Intelligent consultation service system based on knowledge graph

Abstract

The invention relates to the technical field of intelligent consultation service systems, and discloses an intelligent consultation service system based on a knowledge graph, and the system comprises a user terminal, knowledge processing equipment and a cloud reasoning center. The user terminal inputs a consultation question text and context semantic data; the knowledge processing equipment receives an instruction to update a knowledge node state; a knowledge extraction unit of the cloud reasoning center determines an associated area, generates a semantic feature curve and maps the semantic feature curve, a relation matching unit optimizes a reasoning path to generate a decision scheme, and a decision generation unit executes reasoning or intervention operation according to the equipment state. The system can accurately understand semantics, optimize reasoning paths and process exceptions, and provides efficient and accurate intelligent consultation services for users by dynamically updating the knowledge graph.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent consulting service systems, and specifically to an intelligent consulting service system based on a knowledge graph. Background Art

[0002] With the rapid development of information technology, intelligent consulting service systems have been widely used in numerous fields, including finance, healthcare, and education. Traditional intelligent consulting service systems are mostly based on rule-based systems or simple keyword matching technologies. This approach often struggles to accurately understand user intent when dealing with complex semantics and ambiguous questions, resulting in poor accuracy and practicality in consulting results. For example, in medical consultation scenarios, when users describe their symptoms in vague and unprofessional language, traditional systems may not accurately match the corresponding disease knowledge, resulting in incorrect or incomplete advice.

[0003] While some intelligent consulting service systems that utilize machine learning have improved their semantic understanding capabilities to a certain extent, they lack in-depth analysis of knowledge structure and semantic associations, making them difficult to handle complex situations such as semantic jumps and unusual semantics. When users suddenly change topics during a consultation or input logically contradictory information, the system is prone to reasoning errors and is unable to provide effective solutions. Furthermore, the knowledge update mechanisms of existing intelligent consulting service systems are inflexible and unable to adapt to dynamic changes in knowledge in real time, resulting in outdated knowledge graphs that fail to reflect the latest industry knowledge and technological achievements.

[0004] Furthermore, traditional systems lack the ability to optimize reasoning paths and handle exceptions. During the reasoning process, the inability to rationally plan reasoning paths based on the semantic distribution of the knowledge graph can lead to low reasoning efficiency. When the system experiences an anomaly, the lack of an effective intervention mechanism prevents rapid restoration of normal operations, severely impacting user experience and service quality. As user demands for the accuracy, efficiency, and real-time performance of intelligent consulting services continue to increase, the development of an intelligent consulting service system that can deeply understand semantics, flexibly handle exceptions, and dynamically update knowledge is urgent. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent consulting service system based on knowledge graph to solve the problems raised in the above background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solutions: an intelligent consulting service system based on a knowledge graph, the system comprising a user terminal, a knowledge processing device, and a cloud-based reasoning center, the user terminal and the knowledge processing device being respectively communicatively connected to the cloud-based reasoning center, the cloud-based reasoning center comprising a knowledge extraction unit, a relationship matching unit, and a decision generation unit;

[0007] The user terminal is used to input the consultation question text and contextual semantic data in real time;

[0008] The knowledge processing device is used to receive processing instructions from the cloud reasoning center to update the knowledge node status;

[0009] The knowledge extraction unit is used to record the semantic jump period and the logical recovery period when the knowledge processing device is running, determine the associated area according to the semantic jump period and the logical recovery period, control the knowledge processing device to maintain a preset constraint threshold for abnormal semantics and record the change parameters, generate a semantic feature curve according to the change parameters and perform semantic distribution mapping on the knowledge graph;

[0010] The relationship matching unit is used to optimize the reasoning path of the knowledge graph that has completed the semantic distribution mapping and generate a decision solution;

[0011] The decision generation unit includes a conventional reasoning unit and an emergency reasoning unit. The conventional reasoning unit is used to control the knowledge processing device to execute a baseline reasoning mode when the knowledge processing device is in a normal state; the emergency reasoning unit is used to control the knowledge processing device to perform intervention operations in an associated area according to a decision plan when the knowledge processing device detects an abnormal state.

[0012] Preferably, the user terminal includes a terminal housing and a multimodal input device, a semantic analysis device, a logic evaluation module and a communication module arranged in the housing;

[0013] The multimodal input device is used to synchronously acquire text input data and voice input data;

[0014] The semantic parsing device is used to extract context-related features of input data;

[0015] The logic evaluation module is used to calculate the similarity between the current context-related features and the historical standard pattern based on the attention mechanism, determine the semantic coherence index, and trigger the correction signal according to the calculation result;

[0016] The communication module is used to upload input data to the cloud reasoning center.

[0017] Preferably, the knowledge processing device includes a processing chassis and a logic operation module, a parameter detection module, a command response module and a protocol conversion module arranged in the chassis.

[0018] Preferably, the semantic jump period and logic recovery period of recording the knowledge processing device during operation include:

[0019] When the knowledge processing device detects a semantic jump, it records the timestamp data of the jump occurrence time, continuously monitors the logic recovery critical point through the parameter detection module, and records the serial number of the recovery time.

[0020] Preferably, determining the associated area according to the semantic jump period and the logical recovery period includes:

[0021] Generate jump coordinates and recovery coordinates based on timestamp data and sequence numbers;

[0022] The jump coordinates, the restored coordinates and the user terminal position are spatially associated to form a closed area, and the closed area is marked as an associated area.

[0023] Preferably, the control knowledge processing device maintains a preset constraint threshold for abnormal semantics and records change parameters, generates a semantic feature curve according to the change parameters, and performs semantic distribution mapping on the knowledge graph, including the following steps:

[0024] Controlling the knowledge processing device to adjust the output intensity of the associated region and obtaining semantic parameters in real time through the parameter detection module;

[0025] A preset constraint threshold range is set. If the semantic parameter exceeds the preset constraint threshold range, the knowledge processing device is controlled to maintain the preset constraint threshold operation;

[0026] Continuously record the semantic output values ​​of knowledge processing devices to form a set of changing parameters;

[0027] According to the set of changing parameters, the hidden Markov model algorithm is used to generate the semantic dynamic feature curve;

[0028] Extracting a number of key feature points at equal time intervals from the semantic dynamic feature curve, wherein the number of the key feature points is proportional to the duration of the curve;

[0029] The knowledge graph is annotated with semantic distribution mapping based on the extracted key feature points.

[0030] Preferably, the inference path optimization of the knowledge graph that has completed the semantic distribution mapping includes logically isolating the preset inference paths that overlap with the high-correlation segments in the knowledge graph after the knowledge graph is annotated with the semantic distribution mapping.

[0031] Preferably, when the knowledge processing device detects an abnormal state, controlling the knowledge processing device to perform an intervention operation in the associated area according to the decision plan includes:

[0032] S1. Select the associated area access location with the shortest operation process based on the next inference node to be switched;

[0033] S2. Control the knowledge processing device to switch to the access location and perform parameter synchronization;

[0034] S3, controlling the knowledge processing device to enter the associated area from the access position and perform state adjustment according to the reasoning path in the decision solution;

[0035] S4, obtaining the output indicators of the knowledge processing equipment in real time, and dynamically correcting the output indicators using the stochastic gradient descent algorithm;

[0036] S5. After each continuous reasoning operation is completed in the associated area, the knowledge processing device is controlled to pause output and exit the associated area, and S1 is re-executed.

[0037] Preferably, said S1 comprises the following steps:

[0038] The A* algorithm is used to calculate the feasible access locations for each associated area and conduct a comprehensive evaluation based on the operation process and logical distance parameters;

[0039] Based on the comprehensive evaluation results, select the associated area access location with the shortest operation process and the smallest logical distance;

[0040] The comprehensive evaluation based on the operation process and logical distance parameters includes:

[0041] Establish a two-dimensional evaluation matrix containing the number of operation process steps and logical distance values;

[0042] After normalizing the two-dimensional evaluation matrix, factor analysis is used to extract the first principal factor as a comprehensive evaluation indicator;

[0043] The decision generating unit is used to select the associated area access position with the largest comprehensive evaluation index value.

[0044] Preferably, the knowledge extraction unit further includes a dynamic update module, which is used to incrementally update the knowledge graph according to the semantic feature curve after the knowledge processing device completes the semantic distribution mapping.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] The knowledge graph-based intelligent consulting service system provided by this invention significantly improves the quality and efficiency of intelligent consulting services through the collaborative operation of multiple components. The multimodal input device of the user terminal supports simultaneous text and voice input, meeting the usage habits of different users and broadening information acquisition channels. The semantic analysis device extracts contextual features and, combined with the similarity calculation based on the attention mechanism of the logical evaluation module, can accurately determine the semantic coherence index, effectively avoiding semantic misunderstandings and improving the accuracy of understanding user questions.

[0047] The logical operation module and parameter detection module within the knowledge processing device work together to accurately update the status of knowledge nodes and monitor the device's operating status in real time. The knowledge extraction unit identifies associated regions by recording periods of semantic transitions and logical recovery periods. It then constrains abnormal semantics and generates semantic feature curves, mapping the semantic distribution of the knowledge graph. This enables the system to deeply understand complex semantic relationships and mine potential knowledge.

[0048] The relationship matching unit optimizes the reasoning paths of the knowledge graph that completes semantic distribution mapping. By logically isolating pre-set reasoning paths with overlapping highly correlated sections, it improves reasoning efficiency and accuracy. The decision-making unit's regular and emergency reasoning units function under normal and abnormal conditions, respectively. The emergency reasoning unit, in particular, performs intervention operations within the associated area based on the decision plan. Through a series of operations, including scientific access location selection, parameter synchronization, state adjustment, and output indicator correction, it ensures that the system can quickly resume normal operation under abnormal circumstances.

[0049] The dynamic update module of the knowledge extraction unit incrementally updates the knowledge graph based on the semantic feature curve, ensuring that the knowledge graph is always synchronized with the latest knowledge, enabling the system to adapt to the ever-changing knowledge environment and provide users with more accurate, efficient and timely intelligent consulting services, greatly improving the user experience and the practicality of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a working principle diagram of an intelligent consulting service system based on knowledge graph according to the present invention;

[0051] Figure 2 This is the working principle diagram of the user terminal;

[0052] Figure 3 Determine the flow chart for the associated area;

[0053] Figure 4 This is a flowchart of the intervention operation in the associated area under abnormal conditions. DETAILED DESCRIPTION

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0055] See also Figure 1-Figure 4The present invention relates to an intelligent consulting service system based on a knowledge graph. The intelligent consulting service system comprises user terminals, a knowledge processing device, and a cloud-based reasoning center. The user terminals and the knowledge processing device are each connected to the cloud-based reasoning center. The cloud-based reasoning center includes a knowledge extraction unit, a relationship matching unit, and a decision generation unit.

[0056] The user terminal receives the consultation question text and contextual semantic data input by the user in real time and uploads it to the cloud-based reasoning center. The knowledge processing device receives the processing instructions from the cloud-based reasoning center to update the knowledge node status. During the operation of the knowledge processing device, the knowledge extraction unit records the semantic jump period and the logic recovery period, determines the associated area, controls the knowledge processing device to operate according to the preset constraint threshold for abnormal semantics and records the change parameters, generates a semantic feature curve, and completes the semantic distribution mapping of the knowledge graph. The relationship matching unit optimizes the reasoning path of the knowledge graph that has completed the semantic distribution mapping and generates a decision plan. The conventional reasoning unit of the decision generation unit controls the knowledge processing device to execute the baseline reasoning mode when it is in the normal state; when the emergency reasoning unit detects an abnormal state, it controls the knowledge processing device to perform intervention operations in the associated area according to the decision plan.

[0057] Example 1:

[0058] The user terminal serves as the direct interface between the system and the user. Its internal structure comprises a terminal housing, a multimodal input device, a semantic parser, a logic evaluation module, and a communication module. These components work together to collect, process, and upload user inquiries.

[0059] The terminal housing, serving as the physical carrier for each functional module, is manufactured using an integrated engineering plastic molding process, with a frosted surface for enhanced grip comfort. The housing's interior features a multi-layered design. The outermost layer is an electromagnetic shielding layer, constructed from a copper foil composite material with a thickness of 0.1 mm, effectively preventing external electromagnetic interference from affecting the internal circuitry. The middle layer is a structural support layer constructed from high-strength engineering plastic, ensuring the housing will not deform under certain external impacts. The inner layer is a heat dissipation channel layer, designed with a dense microchannel structure to quickly transfer heat generated by internal electronic components to the housing surface. A touchscreen display is located on the front of the housing for text input and information display, and a microphone array is located below the display for voice data collection.

[0060] The multimodal input device integrates both text and voice input. Text input utilizes a combination of a touch keyboard and handwriting recognition. The touch keyboard area features an adaptive layout algorithm that dynamically adjusts key size and position based on user input habits. For example, keys for commonly used characters are appropriately enlarged and optimized for more convenient input. The handwriting recognition function supports a variety of writing styles. The system utilizes a vast library of handwriting fonts and uses a convolutional neural network algorithm to analyze and recognize the user's handwriting in real time. The system automatically adjusts the sampling frequency to ensure recognition accuracy as the user's writing speed changes. Voice input utilizes a four-microphone array design, with four microphones positioned at the four corners of the terminal housing. This layout effectively suppresses ambient noise. Using a beamforming algorithm, the system accurately locates the user's voice source and focuses its primary sound pickup direction on that location. The system also features echo cancellation, which identifies and filters out sound signals re-collected by the microphone after being output by the speaker, ensuring the purity of the captured voice data.

[0061] The semantic parser is responsible for conducting in-depth analysis of the data collected by the multimodal input device. First, the text data undergoes preprocessing, including stop word removal, lexical analysis, and syntactic analysis. The stop word list is dynamically updated based on different application scenarios to ensure the accuracy of preprocessing. Lexical analysis uses a bidirectional LSTM model to more accurately identify word boundaries and parts of speech. Syntactic analysis uses a dependency parsing algorithm to construct a sentence grammatical structure tree, revealing the dependencies between words. For speech data, speech recognition is first performed to convert the speech into text, and then the same processing flow as for text data is applied. When extracting contextual features, the system comprehensively considers word co-occurrence, semantic similarity, and temporal information. By constructing a semantic network, related words and concepts are connected to form a complex semantic association graph. The system also records users' historical consultation records, analyzes their interests and question types, and provides a reference for extracting contextual features for the current question.

[0062] The logical evaluation module evaluates contextual features based on an attention mechanism. This module maintains a library of historical standard patterns, which is continuously updated and refined as the system operates. When a new contextual feature is received, the system calculates its similarity with the historical standard patterns. The attention mechanism automatically focuses on the important parts of a feature, assigning different weights to different features. For example, features closely related to the core of the question are given a higher weight, while less significant features are given a lower weight. This method enables a more accurate assessment of the semantic coherence index. If the semantic coherence index falls below a preset threshold, the system triggers a correction signal. This correction signal is handled differently depending on the specific situation, potentially prompting the user to re-enter their input or providing additional information for selection.

[0063] The communication module is responsible for uploading the processed input data to the cloud-based inference center. This module supports multiple communication protocols, including 4G, 5G, and WiFi. When selecting a communication method, the system will automatically switch according to the current network environment. For example, when the WiFi signal is good, WiFi is used for data transmission to improve transmission speed and stability; when the WiFi signal is unstable or unavailable, it automatically switches to the 4G or 5G network. The communication module also has a data encryption function, using the AES-256 encryption algorithm to encrypt the transmitted data to ensure the security of the data during transmission. At the same time, in order to ensure the reliability of data transmission, the system uses breakpoint resume technology. When an interruption occurs during data transmission, the system can record the location of the transmitted data and continue transmission from the breakpoint after the network is restored to avoid repeated transmission and loss of data.

[0064] During actual operation, the various modules of the user terminal collaborate with each other to form an organic whole. The multimodal input device provides users with flexible and diverse input methods to meet the usage habits of different users. The semantic analysis device deeply explores the semantic information of the input data, providing a solid foundation for subsequent processing. The logical evaluation module ensures the semantic coherence and accuracy of the input data, improving the processing efficiency and quality of the system. The communication module ensures the secure and reliable transmission of data, enabling the cloud-based reasoning center to obtain user consultation information in a timely manner. This collaborative working method enables the user terminal to efficiently and accurately collect and process user consultation questions, providing a strong guarantee for the normal operation of the entire intelligent consulting service system.

[0065] Example 2:

[0066] The internal structure of the knowledge processing device includes a processing chassis and a logic operation module, parameter detection module, command response module, and protocol conversion module installed within the chassis. These modules work together to process and analyze the knowledge graph and interact with the cloud-based reasoning center.

[0067] The processing chassis is constructed from aluminum alloy through precision casting and CNC machining. The chassis features a modular design, with each functional module connected to the mainboard via standardized interfaces, facilitating installation, maintenance, and upgrades. The chassis' interior is coated with an electromagnetic shielding coating, effectively preventing electromagnetic interference generated by internal electronic components from affecting external devices and also protecting internal modules from external electromagnetic interference. Cooling holes on the chassis' bottom, combined with the internal cooling fan, create an excellent airflow path, ensuring stable device temperatures during extended operation.

[0068] The logical operation module, as the core of the knowledge processing device, utilizes a multi-core processor architecture, equipped with a high-performance graphics processing unit (GPU) and a dedicated AI accelerator chip. This module boasts powerful parallel computing capabilities, enabling it to handle multiple complex logical tasks simultaneously. When processing the knowledge graph, the logical operation module first parses the nodes and edges in the graph, extracting node attribute information and edge relationship information. It then infers and calculates this information based on pre-set logical rules. For example, when receiving a user inquiry, the logical operation module searches the knowledge graph for relevant nodes and paths, identifying the most likely answer through logical reasoning and semantic matching. The logical operation module also supports deep learning algorithms, enabling automatic learning and optimization of the knowledge graph. By continuously analyzing and processing large amounts of knowledge data, the module automatically adjusts inference rules and algorithm parameters to improve the accuracy and efficiency of knowledge processing.

[0069] The parameter detection module monitors various operating parameters of the knowledge processing device in real time. This module integrates multiple sensors, including temperature, voltage, current, and memory usage. Temperature sensors are located at key locations within the chassis and monitor the temperature of heat-generating components such as the processor and chipset in real time. When the temperature exceeds a preset threshold, the system automatically activates the cooling fan to increase heat dissipation efficiency. Voltage and current sensors monitor the device's power supply in real time, ensuring stable voltage and current operation. A memory usage sensor monitors system memory usage in real time. When memory usage is excessive, the system automatically optimizes memory allocation, freeing up unnecessary memory space and improving system efficiency. In addition to hardware parameters, the parameter detection module also monitors semantic parameters. For example, during knowledge graph processing, the module monitors parameters such as the strength of associations between nodes and the credibility of inference paths. Changes in these parameters reflect the operating status and effectiveness of the knowledge processing device, providing a basis for system optimization and adjustment.

[0070] The command response module is responsible for receiving and processing commands from the cloud-based inference center. This module utilizes an efficient message queuing mechanism to ensure timely and accurate command reception and processing. Upon receiving a command, the command response module first parses it, extracting its type, parameters, and execution requirements. Then, based on the command type, it forwards it to the appropriate functional module for processing. For example, if it's a knowledge graph update command, the command response module forwards it to the logic operation module, which is responsible for updating the knowledge graph's content and structure. The command response module also manages command priority. Urgent commands are prioritized to ensure timely execution of critical tasks. The module also records the execution status and results of commands and provides feedback to the cloud-based inference center. If an exception occurs during command execution, the command response module automatically takes appropriate action, such as retrying or downgrading, and promptly reports the exception to the cloud-based inference center.

[0071] The protocol conversion module is responsible for converting between different communication protocols. This module supports a variety of common communication protocols, including HTTP, HTTPS, MQTT, and AMQP. When the knowledge processing device communicates with the cloud-based inference center, the protocol conversion module automatically selects the appropriate conversion method based on the protocols supported by both parties. For example, if the cloud-based inference center sends instructions using HTTP, while the knowledge processing device's internal modules communicate using MQTT, the protocol conversion module converts the HTTP instructions into MQTT messages, ensuring that the instructions are correctly received and processed. The protocol conversion module also provides data format conversion. Different communication protocols may use different data formats, such as JSON, XML, and Protobuf. The protocol conversion module recognizes and converts these different data formats, ensuring data consistency and integrity during transmission. Furthermore, the protocol conversion module supports protocol extensions and custom protocols. As technology evolves and application scenarios change, new communication protocols and data formats may emerge. The protocol conversion module provides a flexible extension interface, allowing developers to add support for new protocols as needed.

[0072] During actual operation, the various modules of the knowledge processing device collaborate to form an organic whole. The core logic operation module is responsible for processing and analyzing the knowledge graph. The parameter detection module monitors the device's operating status in real time, ensuring the system's stable operation. The command response module ensures that the device can promptly respond to commands from the cloud-based reasoning center, enabling system collaboration. The protocol conversion module resolves compatibility issues between different communication protocols, ensuring smooth communication between the device and the cloud-based reasoning center as well as other devices. This collaborative approach enables the knowledge processing device to efficiently and accurately process and analyze the knowledge graph, providing strong support for the intelligent consulting service system.

[0073] Example 3:

[0074] When a knowledge processing device performs a knowledge graph inference task, it monitors changes in the semantic features of the input data in real time. Once it detects a change in a semantic feature exceeding a preset threshold, the system determines that a semantic jump has occurred. At this point, the system immediately records the timestamp of the jump, accurate to the millisecond level, to ensure the accuracy of the time information.

[0075] After detecting a semantic jump, the parameter detection module initiates continuous monitoring of the logical recovery critical point. This critical point is the critical point at which the knowledge processing device recovers from a semantic jump state to a normal logical reasoning state. During monitoring, the parameter detection module analyzes the output of the knowledge processing device and evaluates its logical coherence and semantic accuracy. When the logical coherence and semantic accuracy of the output meet the preset recovery criteria, the system determines that the logical recovery critical point has been reached and records the sequence number of the recovery moment. This sequence number uniquely identifies the recovery event in chronological order and is used for subsequent data processing and analysis.

[0076] Based on the recorded timestamp data and sequence number, the system generates transition coordinates and recovery coordinates. Transition coordinates are multidimensional coordinate points composed of the timestamp at the time of the transition and the corresponding semantic feature vector. The semantic feature vector contains key semantic information about the input data at the time of the transition, such as keywords, themes, and sentiment. Recovery coordinates are multidimensional coordinate points composed of the sequence number at the time of the recovery and the corresponding semantic feature vector. These two coordinate points represent the starting and ending points of the semantic transition, respectively, providing basic data for subsequent spatial correlation analysis.

[0077] After generating the jump coordinates and recovery coordinates, the system spatially associates these two coordinate points with the user's terminal location. The user's terminal location refers to the geographic location of the user when they enter their inquiry, typically obtained through GPS positioning or network IP address resolution. This spatial association process utilizes a graph-theoretic association algorithm, treating the jump coordinates, recovery coordinates, and user terminal location as nodes in a graph. The connections between nodes represent their semantic relevance. By calculating the semantic similarity and spatial distance between nodes, the system constructs a semantic association network.

[0078] Based on the semantic association network, the system searches for the minimum closed region that encompasses the jump coordinates, restored coordinates, and the user terminal location. This closed region consists of a set of nodes with high semantic relevance, whose connections form a closed loop. The closed region is determined using a combination of the minimum spanning tree algorithm and the convex hull algorithm. First, the minimum spanning tree algorithm is used to find the shortest path connecting the jump coordinates, restored coordinates, and the user terminal location. Then, the convex hull algorithm is used to calculate the minimum convex polygon that encompasses this path. This convex polygon is the final association region.

[0079] The identification of associated regions is crucial for the subsequent targeted handling of abnormal semantics. Within these regions, knowledge processing equipment can more accurately locate the source and impact of abnormal semantics, enabling more effective action. For example, the system can focus on analyzing knowledge nodes within these regions to identify abnormal knowledge or inference rules that may lead to semantic jumps. Furthermore, the identification of associated regions provides a basis for local optimization of the knowledge graph. The system can adjust and improve the knowledge structure within these regions, enhancing the accuracy and robustness of the knowledge graph.

[0080] In practical applications, the process of detecting semantic transitions and determining associated regions is dynamic and ongoing. As knowledge processing devices continue to operate and users continuously input questions, the system continuously detects semantic transitions, records relevant data, identifies associated regions, and optimizes and updates the knowledge graph based on this information. This dynamic processing mechanism enables the system to adapt to the ever-changing semantic environment, improving the quality and effectiveness of intelligent consulting services.

[0081] Furthermore, to improve the accuracy and efficiency of semantic jump detection, the system employs a multi-dimensional detection strategy. In addition to monitoring the magnitude of semantic feature changes, the system also analyzes factors such as the speed, direction, and frequency of semantic change. For example, if semantic features undergo drastic changes within a short period of time, or if semantic changes exhibit periodic fluctuations, the system will pay closer attention to these situations and conduct in-depth analysis and processing in a timely manner.

[0082] The system uses a hierarchical data structure to represent and store association regions. A region is represented as a subgraph containing multiple knowledge nodes and relationships, with clear boundaries and internal structure. Nodes and relationships in the subgraph are assigned semantic labels and weights, reflecting their importance and relevance within the region. This hierarchical data structure enables more efficient management and manipulation of association regions.

[0083] Semantic transition detection and associated region identification are essential components of intelligent consulting service systems. By accurately detecting semantic transitions, recording relevant data, generating coordinate points, performing spatial associations, and identifying associated regions, the system can more effectively handle unusual semantics, improving the quality of the knowledge graph and the effectiveness of intelligent consulting services. This processing mechanism not only meets users' consulting needs but also continuously learns and optimizes, adapting to the ever-changing semantic environment, providing users with more accurate and efficient intelligent consulting services.

[0084] Example 4:

[0085] After identifying the associated region, the system will fine-tune and map its semantic parameters. For example, in an intelligent consulting scenario involving financial investment, when a user asks about "high-risk investment strategies," the system might detect a semantic shift and identify an associated region encompassing knowledge nodes such as "stock investment," "futures trading," and "risk assessment."

[0086] First, the system controls the knowledge processing device to adjust the output intensity of the associated region. This process is similar to adjusting the brightness of a light, making the knowledge nodes within the associated region more "eye-catching" in the knowledge graph. By increasing the activation threshold of these nodes, they are more easily called up during reasoning, thereby improving the efficiency of retrieving knowledge related to the user's question. For example, the "Futures Trading Risk Control" node, which originally had a low weight in regular searches, has been given a higher weight within the associated region, allowing it to be retrieved more quickly by the system when dealing with questions related to high-risk investments.

[0087] At the same time, the parameter detection module begins acquiring semantic parameters in real time. These parameters include, but are not limited to, the strength of associations between knowledge nodes, the credibility of reasoning paths, and semantic similarity. In the financial investment example, the system monitors the strength of the association between the "Stock Investment" and "Risk Assessment" nodes, as well as the credibility parameters of different investment strategies in the current market environment. Real-time acquisition of these parameters provides the data foundation for subsequent semantic analysis.

[0088] Pre-set constraint thresholds serve as the system's safety line. For example, the system sets a reasonable range for semantic parameters in investment risk assessment. If it detects that the risk assessment parameters of an investment strategy exceed this range, the system immediately controls the knowledge processing equipment to maintain operation within the preset constraint threshold. This is like a speed limiter in a car, ensuring that the system does not deviate from the reasonable range due to excessive pursuit of a specific result. In a financial investment scenario, if the risk assessment parameters of a high-risk investment strategy exceed the safety threshold, the system will automatically adjust the strength of its recommendation for that strategy to avoid providing overly aggressive investment advice to the user.

[0089] The system continuously records the semantic output values ​​of knowledge processing devices, forming a set of variable parameters. These output values ​​reflect the dynamic changes in the knowledge processing devices as they process user questions. In financial investment consulting, these output values ​​may include recommendation indices for different investment strategies, changes in risk assessment results over time, and so on. By recording these output values, the system can capture subtle changes in the knowledge processing process, providing rich data for subsequent semantic feature analysis.

[0090] Based on a set of changing parameters, the system uses a hidden Markov model algorithm to generate a semantic dynamic feature curve. This process is similar to how meteorologists plot weather curves based on meteorological data. Hidden Markov models can uncover underlying patterns and regularities in seemingly random semantic output values. In financial investment consulting, a semantic dynamic feature curve might reflect the fluctuations in the strength of recommendations for different investment strategies as the market changes, or the changing trends in user interest in investment products of varying risk levels.

[0091] In the semantic dynamic feature curve, the system extracts key feature points at equal intervals, with the number of key feature points proportional to the curve's duration. This is analogous to marking important notes at regular intervals in a piece of music. These key feature points represent significant turning points in semantic change or moments of significant characterization. In financial investment consulting, these key feature points might correspond to the timing of significant market events or moments when a user's consulting attitude shifts significantly.

[0092] Finally, based on the extracted key feature points, the system performs semantic distribution mapping on the knowledge graph. This process is similar to marking important locations and routes on a map. By mapping key feature points to the knowledge graph, the system can more accurately reflect the distribution of semantic information within the knowledge graph. In financial investment consulting, semantic distribution mapping annotation may highlight the connections between knowledge nodes related to high-risk investment strategies or mark the most valuable investment recommendations in a specific market environment.

[0093] In this way, the knowledge graph can more accurately reflect semantic information, providing more reliable support for intelligent consulting services. In practical applications, this semantic distribution mapping annotation method can help the system better understand user questions and provide more precise answers and suggestions. For example, in financial investment consulting, the system can recommend products that best suit users' risk tolerance and investment goals based on the results of semantic distribution mapping annotation. It can also adjust recommendation strategies in a timely manner when the market changes, providing users with more timely investment advice.

[0094] Example 5:

[0095] When the knowledge processing device detects an abnormal state, the system immediately activates its emergency response mechanism. This process begins by calculating feasible access locations for each associated area using the A algorithm. Similar to finding the shortest path in a maze, the A algorithm assigns an evaluation value to each possible access location, taking into account the number of steps in the operational process and the logical distance value. The number of steps in the operational process represents the number of operations required to reach the target state from the current state, while the logical distance value measures the degree of semantic connection between two knowledge nodes.

[0096] The system creates a two-dimensional evaluation matrix that includes the number of process steps and logical distance values. This matrix is ​​similar to a coordinate graph, with the horizontal axis representing the number of process steps and the vertical axis representing the logical distance values. Each feasible access location has a corresponding coordinate point in this matrix. To make the indicators of these two different dimensions comparable, the system normalizes the matrix. Normalization is like converting measurements in different units to the same standard unit, allowing the number of process steps and logical distance values ​​to be compared on the same scale.

[0097] After normalization, the system uses factor analysis to extract the first principal factor as a comprehensive evaluation indicator. Factor analysis is similar to identifying hidden key factors from a large amount of observational data. The first principal factor represents the primary factor that explains the majority of data variation. By extracting this factor, the system can simplify the two-dimensional evaluation matrix into a one-dimensional comprehensive evaluation indicator, allowing for more intuitive comparisons of the advantages and disadvantages of different access locations.

[0098] Based on the comprehensive evaluation results, the system selects the access location in the associated area with the shortest operational process and the smallest logical distance. This is like choosing the shortest and easiest path to a destination among multiple roads. Once the access location is determined, the system controls the knowledge processing device to switch to that location and synchronize its parameters. This parameter synchronization process is similar to data transmission and calibration between two devices, ensuring that the knowledge processing device can accurately obtain and process relevant information at the new access location.

[0099] The knowledge processing device then enters the associated region and performs state adjustments based on the reasoning path in the decision plan. This process is similar to driving a route on a map: the knowledge processing device operates and adjusts the knowledge nodes within the associated region along the predetermined reasoning path. During this process, the system obtains real-time output indicators from the knowledge processing device, which reflect the performance and effectiveness of the knowledge processing device during the state adjustment process.

[0100] To ensure the accuracy and stability of output metrics, the system uses a stochastic gradient descent algorithm to dynamically correct them. This algorithm is similar to climbing a mountain, constantly adjusting direction to find the highest peak. Based on real-time output metrics, the system continuously adjusts the operating parameters of the knowledge processing device, gradually bringing the output metrics closer to the ideal value.

[0101] After each continuous reasoning operation in the associated area is completed, the system controls the knowledge processing device to pause output and exit the associated area. This is like taking a break and reviewing the work after completing a task. After exiting the associated area, the system selects a new access location and repeats the above process. This cycle continues until the abnormal state of the knowledge processing device is resolved.

[0102] Furthermore, after the knowledge processing device completes semantic distribution mapping, the knowledge extraction unit's dynamic update module incrementally updates the knowledge graph based on the semantic feature curve. The semantic feature curve records semantic changes during knowledge processing, and the dynamic update module analyzes these changes to identify areas of the knowledge graph that require updating. Incremental updates are like adding a few new strokes of color to a painting. Without disrupting the original knowledge structure, the system adds new knowledge nodes or adjusts the relationships between existing nodes, ensuring the timeliness and accuracy of the knowledge graph.

[0103] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0104] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent consulting service system based on knowledge graph, characterized in that: include: A user terminal, a knowledge processing device, and a cloud-based reasoning center, wherein the user terminal and the knowledge processing device are respectively connected to the cloud-based reasoning center for communication, and the cloud-based reasoning center includes a knowledge extraction unit, a relationship matching unit, and a decision generation unit; The user terminal is used to input the consultation question text and contextual semantic data in real time; The knowledge processing device is used to receive processing instructions from the cloud reasoning center to update the knowledge node status; The knowledge extraction unit is used to record the semantic jump period and the logical recovery period when the knowledge processing device is running, determine the associated area according to the semantic jump period and the logical recovery period, control the knowledge processing device to maintain a preset constraint threshold for abnormal semantics and record the change parameters, generate a semantic feature curve according to the change parameters and perform semantic distribution mapping on the knowledge graph; The relationship matching unit is used to optimize the reasoning path of the knowledge graph that has completed the semantic distribution mapping and generate a decision solution; The decision generation unit includes a conventional reasoning unit and an emergency reasoning unit. The conventional reasoning unit is used to control the knowledge processing device to execute a baseline reasoning mode when the knowledge processing device is in a normal state; the emergency reasoning unit is used to control the knowledge processing device to perform intervention operations in an associated area according to a decision plan when the knowledge processing device detects an abnormal state.

2. The intelligent consulting service system based on knowledge graph according to claim 1, characterized in that: The user terminal includes a terminal housing and a multimodal input device, a semantic analysis device, a logic evaluation module and a communication module arranged in the housing; The multimodal input device is used to synchronously acquire text input data and voice input data; The semantic parsing device is used to extract context-related features of input data; The logic evaluation module is used to calculate the similarity between the current context-related features and the historical standard pattern based on the attention mechanism, determine the semantic coherence index, and trigger the correction signal according to the calculation result; The communication module is used to upload input data to the cloud reasoning center.

3. The intelligent consulting service system based on knowledge graph according to claim 1 is characterized in that: The knowledge processing device includes a processing chassis and a logic operation module, a parameter detection module, an instruction response module and a protocol conversion module arranged in the chassis.

4. The intelligent consulting service system based on knowledge graph according to claim 3 is characterized in that: The semantic jump period and logic recovery period of recording the operation of the knowledge processing device include: When the knowledge processing device detects a semantic jump, it records the timestamp data of the jump occurrence time, continuously monitors the logic recovery critical point through the parameter detection module, and records the serial number of the recovery time.

5. The intelligent consulting service system based on knowledge graph according to claim 4 is characterized in that: Determining the associated area according to the semantic jump period and the logical recovery period includes: Generate jump coordinates and recovery coordinates based on timestamp data and sequence numbers; The jump coordinates, the restored coordinates and the user terminal position are spatially associated to form a closed area, and the closed area is marked as an associated area.

6. The intelligent consulting service system based on knowledge graph according to claim 5 is characterized in that: The control knowledge processing device maintains a preset constraint threshold for abnormal semantics and records change parameters, generates a semantic feature curve according to the change parameters, and performs semantic distribution mapping on the knowledge graph, including the following steps: Controlling the knowledge processing device to adjust the output intensity of the associated region and obtaining semantic parameters in real time through the parameter detection module; A preset constraint threshold range is set. If the semantic parameter exceeds the preset constraint threshold range, the knowledge processing device is controlled to maintain the preset constraint threshold operation; Continuously record the semantic output values ​​of knowledge processing devices to form a set of changing parameters; According to the set of changing parameters, the hidden Markov model algorithm is used to generate the semantic dynamic feature curve; Extracting a number of key feature points at equal time intervals from the semantic dynamic feature curve, wherein the number of the key feature points is proportional to the duration of the curve; The knowledge graph is annotated with semantic distribution mapping based on the extracted key feature points.

7. The intelligent consulting service system based on knowledge graph according to claim 6, characterized in that: The inference path optimization of the knowledge graph that has completed the semantic distribution mapping includes logically isolating the preset inference paths that overlap with the high-correlation segments in the knowledge graph after the semantic distribution mapping is annotated on the knowledge graph.

8. The intelligent consulting service system based on knowledge graph according to claim 1, characterized in that: When the knowledge processing device detects an abnormal state, controlling the knowledge processing device to perform an intervention operation in the associated area according to the decision plan includes: S1. Select the associated area access location with the shortest operation process based on the next inference node to be switched; S2. Control the knowledge processing device to switch to the access location and perform parameter synchronization; S3, controlling the knowledge processing device to enter the associated area from the access position and perform state adjustment according to the reasoning path in the decision solution; S4, real-time acquisition of output indicators of the knowledge processing device, and dynamic correction of the output indicators using the stochastic gradient descent algorithm; S5. After each continuous reasoning operation is completed in the associated area, the knowledge processing device is controlled to suspend output and exit the associated area, and S1 is re-executed.

9. The intelligent consulting service system based on knowledge graph according to claim 8, characterized in that: Said S1 comprises the following steps: The feasible access locations for each associated area are calculated using the A* algorithm and a comprehensive evaluation is performed based on the operation process and logical distance parameters; Based on the comprehensive evaluation results, select the associated area access location with the shortest operation process and the smallest logical distance; The comprehensive evaluation based on the operation process and logical distance parameters includes: Establish a two-dimensional evaluation matrix containing the number of operation process steps and logical distance values; After normalizing the two-dimensional evaluation matrix, factor analysis is used to extract the first principal factor as a comprehensive evaluation indicator; The decision generating unit is used to select the associated area access position with the largest comprehensive evaluation index value.

10. The intelligent consulting service system based on knowledge graph according to claim 1, characterized in that: The knowledge extraction unit also includes a dynamic update module, which is used to incrementally update the knowledge graph according to the semantic feature curve after the knowledge processing device completes the semantic distribution mapping.

Images