time identify the user's scenario tags and context associations, classify the user's decision-making style, and implement a hierarchical portrait strategy. Then, the system can associate the user portrait and analysis results with the internal knowledge base, provide accurate knowledge services according to the user portrait, and output task promotion suggestions, professional knowledge answers, or recommended contacts, etc. Throughout the process, the dynamic update mechanism continuously monitors and adjusts the user portrait to ensure the accuracy and personalization of knowledge recommendations.

[0052] The embodiments of this specification can analyze user behaviors and historical data based on multi-dimensional information such as the user's historical records, interaction behaviors, and scenario tags, construct accurate user portraits, precisely match user needs, provide personalized knowledge recommendations, and significantly improve the accuracy of knowledge services; the embodiments of this specification can also provide decision-making support based on the comprehensive user portrait. By analyzing the user's decision-making style and context correlation, it is possible to predict the user's possible decision-making needs, prepare relevant support information in advance, thereby shortening the decision-making time and effectively improving the decision-making efficiency and quality; the embodiments of this specification can also identify the user's knowledge gaps and learning preferences based on the user's historical learning records, recommend relevant courses or documents, and generate ability improvement suggestions in combination with job requirements to help employees improve their skills targeted, form personalized learning paths, and assist in the career development of employees; the embodiments of this specification can also screen and recommend knowledge by analyzing information such as permission levels and content preferences to ensure that users can access knowledge base content that matches their responsibilities and interests. While avoiding information overload, users can quickly locate the information they need, improving the utilization efficiency of the knowledge base.

[0053] The system of the embodiments of this specification can be able to identify the user's current task stage and device type in real time and provide scenario-based knowledge support, such as differential services for in-depth reading requirements on the PC side and quick query on the mobile side. The system of the embodiments of this specification can ensure that users can only access the knowledge base content within their permission scope through permission judgment and dynamic update mechanisms, prevent information leakage, and strengthen permission management and security. The system of the embodiments of this specification can recommend relevant team knowledge or contacts by analyzing the user's recent collaborative personnel, promote team collaboration and knowledge sharing, and improve team collaboration efficiency. The system of the embodiments of this specification can automatically push relevant knowledge and documents by identifying the enterprise's strategic dynamics to help employees understand the enterprise's strategy and implement it in their daily work, enabling employees to better understand the enterprise's strategy and translate it into actual actions. The system of the embodiments of this specification can continuously optimize knowledge recommendations and permission settings by periodically refreshing the user portrait and monitoring abnormal behaviors, maintaining the dynamic accuracy of knowledge management, thereby continuously optimizing knowledge recommendations and permission management, achieving the continuous optimization of knowledge management while maintaining the timeliness and accuracy of the system.

[0054] The system of the embodiments of this specification is easy to expand and adapt to new business requirements through modular design, such as adding a knowledge base, adjusting permission rules, etc. The system adopts modular design, and each module is independent and works in cooperation with each other, enabling the system to flexibly expand and adapt to new business requirements, reducing the later maintenance cost, and improving the adaptability and expandability of the system.

[0055] The above describes specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve the desired result. Additionally, the processes depicted in the drawings do not necessarily require the specific order or sequential order shown to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0056] Please refer to Figure 5 , Figure 5 which is a schematic flowchart of an intelligent answering method based on a large model provided by an embodiment of this specification.

[0057] As Figure 5 shown, the intelligent answering method based on a large model may at least include:

[0058] 500. Obtain the input information of the user;

[0059] 510. Extract the key information corresponding to the input information through the large model;

[0060] 520. Obtain the historical record information of the user, and determine several dimensional features of the user according to the key information and the historical record information; the several dimensional features include position information, basic attributes, interaction behaviors, skill tags, knowledge short - boards, and decision - making styles;

[0061] 530. Determine the permission level of the user according to the position information of the user, and determine the knowledge base to which the user has access rights according to the permission level;

[0062] 540. Generate a dynamic user portrait according to the several dimensional features of the user;

[0063] 550. Output a personalized reply for the user according to the dynamic user portrait and the knowledge base to which the user has access rights.

[0064] In some embodiments, determining several dimensional features of the user according to the key information and the historical record information further includes: obtaining the basic attributes of the user according to the historical record information of the user, where the basic attributes include years of employment, educational background, and skill certifications; obtaining the interaction behaviors of the user according to the historical record information of the user, where the interaction behaviors include high - frequency operation modes, knowledge base usage time periods, and knowledge base usage frequencies; determining the professional field portrait of the user according to the historical record information of the user and the interaction behaviors of the user, where the professional field portrait includes skill tags and knowledge short - boards; obtaining the historical decision - making behaviors of the user according to the historical record information of the user, and determining the decision - making style of the user according to the historical decision - making behaviors of the user.

[0065] In some embodiments, the large model-based intelligent response method further includes: obtaining the user's historical learning records and the job requirement information corresponding to the user's job position, predicting associated courses or documents based on the historical learning records, and generating ability improvement suggestions based on the job requirement information.

[0066] In some embodiments, the large model-based intelligent response method further includes: obtaining multi-round context conversations, real-time identifying the user's current task stage and device type according to the multi-round context conversations, and generating scenario-based recommendations for the user according to the user's current task stage and device type.

[0067] In some embodiments, generating scenario-based recommendations for the user according to the user's current task stage and device type includes: when the user's current task stage is the project startup stage, recommending process guidance documents to the user; when the user's current task stage is the project delivery stage, recommending cross-departmental collaboration data to the user; and adjusting the depth of the response content according to the device type.

[0068] In some embodiments, the large model-based intelligent response method further includes: obtaining the user's recent collaborative personnel and corporate strategic dynamics, providing context-related knowledge recommendations to the user according to the user's recent collaborative personnel and corporate strategic dynamics; determining the user's job position information and permission level, implementing a hierarchical profiling strategy according to the user's job position information and permission level, and providing different knowledge services to the user.

[0069] In some embodiments, determining the user's permission level according to the user's job position information and determining the knowledge base to which the user has access rights according to the permission level further includes performing the following operations through a dynamic update mechanism: periodically reviewing the user's access rights; when detecting abnormal behavior of the user, triggering permission adjustment or knowledge recommendation content update.

[0070] In some embodiments, outputting a personalized response for the user according to the dynamic user profile and the knowledge base to which the user has access rights further includes supporting multi-round context conversations, including: storing user interaction data, constructing a temporally correlated dialogue context graph according to the user interaction data; the user interaction data includes natural language input, operation behavior, and system response; identifying the explicit and implicit intentions in the user's input information through a multi-level intention parsing model and mapping them into a structured intention tree; when detecting a user's follow-up behavior, extracting the unexposed nodes strongly correlated with the structured intention tree from the temporally correlated dialogue context graph, and generating a differentiated response based on the user's device type.

[0071] The embodiments of this specification can analyze user behavior and historical data based on multi-dimensional information such as the user's historical records, interaction behaviors, and scenario tags, construct a precise user portrait, accurately match user needs, provide personalized knowledge recommendations, and significantly improve the accuracy of knowledge services. The embodiments of this specification can also provide decision-making support based on the comprehensive user portrait. By analyzing the user's decision-making style and context correlation, it is possible to predict the user's possible decision-making needs, prepare relevant support information in advance, thereby shortening the decision-making time and effectively improving the decision-making efficiency and quality. The embodiments of this specification can also identify the user's knowledge gaps and learning preferences based on the user's historical learning records, recommend relevant courses or documents, and generate ability improvement suggestions in combination with job requirements to help employees improve their skills targetedly, form a personalized learning path, and assist in the career development of employees. The embodiments of this specification can also screen and recommend knowledge by analyzing information such as permission levels and content preferences to ensure that users can access knowledge base content that matches their responsibilities and interests. While avoiding information overload, users can quickly locate the information they need, improving the utilization efficiency of the knowledge base.

[0072] The system of the embodiments of this specification can be able to identify the user's current task stage and device type in real time, and provide scenario-based knowledge support, such as differentiated services for in-depth reading requirements on the PC side and quick query on the mobile side. The system of the embodiments of this specification can ensure that users can only access the knowledge base content within their permission scope through permission judgment and dynamic update mechanisms, prevent information leakage, and strengthen permission management and security. The system of the embodiments of this specification can recommend relevant team knowledge or contacts by analyzing the user's recent collaborative personnel, promote team collaboration and knowledge sharing, and improve team collaboration efficiency. The system of the embodiments of this specification can automatically push relevant knowledge and documents by identifying the enterprise strategic dynamics, help employees understand the enterprise strategy and implement it in their daily work, enabling employees to better understand the enterprise strategy and translate it into actual actions. The system of the embodiments of this specification can continuously optimize knowledge recommendations and permission settings by periodically refreshing the user portrait and monitoring abnormal behaviors, maintaining the dynamic precision of knowledge management, thereby continuously optimizing knowledge recommendations and permission management, and achieving the continuous optimization of knowledge management while maintaining the timeliness and accuracy of the system.

[0073] The system of the embodiments of this specification is easy to expand and adapt to new business requirements through modular design, such as adding a knowledge base, adjusting permission rules, etc. The system adopts modular design, and each module is independent and collaborative with each other, enabling the system to flexibly expand and adapt to new business requirements, reducing the later maintenance cost, and improving the system adaptability and expandability.

[0074] Each embodiment in this specification is described in a progressive manner. For the parts that are the same or similar among the embodiments, reference can be made to each other, and the key point of each embodiment is to illustrate the differences from other embodiments. In particular, for the embodiment of the intelligent response system based on the large model, since it is basically similar to the embodiment of the intelligent response system based on the large model, the description is relatively simple, and reference can be made to the relevant parts of the method embodiment for the relevant content.

[0075] Please refer to Figure 6 FIG. shows a schematic structural diagram of an electronic device of a server in an intelligent response system based on a large model provided by an embodiment of this specification.

[0076] As Figure 6 shown, the electronic device 600 may include: at least one processor 610, at least one network interface 640, a user interface 630, a memory 650, and at least one communication bus 620.

[0077] Among them, the communication bus 620 can be used to realize the connection and communication of the above-mentioned components.

[0078] Among them, the user interface 630 may include buttons, and the optional user interface may further include a standard wired interface and a wireless interface.

[0079] Among them, the network interface 640 may include, but is not limited to, a Bluetooth module, an NFC module, a ZigBee module, and a UWB module, etc.

[0080] Among them, the processor 610 may include one or more processing cores. The processor 610 connects various parts within the entire electronic device 600 through various interfaces and lines, and by running or executing instructions, programs, code sets, or instruction sets stored in the memory 650, and calling data stored in the memory 650, it executes various functions of the electronic device 600 and processes data. Optionally, the processor 610 may be implemented in at least one hardware form of DSP, FPGA, or PLA. The processor 610 may integrate one or a combination of several of CPU and GPU, etc. Among them, the CPU mainly processes the operating system, user interface, and application programs, etc.; the GPU is responsible for the rendering and drawing of the content to be displayed on the display screen.

[0081] Among them, the memory 650 may include RAM or ROM. Optionally, the memory 650 includes a non-transitory computer-readable medium. The memory 650 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 650 may include a program storage area and a data storage area. Among them, the program storage area can store instructions for implementing the operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-mentioned method embodiments, etc.; the data storage area can store the data involved in the above-mentioned method embodiments. Optionally, the memory 650 may also be at least one storage device located far from the aforementioned processor 610. The memory 650 as a computer storage medium may include an operating system, a communication module, a user interface module, and an intelligent response application program based on a large model. The processor 610 can be used to call the intelligent response application program based on the large model stored in the memory 650 and execute the steps in an intelligent response method based on a large model mentioned in the foregoing embodiments.

[0082] The embodiments of this specification also provide a computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When it runs on a computer or a processor, it causes the computer or the processor to execute the above Figure 5 One or more steps in the illustrated embodiments. If each component module of the above electronic device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.

[0083] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions according to the embodiments of this specification are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired manner (such as coaxial cable, optical fiber, Digital Subscriber Line (DSL)) or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server or data center that includes one or more integrated available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a Digital Versatile Disc (DVD)), or a semiconductor medium (such as a Solid State Disk (SSD)), etc.

[0084] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. The foregoing storage media include: various media such as ROM, RAM, magnetic disks, or optical discs that can store program codes. Without conflict, the technical features in this embodiment and the implementation solutions can be combined arbitrarily.

[0085] The above-described embodiments are merely described in terms of the preferred embodiments of this specification, and do not limit the scope of this specification. Without departing from the design spirit of this specification, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of this specification shall fall within the protection scope determined by the claims of this specification.

Claims

1. An intelligent response system based on a large model, characterized in that, including: a user input module for obtaining user input information; a keyword extraction module for extracting key information corresponding to the input information through a large model; a historical record joint analysis module for obtaining the user's historical record information and determining several dimensional features of the user according to the key information and the historical record information; the several dimensional features include position information, basic attributes, interaction behaviors, skill tags, knowledge gaps, and decision-making styles; a permission judgment module for determining the user's permission level according to the user's position information and determining the knowledge base to which the user has access permission according to the permission level; a comprehensive profile building module for generating a dynamic user profile according to the user's several dimensional features; a knowledge recommendation module for outputting a personalized reply for the user according to the dynamic user profile and the knowledge base to which the user has access permission; the historical record joint analysis module includes: an attribute acquisition module for obtaining the user's basic attributes according to the user's historical record information, and the basic attributes include years of employment, educational background, and skill certifications; a behavior analysis module for obtaining the user's interaction behaviors according to the user's historical record information, and the interaction behaviors include high-frequency operation modes, knowledge base usage time periods, and knowledge base usage frequencies; a professional field profile module for determining the user's professional field profile according to the user's historical record information and the user's interaction behaviors, and the professional field profile includes skill tags and knowledge gaps; a decision-making style classification module for obtaining the user's historical decision-making behaviors according to the user's historical record information and determining the user's decision-making style according to the user's historical decision-making behaviors.

2. The intelligent response system based on the large model according to claim 1, wherein The system further includes: a learning path prediction module for obtaining the user's historical learning records and the job requirement information corresponding to the user's position information, predicting associated courses or associated documents according to the historical learning records, and generating ability improvement suggestions according to the job requirement information.

3. The intelligent answering system based on a large model according to claim 1, characterized in that, The system further includes: a scenario label module for obtaining multi-round context conversations, real-time identifying the user's current task stage and device type according to the multi-round context conversations, and generating scenario-based recommendations for the user according to the user's current task stage and device type.

4. The intelligent response system based on a large model according to claim 3, wherein The scenario label module generates scenario-based recommendations in the following manner: when the user's current task stage is the project startup stage, recommending process guidance documents to the user; when the user's current task stage is the project delivery stage, recommending cross-departmental collaboration data to the user; adjusting the depth of the reply content according to the device type.

5. The intelligent response system based on a large model according to claim 1, wherein The system further includes: a context association module for obtaining the user's recent collaborative personnel and corporate strategic dynamics, and providing context-related knowledge recommendations to the user according to the user's recent collaborative personnel and corporate strategic dynamics; a hierarchical strategy module for determining the user's position information and permission level, implementing a hierarchical profiling strategy according to the user's position information and permission level, and providing different knowledge services to the user.

6. The intelligent response system based on a large model according to claim 1, wherein The permission judgment module further includes performing the following operations through a dynamic update mechanism: Periodically review the user's access permissions; When detecting abnormal behavior of the user, trigger permission adjustment or update of knowledge recommendation content.

7. The intelligent response system based on a large model according to claim 1, characterized in that, The knowledge recommendation module supports multi-round context conversations, including: A conversation history record module for storing user interaction data and constructing a temporally associated conversation context graph according to the user interaction data; the user interaction data includes natural language input, operation behavior, and system response; A composite intention association engine for identifying explicit and implicit intentions in the user's input information through a multi-level intention parsing model and mapping them into a structured intention tree; A knowledge node dynamic completion module for, when detecting a user's follow-up behavior, extracting unshown nodes strongly associated with the structured intention tree from the temporally associated conversation context graph and generating a differentiated response based on the user device type.

8. An intelligent answering method for an intelligent answering system based on a large model according to any one of claims 1-7, characterized in that, Including: Obtain the user's input information; Extract key information corresponding to the input information through a large model; Obtain the user's historical record information, and determine several dimensional features of the user according to the key information and the historical record information; the several dimensional features include position information, basic attributes, interaction behavior, skill tags, knowledge gaps, and decision-making styles; determine the user's permission level according to the user's position information, and determine the knowledge base to which the user has access rights according to the permission level; Generate a dynamic user profile according to the several dimensional features of the user; Output a personalized response for the user according to the dynamic user profile and the knowledge base to which the user has access rights; The historical record joint analysis module includes: An attribute acquisition module for obtaining the user's basic attributes according to the user's historical record information, where the basic attributes include years of service, educational background, and skill certifications; A behavior analysis module for obtaining the user's interaction behavior according to the user's historical record information, where the interaction behavior includes high-frequency operation patterns, knowledge base usage time periods, and knowledge base usage frequencies; A professional field profile module for determining the user's professional field profile according to the user's historical record information and the user's interaction behavior, where the professional field profile includes skill tags and knowledge gaps; A decision-making style classification module for obtaining the user's historical decision-making behavior according to the user's historical record information and determining the user's decision-making style according to the user's historical decision-making behavior.

9. A computer-readable storage medium storing a computer program, characterized in that, When the program is executed by a processor, all steps of the method described in claim 8 are implemented.

Images