A human-computer interaction method based on intelligent data analysis

By extracting and aligning multimodal features from user input data, performing cross-modal data matching and intent recognition, constructing an interaction model and network, and optimizing the interaction engine, the problem of low accuracy in multimodal data interaction in traditional methods is solved, thereby improving user experience and system performance.

CN118349147BActive Publication Date: 2024-11-08JOINT WARFARE COLLEGE NAT DEFENSE UNIV OF THE CHINESE PEOPLES LIBERATION ARMY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410513644.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-26
Publication Date
2024-11-08
Estimated Expiration
2044-04-26

AI Technical Summary

Technical Problem

Traditional human-computer interaction methods based on intelligent data analysis have low accuracy in multimodal data interaction, resulting in an uncomfortable user experience.

Method used

By acquiring user input data, we perform multimodal feature extraction and alignment, cross-modal input data matching and intent recognition, construct cross-modal data interaction models and graphs, design cross-modal data interaction networks and architectures, and optimize the interaction engine to improve interaction accuracy.

Benefits of technology

It improves the accuracy of multimodal data interaction and user experience comfort, enhances the robustness, stability and responsiveness of the system, and supports distributed and cross-platform application scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118349147B_ABST
    Figure CN118349147B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of interactive technology, and in particular to a human-computer interaction method based on intelligent data analysis. The method comprises the following steps: extracting multimodal features from user operation input data and aligning features to obtain multimodal feature alignment data; identifying human-computer interaction intentions based on the multimodal feature alignment data to obtain human-computer interaction intention data; constructing a cross-modal data interaction graph based on the human-computer interaction intention data to obtain a cross-modal data interaction graph; constructing a cross-modal data interaction network for the cross-modal data interaction graph and performing interactive architecture design to obtain a cross-modal data interaction architecture; performing architectural performance optimization processing on the cross-modal data interaction architecture and performing interactive engine design to obtain a cross-modal human-computer interaction engine, and sending the cross-modal human-computer interaction engine to a cloud platform to perform human-computer interaction of intelligent data analysis. The present invention makes human-computer interaction more accurate by optimizing human-computer interaction technology.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of interactive technology, and in particular to a human-computer interaction method based on intelligent data analysis. Background Art

[0002] Intelligent data analysis uses machine learning, artificial intelligence and other technologies to automatically discover patterns, trends and anomalies in data, thereby providing decision makers with more accurate and comprehensive information support. Human-computer interaction, as an important part of intelligent data analysis, aims to establish a closer connection between users and data through an intuitive and friendly interface, thereby achieving a more efficient and intelligent data analysis and decision-making process. This method can not only help users quickly understand the information behind the data, but also provide personalized analysis suggestions, providing users with reliable reference and support when facing complex problems. Therefore, the human-computer interaction method based on intelligent data analysis has become an important engine for promoting the development of various fields such as enterprises and scientific research institutions, and has provided strong support for data-driven decision-making and innovative development. However, the traditional human-computer interaction method based on intelligent data analysis has the problem of low accuracy in multimodal data interaction, which makes it difficult for multimodal data interaction to maintain the comfort of user experience. Summary of the invention

[0003] Based on this, it is necessary to provide a human-computer interaction method based on intelligent data analysis to solve at least one of the above technical problems.

[0004] To achieve the above object, a human-computer interaction method based on intelligent data analysis is provided, the method comprising the following steps:

[0005] Step S1: obtaining user operation input data; performing multimodal feature extraction on the user operation input data to obtain user input multimodal feature data; performing feature alignment on the user input multimodal feature data to obtain multimodal feature alignment data;

[0006] Step S2: performing cross-modal input data matching on the user operation input data according to the multimodal feature alignment data to obtain cross-modal input matching data; performing human-computer interaction intention recognition on the cross-modal input matching data to obtain human-computer interaction intention data; constructing a cross-modal data interaction model according to the human-computer interaction intention data to obtain a cross-modal data interaction model; constructing a cross-modal data interaction graph according to the cross-modal data interaction model to obtain a cross-modal data interaction graph;

[0007] Step S3: constructing a cross-modal data interaction network for the cross-modal data interaction graph to obtain a cross-modal data interaction network; designing an interaction architecture based on the cross-modal data interaction network to obtain a cross-modal data interaction architecture; performing architecture performance optimization processing on the cross-modal data interaction architecture to obtain a cross-modal data interaction performance optimization architecture;

[0008] Step S4: Design an interaction engine based on the cross-modal data interaction performance optimization architecture and the cross-modal data interaction model to obtain a cross-modal human-computer interaction engine, and send the cross-modal human-computer interaction engine to the cloud platform to perform human-computer interaction for intelligent data analysis.

[0009] The present invention can capture multiple input methods of users, such as text, image, audio, etc., by extracting multimodal features of user operation input data. The advantage of doing so is that the content of user input can be understood more comprehensively, the system's recognition accuracy of user intentions can be improved, and the feature data of different modes can be aligned, which helps to eliminate data differences between different modes, making subsequent data processing more accurate and reliable, and improving the robustness and stability of the system; by matching multimodal feature alignment data, input data of different modes can be effectively associated, so as to better understand the user's operation intentions, which helps to improve the system's understanding and response capabilities to user needs; by performing intent recognition on cross-modal input matching data, the system can more accurately understand the user's intentions, and thus better provide services and support for users. This improves the interactive experience between the system and the user, and enhances the user's satisfaction with the system. Based on the identified human-computer interaction intention, building a cross-modal data interaction model helps the system better organize and manage data of different modalities, achieve more efficient data interaction and processing, and improve the overall performance and efficiency of the system. By building a cross-modal data interaction map, the data relationship of different modalities can be visualized, helping the system to better understand the association and connection between data, and providing more possibilities and flexibility for subsequent data analysis and application; by building a cross-modal data interaction network, the system can model the data interaction relationship of different modalities, so as to better understand and process the interaction relationship between multimodal data. This helps to improve the system's ability to understand complex data relationships and enhance the system's ability to integrate and utilize data. Based on the cross-modal data interaction network, a reasonable interaction architecture is designed, which can effectively organize and manage the various modules of the system and achieve efficient operation and management of data interaction. This helps to improve the maintainability and scalability of the system, and is also conducive to the optimization and upgrade of the subsequent system. Performance optimization of the interaction architecture can improve the system's operating efficiency and response speed, so as to better meet user needs. By optimizing the system's architecture design, the system's resource consumption can be reduced and the system's overall performance can be improved. Based on the optimized architecture and model, an efficient cross-modal human-computer interaction engine can be designed to achieve rapid processing and interactive response to multimodal data. This helps to improve the real-time and responsiveness of the system and provide users with a smoother interactive experience. By deploying the interactive engine on the cloud platform, human-computer interaction for intelligent data analysis can be achieved, allowing the system to process data and interact anytime and anywhere. This helps to improve the flexibility and availability of the system, while also better supporting distributed and cross-platform application scenarios.Therefore, the present invention provides a human-computer interaction method based on intelligent data analysis, which is an optimization processing of the traditional human-computer interaction method based on intelligent data analysis. It solves the problem that the traditional human-computer interaction method based on intelligent data analysis has low accuracy of multimodal data interaction, so that multimodal data interaction is difficult to maintain the comfort of user experience. It improves the accuracy of multimodal data interaction and ensures the consistency of multimodal data interaction and user experience comfort.

[0010] Preferably, step S1 comprises the following steps:

[0011] Step S11: Obtain user operation input data;

[0012] Step S12: classifying the input data type of the user operation input data to obtain input data type data;

[0013] Step S13: extracting multimodal features of the user operation input data according to the input data type data to obtain user input multimodal feature data;

[0014] Step S14: performing feature alignment on the multimodal feature data input by the user to obtain multimodal feature alignment data.

[0015] The present invention obtains the user's operation input data through various input channels (such as text input boxes, voice recognition, cameras, etc.). This ensures that the system can capture the user's diverse input, classify the user's input data, and determine the type of each input data, such as text, image, audio, etc. This helps the system to better understand the characteristics and processing methods of different types of data, prepare for subsequent feature extraction and processing, and select appropriate feature extraction methods for different types of input data, such as word embedding or TF-IDF processing for text data, convolutional neural network feature extraction for image data, and Mel spectrum feature extraction for audio data. This helps to convert the original input data into a multimodal feature representation with semantic information, and perform feature alignment on the user input multimodal feature data to obtain multimodal feature alignment data: align the feature data of different modes, eliminate the data differences between different modes, and ensure that they have consistent expressions in the feature space. This helps to improve the generalization ability and robustness of subsequent models, so that the system can understand the user's intentions more accurately.

[0016] Preferably, step S2 comprises the following steps:

[0017] Step S21: performing cross-modal input data matching on the user operation input data according to the multi-modal feature alignment data to obtain cross-modal input matching data;

[0018] Step S22: performing data logical association on the cross-modal input matching data to obtain cross-modal logical association data;

[0019] Step S23: performing human-computer interaction intention recognition on the cross-modal input matching data according to the cross-modal logical association data to obtain human-computer interaction intention data;

[0020] Step S24: constructing a cross-modal data interaction model according to the human-computer interaction intention data and the cross-modal logic association data to obtain a cross-modal data interaction model;

[0021] Step S25: construct a cross-modal data interaction graph according to the cross-modal data interaction model to obtain a cross-modal data interaction graph.

[0022] The present invention uses multimodal feature alignment data to match different modal data input by the user and find data associated with each other. This helps to convert the user's cross-modal input into a higher-level semantic representation, provide richer information to support subsequent processing, logically associate the cross-modal input matching data, analyze the relationship between them, and form cross-modal logical association data. This helps the system understand the logical relationship between different modalities, further improves the ability to understand user intentions, and based on cross-modal logical association data, performs human-computer interaction intention recognition on cross-modal input matching data, that is, infers the user's interaction intention between different modalities. This helps the system to understand the user's intention more accurately, thereby providing more personalized and accurate services, combining human-computer interaction intention data and cross-modal logical association data, and constructing a cross-modal data interaction model for describing and predicting cross-modal interactions between users and systems. This helps the system better understand the user's behavior pattern, provide a more intelligent and personalized interactive experience, and use the cross-modal data interaction model to construct a cross-modal data interaction map to describe the relationship between different modal data, user intentions, and system responses. This helps the system better manage and understand the interaction process of cross-modal data, and provide users with more intelligent services and support.

[0023] Preferably, step S23 includes the following steps:

[0024] Step S231: performing context-aware analysis on the cross-modal logical association data to obtain logical association context-aware data;

[0025] Step S232: performing a perceptual multidimensional analysis on the cross-modal logical association data according to the logical association context perception data to obtain logical association multidimensional perception data;

[0026] Step S233: performing cluster analysis on the logically associated multidimensional perception data to obtain logically associated multidimensional cluster data; performing multidimensional category structure analysis on the logically associated multidimensional perception data according to the logically associated multidimensional cluster data to obtain logically associated multidimensional category structure data;

[0027] Step S234: performing structural linear logic adjustment on the logically associated multidimensional perception data according to the logically associated multidimensional category structure data to obtain logical structural linear adjustment data;

[0028] Step S235: performing cross-modal input content linear correlation analysis on the cross-modal input matching data according to the logical structure linear adjustment data to obtain cross-modal content linear correlation data; performing content correlation time interval exponential distribution calculation on the cross-modal content linear correlation data to obtain content correlation exponential distribution data;

[0029] Step S236: Based on the content-related exponential distribution data, a content distribution Poisson process simulation is performed on the cross-modal content linear correlation data to obtain content distribution Poisson process data; based on the content distribution Poisson process data, human-computer interaction intention recognition is performed on the cross-modal content linear correlation data to obtain human-computer interaction intention data.

[0030] In this step of the present invention, the system takes into account contextual information, such as the environment, time, and location of the user's operation, when analyzing the cross-modal logical association data. This helps to better understand the context and user intention behind the data. By performing multidimensional analysis on the logical association data, the system can identify different dimensions in the data, thereby more comprehensively understanding the characteristics and meaning of the data. This helps to improve the system's ability to understand user behavior. Through cluster analysis, the system divides the logical association multidimensional perception data into different categories, and further clarifies the inherent structure and relationship between the data. This helps the system to better organize and manage data, and provides a basis for subsequent analysis and processing. By linearly adjusting the logical association multidimensional category structure data, the system can optimize the structure of the data to make it more in line with the needs of analysis and processing. This helps to improve the efficiency and accuracy of the system's use of data. By performing content linear association analysis and time interval exponential distribution calculation on cross-modal input matching data, the system can identify the degree and frequency of association between different data. This helps to understand the temporal relationship and correlation between data, and provides an important reference for subsequent analysis and modeling. Through content distribution Poisson process simulation, the system can simulate the interaction process between the user and the system, and then identify the user's interaction intention. This helps the system understand user needs and intentions more accurately and provide smarter and more personalized services.

[0031] Preferably, step S24 includes the following contents:

[0032] Step S241: capturing dependency relationships of cross-modal logic association data based on a self-attention mechanism to obtain cross-modal logic dependency relationship data;

[0033] Step S242: performing cross-modal causal relationship analysis on the cross-modal logic dependency relationship data according to the human-computer interaction intention data to obtain cross-modal interaction logic causal data;

[0034] Step S243: performing interactive dependency transfer learning on the human-computer interaction intention data and the cross-modal logical dependency relationship data according to the interactive logic dependency causal data to obtain interactive dependency transfer learning data;

[0035] Step S244: performing data volume incremental learning processing on the interactive dependency transfer learning data to obtain transfer learning incremental data;

[0036] Step S245: construct a cross-modal data interaction model based on the transfer learning incremental data and the cross-modal interaction logic causal data according to the recursive neural network to obtain a cross-modal data interaction model.

[0037] The self-attention mechanism of the present invention can help the system effectively capture the dependencies between various data when processing cross-modal logically associated data. The capture of such dependencies helps to understand the intrinsic connection and influence between data, and provides an important basis for subsequent analysis and modeling. Through the causal analysis of logical dependencies on human-computer interaction intention data, the system can identify the causal relationship between different data, that is, how the change of one data leads to the change of another data. This helps to understand the driving factors behind user behavior and provide more accurate predictions and responses for the system. Interactive dependency transfer learning can help the system transfer knowledge between different data sets, thereby improving the performance of the model on new tasks. By performing transfer learning on interactive logical dependency causal data, the system can better utilize existing knowledge to improve the understanding and analysis capabilities of human-computer interaction intention data and cross-modal logical dependency data. Through incremental learning of data volume, the system can continuously improve and optimize the model while continuously receiving new data. This helps to maintain the robustness and adaptability of the model to cope with changing environments and needs. Recursive neural networks can effectively process sequence data and complex data dependencies when building cross-modal data interaction models. By inputting transfer learning incremental data and cross-modal interaction logic causal data into the recursive neural network, the system can build a more intelligent and flexible model, thereby achieving more accurate understanding and response to human-computer interaction intentions.

[0038] Preferably, step S3 comprises the following steps:

[0039] Step S31: constructing a cross-modal data interaction network for the cross-modal data interaction graph to obtain a cross-modal data interaction network;

[0040] Step S32: Designing an interaction architecture according to the cross-modal data interaction network and the cross-modal data interaction graph to obtain a cross-modal data interaction architecture;

[0041] Step S33: Designing a parallel framework for the cross-modal data interaction architecture to obtain an interactive parallel framework;

[0042] Step S34: Optimize the architecture performance of the cross-modal data interaction architecture according to the interactive parallel framework to obtain a cross-modal data interaction performance optimized architecture.

[0043] In this step of the present invention, by constructing a cross-modal data interaction network, the system can effectively integrate and interact data of different modes. The establishment of such a network helps to achieve cross-modal association and information transmission between data, thereby improving the utilization efficiency and overall performance of data. By designing an interactive architecture based on a cross-modal data interaction network and a data interaction graph, the system can better organize and manage the relationship between data and improve the scalability and flexibility of the system. This architecture design can effectively support the interaction and integration between different modal data, laying the foundation for subsequent parallel processing and performance optimization. By designing a parallel framework, the system can make full use of multi-core processors or distributed computing resources to achieve efficient processing and calculation of the cross-modal data interaction architecture. This parallel framework design helps to improve the processing speed and throughput of the system, thereby responding to user requests and providing services more quickly. By performing performance optimization processing on the interactive parallel framework, the system can further improve the performance and efficiency of the cross-modal data interaction architecture. This includes optimization measures in terms of optimizing algorithms, adjusting parameters, reducing computing and communication overhead, etc., to ensure that the system can still operate efficiently in large-scale data and complex task scenarios.

[0044] Preferably, step S32 includes the following steps:

[0045] Step S321: performing a node topology structure analysis on the cross-modal data interaction graph to obtain the interaction graph node topology data; constructing a node connectivity graph on the interaction graph node topology data to obtain the interaction graph node connectivity graph;

[0046] Step S322: performing a depth search on the interaction graph node connectivity graph to obtain node connectivity graph depth data; performing a node depth-first search on the interaction graph node connectivity graph according to the node connectivity graph depth data to obtain the interaction graph node connectivity data;

[0047] Step S323: Perform structural sparse distribution law analysis on the interaction graph node connectivity data to obtain structural sparse distribution data; perform degree distribution structure analysis on the structural sparse distribution data to obtain structural sparse degree distribution data;

[0048] Step S324: establishing a data connection channel for the structure sparsity distribution data according to the cross-modal data interaction network to obtain a cross-modal data connection channel;

[0049] Step S325: performing bandwidth matching on the cross-modal data connection channel according to the structural sparsity distribution data to obtain connection bandwidth matching data;

[0050] Step S326: Designing a connection route for the cross-modal data connection channel according to the connection bandwidth matching data and the structure sparsity distribution data to obtain cross-modal structure routing data;

[0051] Step S327: Design an interaction architecture according to the cross-modal data connection channel, the connection bandwidth matching data, and the cross-modal structure routing data to obtain a cross-modal data interaction architecture.

[0052] The present invention obtains the node topology data of the interaction graph through the node topology structure analysis of the cross-modal data interaction graph, and constructs the node connectivity graph, so that the system can better understand the relationship and connection mode between different data. This helps to identify the important nodes and key paths between the data, thereby providing an important reference for subsequent data interaction and processing. By performing a deep search on the node connectivity graph of the interaction graph and performing a node depth-first search, the system can deeply explore the connection structure and hierarchical relationship between the data. This helps to discover the potential associations hidden between the data and provide support for the effectiveness and accuracy of data interaction. By performing a structural sparse distribution law analysis and a degree distribution structure analysis on the node connectivity data of the interaction graph, the system can understand the distribution law and characteristics between the data. This helps to identify the structural sparsity in the data interaction, thereby providing guidance for the establishment and optimization of subsequent data connection channels. According to the cross-modal data interaction network, a data connection channel is established so that data of different modes can be effectively interacted and transmitted. This helps to achieve rapid communication and information exchange between data, and improves data utilization and overall performance. By matching the bandwidth of the structured sparsity distribution data, the system can better configure the bandwidth resources of the connection channel to ensure efficient transmission and communication during data interaction. This helps to avoid data congestion and transmission delays, and improve the system's response speed and user experience. By connecting bandwidth matching data and structured sparsity distribution data for connection routing design, the system can optimize the data transmission path and achieve the best data interaction effect. This helps to reduce the path length and transmission delay of data transmission, and improve the efficiency and reliability of data interaction. By comprehensively considering cross-modal data connection channels, connection bandwidth matching data, and cross-modal structure routing data, the system can design an interaction architecture to achieve efficient interaction and integration of cross-modal data. This helps to build an optimized data interaction architecture, improve the performance and scalability of the system, and provide users with a better service experience.

[0053] Preferably, designing a connection route for a cross-modal data connection channel includes:

[0054] According to the connection bandwidth matching data and the structure sparsity distribution data, the structure demand capacity is evaluated to obtain the structure connection demand capacity data;

[0055] Based on the structural connection demand capacity data, the structural sparsity distribution data is sorted for connection structure demand to obtain connection structure demand sorting data; according to the connection structure demand sorting data, the structural sparsity distribution data is sorted and extracted to obtain structure distribution rank data;

[0056] Sorting and indexing the structure distribution ranking data based on the hash algorithm to obtain the structure data distribution index;

[0057] Combining routing addresses according to the structural distribution ranking data and the structural data distribution index to obtain structural distribution routing address data;

[0058] The connection routing is designed for the cross-modal data connection channel according to the structural distribution routing address data to obtain the cross-modal structural routing data.

[0059] The present invention can accurately determine the structural connection capacity required by the cross-modal data connection channel by evaluating the connection bandwidth matching data and the structural sparsity distribution data. This helps to ensure that there will be no bottlenecks or delays in the data transmission process, and ensure the efficiency and reliability of data interaction. According to the structural connection demand capacity data, the system can sort the structural sparsity distribution data to determine the priority and demand of the connection channel. This helps to optimize the data transmission path and improve the efficiency and priority of data interaction. By extracting the position of the connection structure demand sorting data, the system can understand the priority and importance of different data connection channels, so as to better plan the data transmission path and resource allocation. Based on the hash algorithm to generate the sorting index of the structural distribution position data, the system can quickly locate and access the data connection channel, and improve the speed and efficiency of data interaction. According to the structural distribution position data and the structural data distribution index, the system can effectively combine the routing address and determine the best data transmission path and routing strategy. This helps to reduce the delay and energy consumption of data transmission and improve the overall performance and stability of the system. Based on the structural distribution routing address data, the system can design the optimal cross-modal data connection route to ensure efficient interaction and communication between different data modes. This helps improve the scalability and flexibility of the system and meet the needs of different data processing scenarios.

[0060] Preferably, step S33 includes the following steps:

[0061] Step S331: performing high-concurrency traffic simulation on the cross-modal data interaction architecture to obtain high-concurrency traffic simulation data;

[0062] Step S332: performing concurrent transmission channel configuration according to high concurrent traffic simulation data to obtain concurrent transmission channels;

[0063] Step S333: Designing a buffer mechanism for high concurrent traffic simulation data to obtain a concurrent traffic buffer mechanism;

[0064] Step S334: constructing a buffered transmission channel for the concurrent transmission channel according to the concurrent traffic buffering mechanism to obtain a concurrent buffered transmission channel;

[0065] Step S335: Optimizing the concurrent buffer structure of the Apache Flink framework based on the concurrent buffer transmission channel to obtain a structurally optimized Apache Flink framework;

[0066] Step S336: Design a parallel framework based on the structurally optimized Apache Flink framework to obtain an interactive parallel framework.

[0067] The present invention simulates high-concurrency traffic in a cross-modal data interaction architecture. The system can simulate high-concurrency traffic conditions in various data interaction scenarios, including information such as data transmission rate and request response time. This helps the system evaluate its performance under high load conditions, discover potential bottlenecks and performance bottlenecks, and provide a reference for subsequent optimization work. According to the high-concurrency traffic simulation data, the system can configure concurrent transmission channels to meet the high-concurrency data transmission requirements. By reasonably configuring the transmission channels, the data transmission efficiency and stability can be improved, and the reliability and performance of the system under high load conditions can be ensured. A buffering mechanism is designed for the high-concurrency traffic simulation data. The system can achieve smooth flow of data transmission and avoid data congestion and loss. The design of the buffering mechanism helps to adjust the data transmission rate, balance the data interaction between producers and consumers, and improve the stability and throughput of the system. According to the concurrent traffic buffering mechanism, the system can construct a concurrent buffering transmission channel, combine the buffering mechanism with the transmission channel, and achieve efficient data transmission, which helps to reduce delays and fluctuations in the data transmission process and improve the reliability and stability of data transmission. Based on the concurrent buffering transmission channel, Apache Structural optimization of the Flink framework can improve the performance of the framework in high-concurrency scenarios. Optimizing the structure of the Apache Flink framework can accelerate the data processing and analysis process and improve the real-time and responsiveness of the system. Based on the structurally optimized Apache Flink framework, the system can design an interactive parallel framework to achieve parallel and concurrency in data interaction and processing, which helps to improve the system's processing power and efficiency, support more complex data processing tasks and scenarios, and meet users' real-time data interaction needs.

[0068] Preferably, the buffering mechanism design for high concurrent traffic simulation data includes:

[0069] Perform buffer pool capacity evaluation on high concurrent traffic simulation data to obtain buffer pool capacity evaluation data;

[0070] Memory space allocation is performed according to the buffer pool capacity assessment data to obtain memory space allocation data;

[0071] Dynamically adjust temporary memory space for high concurrent traffic simulation data according to buffer pool capacity assessment data and memory space allocation data to obtain temporary memory space adjustment data;

[0072] A buffer mechanism is designed for high concurrent traffic simulation data based on buffer pool capacity assessment data, memory space allocation data and temporary memory space adjustment data to obtain a concurrent traffic buffer mechanism.

[0073] The present invention evaluates the buffer pool capacity for high-concurrency flow simulation data, so that the system can understand the amount of data that needs to be buffered under high load conditions and the fluctuation of data flow, which helps to determine the appropriate buffer pool capacity and avoid data loss or system performance degradation caused by insufficient buffer pool capacity. According to the buffer pool capacity evaluation data, the system can allocate memory space to ensure that the system has enough memory for buffering high-concurrency flow simulation data. Correct memory space allocation can avoid system crash or performance degradation caused by memory overflow or insufficient memory, and ensure the stability and reliability of the system. According to the buffer pool capacity evaluation data and memory space allocation data, the system can dynamically adjust the temporary memory space for high-concurrency flow simulation data, which means that the system can dynamically adjust the memory allocation according to actual conditions to adapt to flow fluctuations or sudden increases in data volume, thereby maintaining the stability and efficiency of the system. In combination with the buffer pool capacity evaluation data, the memory space allocation data and the temporary memory space adjustment data, the system can design a buffer mechanism for high-concurrency flow, and this buffer mechanism can ensure stable transmission of data under high concurrency conditions, avoid data congestion or loss, and ensure the high availability and reliability of the system.

[0074] The beneficial effect of the present invention is that by extracting the multimodal features of the user's operation input data, a variety of user input methods, such as text, image, audio, etc., can be captured. The advantage of doing so is that the content of the user's input can be understood more comprehensively, the system's recognition accuracy of the user's intention can be improved, and the feature data of different modes can be aligned, which helps to eliminate the data differences between different modes, making subsequent data processing more accurate and reliable, and improving the robustness and stability of the system; by matching the multimodal feature alignment data, the input data of different modes can be effectively associated, so as to better understand the user's operation intention, which helps to improve the system's understanding and response capabilities to user needs; by performing intent recognition on cross-modal input matching data, the system can more accurately understand the user's intention, and thus better provide services and support for users. This improves the interactive experience between the system and the user, and enhances the user's satisfaction with the system. Based on the identified human-computer interaction intention, building a cross-modal data interaction model helps the system better organize and manage data of different modalities, achieve more efficient data interaction and processing, and improve the overall performance and efficiency of the system. By building a cross-modal data interaction map, the data relationship of different modalities can be visualized, helping the system to better understand the association and connection between data, and providing more possibilities and flexibility for subsequent data analysis and application; by building a cross-modal data interaction network, the system can model the data interaction relationship of different modalities, so as to better understand and process the interaction relationship between multimodal data. This helps to improve the system's ability to understand complex data relationships and enhance the system's ability to integrate and utilize data. Based on the cross-modal data interaction network, a reasonable interaction architecture is designed, which can effectively organize and manage the various modules of the system and achieve efficient operation and management of data interaction. This helps to improve the maintainability and scalability of the system, and is also conducive to the optimization and upgrade of the subsequent system. Performance optimization of the interaction architecture can improve the system's operating efficiency and response speed, so as to better meet user needs. By optimizing the system's architecture design, the system's resource consumption can be reduced and the system's overall performance can be improved. Based on the optimized architecture and model, an efficient cross-modal human-computer interaction engine can be designed to achieve rapid processing and interactive response to multimodal data. This helps to improve the real-time and responsiveness of the system and provide users with a smoother interactive experience. By deploying the interactive engine on the cloud platform, human-computer interaction for intelligent data analysis can be achieved, allowing the system to process data and interact anytime and anywhere. This helps to improve the flexibility and availability of the system, while also better supporting distributed and cross-platform application scenarios.Therefore, the present invention provides a human-computer interaction method based on intelligent data analysis, which is an optimization processing of the traditional human-computer interaction method based on intelligent data analysis. It solves the problem that the traditional human-computer interaction method based on intelligent data analysis has low accuracy of multimodal data interaction, so that multimodal data interaction is difficult to maintain the comfort of user experience. It improves the accuracy of multimodal data interaction and ensures the consistency of multimodal data interaction and user experience comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 A schematic diagram of the steps of a human-computer interaction method based on intelligent data analysis;

[0076] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0077] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0078] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0079] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are 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 technicians in this field without creative work are within the scope of protection of the present invention.

[0080] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0081] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0082] To achieve this, please refer to Figures 1 to 3 , a human-computer interaction method based on intelligent data analysis, the method comprising the following steps:

[0083] Step S1: obtaining user operation input data; performing multimodal feature extraction on the user operation input data to obtain user input multimodal feature data; performing feature alignment on the user input multimodal feature data to obtain multimodal feature alignment data;

[0084] Step S2: performing cross-modal input data matching on the user operation input data according to the multimodal feature alignment data to obtain cross-modal input matching data; performing human-computer interaction intention recognition on the cross-modal input matching data to obtain human-computer interaction intention data; constructing a cross-modal data interaction model according to the human-computer interaction intention data to obtain a cross-modal data interaction model; constructing a cross-modal data interaction graph according to the cross-modal data interaction model to obtain a cross-modal data interaction graph;

[0085] Step S3: constructing a cross-modal data interaction network for the cross-modal data interaction graph to obtain a cross-modal data interaction network; designing an interaction architecture based on the cross-modal data interaction network to obtain a cross-modal data interaction architecture; performing architecture performance optimization processing on the cross-modal data interaction architecture to obtain a cross-modal data interaction performance optimization architecture;

[0086] Step S4: Design an interaction engine based on the cross-modal data interaction performance optimization architecture and the cross-modal data interaction model to obtain a cross-modal human-computer interaction engine, and send the cross-modal human-computer interaction engine to the cloud platform to perform human-computer interaction for intelligent data analysis.

[0087] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of a human-computer interaction method based on intelligent data analysis of the present invention. In this example, the human-computer interaction method based on intelligent data analysis includes the following steps:

[0088] Step S1: obtaining user operation input data; performing multimodal feature extraction on the user operation input data to obtain user input multimodal feature data; performing feature alignment on the user input multimodal feature data to obtain multimodal feature alignment data;

[0089] In an embodiment of the present invention, user operation data is obtained through a user interface or other input methods, such as text input, voice input, and image input. For different types of input data, corresponding feature extraction methods are used. For example, for text input, natural language processing technology is used to extract text features, such as a bag-of-words model and Word2Vec; for voice input, voice signal processing technology is used to extract voice features, such as MFCC (Mel-frequency cepstral coefficients); for image input, computer vision technology is used to extract image features, such as a feature extraction layer of a convolutional neural network (CNN); features of different modalities are aligned to ensure that they are represented in the same feature space for subsequent data processing and analysis, involving some data preprocessing steps, such as normalization and standardization, to ensure consistency and comparability between features.

[0090] Step S2: performing cross-modal input data matching on the user operation input data according to the multimodal feature alignment data to obtain cross-modal input matching data; performing human-computer interaction intention recognition on the cross-modal input matching data to obtain human-computer interaction intention data; constructing a cross-modal data interaction model according to the human-computer interaction intention data to obtain a cross-modal data interaction model; constructing a cross-modal data interaction graph according to the cross-modal data interaction model to obtain a cross-modal data interaction graph;

[0091] In an embodiment of the present invention, multimodal feature alignment data is used to perform cross-modal input data matching, and input data of different modalities are associated and matched to find out the similarities and correlations between them. Based on the cross-modal input matching data, machine learning or deep learning technology is used to identify human-computer interaction intentions, that is, to determine the user's operating intentions or needs. According to the identified human-computer interaction intention data, a cross-modal data interaction model is constructed. The model can be rule-based or based on machine learning or deep learning. The cross-modal data interaction model is used to construct a cross-modal data interaction graph, which describes the relationship between different modal data, interaction methods, and possible response strategies.

[0092] Step S3: constructing a cross-modal data interaction network for the cross-modal data interaction graph to obtain a cross-modal data interaction network; designing an interaction architecture based on the cross-modal data interaction network to obtain a cross-modal data interaction architecture; performing architecture performance optimization processing on the cross-modal data interaction architecture to obtain a cross-modal data interaction performance optimization architecture;

[0093] In an embodiment of the present invention, a cross-modal data interaction network is constructed using graph theory or network analysis technology based on the cross-modal data interaction graph obtained above. This may involve the definition of nodes and edges, as well as the connection relationship between nodes. Based on the structure and characteristics of the cross-modal data interaction network, a cross-modal data interaction architecture is designed, components and modules involved in the interaction process, and the interaction methods and protocols between them are determined, and the performance of the cross-modal data interaction architecture is optimized, including optimizing algorithms and data structures, improving the efficiency and speed of data interaction, considering the scalability and fault tolerance of the system, ensuring normal operation under different scales and loads, and considering resource utilization, such as the rational use of memory and computing resources.

[0094] Step S4: Design an interaction engine based on the cross-modal data interaction performance optimization architecture and the cross-modal data interaction model to obtain a cross-modal human-computer interaction engine, and send the cross-modal human-computer interaction engine to the cloud platform to perform human-computer interaction for intelligent data analysis.

[0095] In an embodiment of the present invention, a cross-modal human-computer interaction engine is designed based on a cross-modal data interaction performance optimization architecture and a cross-modal data interaction model. This may include the following steps: defining the input and output interfaces of the engine, determining the interaction mode between the engine and other components, implementing the processing and conversion of cross-modal data, ensuring effective interaction and conversion between different types of data, designing algorithms and models for parsing user intent, generating responses, and performing corresponding operations, uploading the engine code and related resources to the cloud platform, configuring the cloud platform environment, ensuring that the engine can operate normally and provide the required services, testing and monitoring, ensuring that the engine runs stably on the cloud platform, and adjusting and optimizing as needed.

[0096] The present invention can capture multiple input methods of users, such as text, image, audio, etc., by extracting multimodal features of user operation input data. The advantage of doing so is that the content of user input can be understood more comprehensively, the system's recognition accuracy of user intentions can be improved, and the feature data of different modes can be aligned, which helps to eliminate data differences between different modes, making subsequent data processing more accurate and reliable, and improving the robustness and stability of the system; by matching multimodal feature alignment data, input data of different modes can be effectively associated, so as to better understand the user's operation intentions, which helps to improve the system's understanding and response capabilities to user needs; by performing intent recognition on cross-modal input matching data, the system can more accurately understand the user's intentions, and thus better provide services and support for users. This improves the interactive experience between the system and the user, and enhances the user's satisfaction with the system. Based on the identified human-computer interaction intention, building a cross-modal data interaction model helps the system better organize and manage data of different modalities, achieve more efficient data interaction and processing, and improve the overall performance and efficiency of the system. By building a cross-modal data interaction map, the data relationship of different modalities can be visualized, helping the system to better understand the association and connection between data, and providing more possibilities and flexibility for subsequent data analysis and application; by building a cross-modal data interaction network, the system can model the data interaction relationship of different modalities, so as to better understand and process the interaction relationship between multimodal data. This helps to improve the system's ability to understand complex data relationships and enhance the system's ability to integrate and utilize data. Based on the cross-modal data interaction network, a reasonable interaction architecture is designed, which can effectively organize and manage the various modules of the system and achieve efficient operation and management of data interaction. This helps to improve the maintainability and scalability of the system, and is also conducive to the optimization and upgrade of the subsequent system. Performance optimization of the interaction architecture can improve the system's operating efficiency and response speed, so as to better meet user needs. By optimizing the system's architecture design, the system's resource consumption can be reduced and the system's overall performance can be improved. Based on the optimized architecture and model, an efficient cross-modal human-computer interaction engine can be designed to achieve rapid processing and interactive response to multimodal data. This helps to improve the real-time and responsiveness of the system and provide users with a smoother interactive experience. By deploying the interactive engine on the cloud platform, human-computer interaction for intelligent data analysis can be achieved, allowing the system to process data and interact anytime and anywhere. This helps to improve the flexibility and availability of the system, while also better supporting distributed and cross-platform application scenarios.Therefore, the present invention provides a human-computer interaction method based on intelligent data analysis, which is an optimization processing of the traditional human-computer interaction method based on intelligent data analysis. It solves the problem that the traditional human-computer interaction method based on intelligent data analysis has low accuracy of multimodal data interaction, so that multimodal data interaction is difficult to maintain the comfort of user experience. It improves the accuracy of multimodal data interaction and ensures the consistency of multimodal data interaction and user experience comfort.

[0097] Preferably, step S1 comprises the following steps:

[0098] Step S11: Obtain user operation input data;

[0099] Step S12: classifying the input data type of the user operation input data to obtain input data type data;

[0100] Step S13: extracting multimodal features of the user operation input data according to the input data type data to obtain user input multimodal feature data;

[0101] Step S14: performing feature alignment on the multimodal feature data input by the user to obtain multimodal feature alignment data.

[0102] In an embodiment of the present invention, the user's operation input data is obtained through a suitable input device (such as a keyboard, a mouse, a touch screen, etc.), and the user's voice, image and other non-text input data are captured using corresponding sensors or interfaces, such as a camera, a microphone, etc. The obtained user operation input data is classified using data processing technology to distinguish different types of input data, such as text, image, and voice. Machine learning or rule engines and other methods can be used to classify data according to its characteristics, and feature extraction is performed on each type of input data to convert it into a vector or descriptor representing its characteristics. For text data, features can be extracted using methods such as bag-of-words models and word embedding. For image data, features can be extracted using methods such as convolutional neural networks (CNN). For voice data, features can be extracted using methods such as spectrograms. The user's input data contains multiple types (text, images, voice, etc.), and their features need to be aligned for subsequent processing and analysis. Feature fusion or alignment technology can be used to map different types of features to the same representation space to maintain their consistency and comparability, involving feature normalization, dimensionality reduction, alignment and other operations to ensure that multimodal feature data can effectively represent the user's input information.

[0103] The present invention obtains the user's operation input data through various input channels (such as text input boxes, voice recognition, cameras, etc.). This ensures that the system can capture the user's diverse input, classify the user's input data, and determine the type of each input data, such as text, image, audio, etc. This helps the system to better understand the characteristics and processing methods of different types of data, prepare for subsequent feature extraction and processing, and select appropriate feature extraction methods for different types of input data, such as word embedding or TF-IDF processing for text data, convolutional neural network feature extraction for image data, and Mel spectrum feature extraction for audio data. This helps to convert the original input data into a multimodal feature representation with semantic information, and perform feature alignment on the user input multimodal feature data to obtain multimodal feature alignment data: align the feature data of different modes, eliminate the data differences between different modes, and ensure that they have consistent expressions in the feature space. This helps to improve the generalization ability and robustness of subsequent models, so that the system can understand the user's intentions more accurately.

[0104] Preferably, step S2 comprises the following steps:

[0105] Step S21: performing cross-modal input data matching on the user operation input data according to the multi-modal feature alignment data to obtain cross-modal input matching data;

[0106] Step S22: performing data logical association on the cross-modal input matching data to obtain cross-modal logical association data;

[0107] Step S23: performing human-computer interaction intention recognition on the cross-modal input matching data according to the cross-modal logical association data to obtain human-computer interaction intention data;

[0108] Step S24: constructing a cross-modal data interaction model according to the human-computer interaction intention data and the cross-modal logic association data to obtain a cross-modal data interaction model;

[0109] Step S25: construct a cross-modal data interaction graph according to the cross-modal data interaction model to obtain a cross-modal data interaction graph.

[0110] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0111] Step S21: performing cross-modal input data matching on the user operation input data according to the multi-modal feature alignment data to obtain cross-modal input matching data;

[0112] In an embodiment of the present invention, different features in the multimodal feature alignment data are matched to ensure that they can establish an effective correspondence between different modalities. Various matching algorithms, such as similarity matching, feature vector-based matching, etc., can be used to determine the similarity or correlation between cross-modal features. According to the results of feature matching, data of different modalities are matched to establish an association between cross-modal input data, which involves aligning and matching data in time, space or other dimensions to ensure that they represent the same event or behavior, integrating and merging the matched cross-modal data to obtain cross-modal input matching data, which involves data aggregation, screening or other processing to generate a cross-modal input matching data set with complete information.

[0113] Step S22: performing data logical association on the cross-modal input matching data to obtain cross-modal logical association data;

[0114] In an embodiment of the present invention, based on cross-modal input matching data, data logical association is performed to establish relationships and connections between data. Appropriate association algorithms or rules can be used according to business needs and data characteristics to ensure that the association between data is reasonable and effective, define rules for cross-modal logical association, and clarify the relationships and logical connection methods between data of different modalities. This involves defining conditions, constraints, and logical operations for association to ensure that cross-modal logical association data can meet the needs of actual applications. According to the defined logical association rules, the cross-modal input matching data is processed to generate cross-modal logical association data, including filtering, converting, or aggregating the data, and applying logical association rules to connect and associate the data.

[0115] Step S23: performing human-computer interaction intention recognition on the cross-modal input matching data according to the cross-modal logical association data to obtain human-computer interaction intention data;

[0116] In an embodiment of the present invention, cross-modal logical association data and cross-modal input matching data are collected, and data cleaning and preprocessing are performed to ensure data quality and consistency. Features are extracted from the cross-modal input matching data. These features may involve multiple data types such as text, images, and audio. Feature extraction may adopt traditional rule-based or statistical methods, or deep learning technology may be used for end-to-end feature learning. The cross-modal logical association data is used to annotate the cross-modal input matching data with human-computer interaction intentions. The annotation can be based on previously defined intent categories, such as extracting information and performing operations. The annotated data set is used to train a human-computer interaction intention recognition model. Common models include support vector machines, logistic regression, decision trees, and deep neural networks. The trained model is evaluated using a validation set or cross-validation technology to measure its performance in intent recognition. The human-computer interaction intention data is output, and the prediction result of each sample is usually represented in the form of a label or probability distribution.

[0117] Step S24: constructing a cross-modal data interaction model according to the human-computer interaction intention data and the cross-modal logic association data to obtain a cross-modal data interaction model;

[0118] In an embodiment of the present invention, feature engineering is performed in combination with human-computer interaction intention data and cross-modal logical association data to extract effective feature information, which includes feature selection, dimensionality reduction, combination and other processing to improve the performance of the cross-modal data interaction model. The cross-modal data interaction model is trained using labeled human-computer interaction intention data and cross-modal logical association data. Various machine learning or deep learning methods can be used to train the model to achieve effective interaction between cross-modal data. The trained cross-modal data interaction model is evaluated to test its performance in intent recognition and data interaction. According to the evaluation results, the model is optimized and adjusted to improve its accuracy and generalization ability.

[0119] Step S25: construct a cross-modal data interaction graph according to the cross-modal data interaction model to obtain a cross-modal data interaction graph.

[0120] In an embodiment of the present invention, a cross-modal data interaction model and cross-modal logical association data are combined to establish an association relationship between data. This can be achieved by identifying the interaction patterns and rules between data. Based on the associated data, a cross-modal data interaction graph is constructed. The graph can be a structured graphical representation, in which nodes represent data elements and edges represent the relationship between data. The constructed cross-modal data interaction graph is optimized and improved to improve its usability and efficiency. The graph is applied to actual human-computer interaction systems to support effective interaction and application between cross-modal data.

[0121] The present invention uses multimodal feature alignment data to match different modal data input by the user and find data associated with each other. This helps to convert the user's cross-modal input into a higher-level semantic representation, provide richer information to support subsequent processing, logically associate the cross-modal input matching data, analyze the relationship between them, and form cross-modal logical association data. This helps the system understand the logical relationship between different modalities, further improves the ability to understand user intentions, and based on cross-modal logical association data, performs human-computer interaction intention recognition on cross-modal input matching data, that is, infers the user's interaction intention between different modalities. This helps the system to understand the user's intention more accurately, thereby providing more personalized and accurate services, combining human-computer interaction intention data and cross-modal logical association data, and constructing a cross-modal data interaction model for describing and predicting cross-modal interactions between users and systems. This helps the system better understand the user's behavior pattern, provide a more intelligent and personalized interactive experience, and use the cross-modal data interaction model to construct a cross-modal data interaction map to describe the relationship between different modal data, user intentions, and system responses. This helps the system better manage and understand the interaction process of cross-modal data, and provide users with more intelligent services and support.

[0122] Preferably, step S23 includes the following steps:

[0123] Step S231: performing context-aware analysis on the cross-modal logical association data to obtain logical association context-aware data;

[0124] Step S232: performing a perceptual multidimensional analysis on the cross-modal logical association data according to the logical association context perception data to obtain logical association multidimensional perception data;

[0125] Step S233: performing cluster analysis on the logically associated multidimensional perception data to obtain logically associated multidimensional cluster data; performing multidimensional category structure analysis on the logically associated multidimensional perception data according to the logically associated multidimensional cluster data to obtain logically associated multidimensional category structure data;

[0126] Step S234: performing structural linear logic adjustment on the logically associated multidimensional perception data according to the logically associated multidimensional category structure data to obtain logical structural linear adjustment data;

[0127] Step S235: performing cross-modal input content linear correlation analysis on the cross-modal input matching data according to the logical structure linear adjustment data to obtain cross-modal content linear correlation data; performing content correlation time interval exponential distribution calculation on the cross-modal content linear correlation data to obtain content correlation exponential distribution data;

[0128] Step S236: Based on the content-related exponential distribution data, a content distribution Poisson process simulation is performed on the cross-modal content linear correlation data to obtain content distribution Poisson process data; based on the content distribution Poisson process data, human-computer interaction intention recognition is performed on the cross-modal content linear correlation data to obtain human-computer interaction intention data.

[0129] In the embodiment of the present invention, cross-modal logically associated data are collected, including multiple data types such as text, image, audio, etc., and the collection of context information is considered, which may include time, location, and user status. Context-related features such as timestamp, geographic location, and user behavior are extracted from the cross-modal logically associated data. These features help to identify the correlation and context information between data. The extracted context features are used to analyze the context information in the cross-modal logically associated data. Statistical analysis, association rule mining, and other technologies can be used to identify the context correlation between data. The logically associated context-aware data obtained in step S231 is used as input to extract context-related features from the logically associated data. Extract multidimensional features from context-sensitive data. These features involve multiple data types such as text, images, and audio. Use the extracted multidimensional features to analyze logically related data and explore changes and associations in multiple dimensions. Statistical methods, machine learning techniques, or data mining algorithms can be used for analysis. The logically related multidimensional perception data obtained in step S232 is used as input to perform cluster analysis on the multidimensional perception data and classify similar data points into the same category. Traditional clustering algorithms such as K-means clustering and hierarchical clustering can be used. Clustering methods in deep learning can also be tried. According to the clustering results, the logically related multidimensional perception data can be analyzed. The multidimensional category structure of the data, that is, the relationship and characteristics between different categories, can be understood through visualization technology or statistical analysis to understand the similarities and differences between different categories; the logically associated multidimensional category structure data obtained in step S233 is used as input, and the logically associated multidimensional perception data is structurally linearly logically adjusted according to the logically associated multidimensional category structure data. The purpose of the adjustment is to make the data more consistent with the expected linear relationship in the logical structure, so as to facilitate subsequent analysis, collect the logically associated multidimensional category structure data, and organize it into a usable data format, analyze the structure of the logically associated multidimensional perception data, and determine the linear adjustment operation that needs to be performed, according to the collected The method comprises the following steps: first, applying appropriate linear adjustment algorithms, such as linear regression or principal component analysis, to obtain logical structure linear adjustment data from a set of structural data, and then obtaining cross-modal input matching data, such as text-image matching data or audio-text matching data. The logical structure linear adjustment data is used to perform linear association analysis on the cross-modal input matching data to discover the potential association between them. The content association time interval exponential distribution of the cross-modal content linear association data is calculated based on the analysis results. Statistical tools or algorithms can be used to complete this step. Based on the content-related exponential distribution data, the Poisson process simulation technology is used to simulate the content distribution of the cross-modal content linear association data.This may involve using a random number generator and a simulation algorithm to simulate the characteristics of the content distribution, comparing the simulated content distribution Poisson process data with the actual cross-modal content linear association data to verify the accuracy of the simulation, and based on the simulated data and actual data, using machine learning or rule engines and other methods to perform human-computer interaction intention recognition on the cross-modal content linear association data to determine the user's intentions or behavior patterns.

[0130] In this step of the present invention, the system takes into account contextual information, such as the environment, time, and location of the user's operation, when analyzing the cross-modal logical association data. This helps to better understand the context and user intention behind the data. By performing multidimensional analysis on the logical association data, the system can identify different dimensions in the data, thereby more comprehensively understanding the characteristics and meaning of the data. This helps to improve the system's ability to understand user behavior. Through cluster analysis, the system divides the logical association multidimensional perception data into different categories, and further clarifies the inherent structure and relationship between the data. This helps the system to better organize and manage data, and provides a basis for subsequent analysis and processing. By linearly adjusting the logical association multidimensional category structure data, the system can optimize the structure of the data to make it more in line with the needs of analysis and processing. This helps to improve the efficiency and accuracy of the system's use of data. By performing content linear association analysis and time interval exponential distribution calculation on cross-modal input matching data, the system can identify the degree and frequency of association between different data. This helps to understand the temporal relationship and correlation between data, and provides an important reference for subsequent analysis and modeling. Through content distribution Poisson process simulation, the system can simulate the interaction process between the user and the system, and then identify the user's interaction intention. This helps the system understand user needs and intentions more accurately and provide smarter and more personalized services.

[0131] Preferably, step S24 includes the following contents:

[0132] Step S241: capturing dependency relationships of cross-modal logic association data based on a self-attention mechanism to obtain cross-modal logic dependency relationship data;

[0133] Step S242: performing cross-modal causal relationship analysis on the cross-modal logic dependency relationship data according to the human-computer interaction intention data to obtain cross-modal interaction logic causal data;

[0134] Step S243: performing interactive dependency transfer learning on the human-computer interaction intention data and the cross-modal logical dependency relationship data according to the interactive logic dependency causal data to obtain interactive dependency transfer learning data;

[0135] Step S244: performing data volume incremental learning processing on the interactive dependency transfer learning data to obtain transfer learning incremental data;

[0136] Step S245: construct a cross-modal data interaction model based on the transfer learning incremental data and the cross-modal interaction logic causal data according to the recursive neural network to obtain a cross-modal data interaction model.

[0137] In an embodiment of the present invention, a self-attention mechanism model, such as a Transformer model, is used to process cross-modal logical association data to capture the dependency relationship between each modality, the data of each modality is input into the self-attention mechanism model, and the attention mechanism of the model is used to capture the association information between different modalities to obtain cross-modal logical dependency relationship data, and the existing human-computer interaction intention data is used in combination with the cross-modal logical dependency relationship data to perform cross-modal causal relationship analysis, and the causal relationship between each modality is analyzed using statistical methods or causal reasoning algorithms to obtain cross-modal interactive logical causal data, and based on the interactive logical dependency causal data, the human-computer interaction intention data and the cross-modal logical dependency relationship data are interactively dependently transferred using a transfer learning method, and in the transfer During the learning process, the existing interactive logic-dependent causal data is used to guide new task learning to improve the performance of the model on new tasks and obtain interactive dependency transfer learning data. The interactive dependency transfer learning data is processed with incremental learning to further improve the generalization ability of the model. An incremental learning algorithm can be used to gradually introduce new data and update model parameters so that the model can adapt to the changing data. Based on recursive neural networks (RNNs) or their variants, such as long short-term memory networks (LSTMs) or gated recurrent units (GRUs), the transfer learning incremental data and cross-modal interactive logic causal data are processed to construct a cross-modal data interaction model. RNNs or their variants are used to model the interaction relationship between different modalities to achieve more accurate cross-modal data interaction analysis.

[0138] The self-attention mechanism of the present invention can help the system effectively capture the dependencies between various data when processing cross-modal logically associated data. The capture of such dependencies helps to understand the intrinsic connection and influence between data, and provides an important basis for subsequent analysis and modeling. Through the causal analysis of logical dependencies on human-computer interaction intention data, the system can identify the causal relationship between different data, that is, how the change of one data leads to the change of another data. This helps to understand the driving factors behind user behavior and provide more accurate predictions and responses for the system. Interactive dependency transfer learning can help the system transfer knowledge between different data sets, thereby improving the performance of the model on new tasks. By performing transfer learning on interactive logical dependency causal data, the system can better utilize existing knowledge to improve the understanding and analysis capabilities of human-computer interaction intention data and cross-modal logical dependency data. Through incremental learning of data volume, the system can continuously improve and optimize the model while continuously receiving new data. This helps to maintain the robustness and adaptability of the model to cope with changing environments and needs. Recursive neural networks can effectively process sequence data and complex data dependencies when building cross-modal data interaction models. By inputting transfer learning incremental data and cross-modal interaction logic causal data into the recursive neural network, the system can build a more intelligent and flexible model, thereby achieving more accurate understanding and response to human-computer interaction intentions.

[0139] Preferably, step S3 comprises the following steps:

[0140] Step S31: constructing a cross-modal data interaction network for the cross-modal data interaction graph to obtain a cross-modal data interaction network;

[0141] Step S32: Designing an interaction architecture according to the cross-modal data interaction network and the cross-modal data interaction graph to obtain a cross-modal data interaction architecture;

[0142] Step S33: Designing a parallel framework for the cross-modal data interaction architecture to obtain an interactive parallel framework;

[0143] Step S34: Optimize the architecture performance of the cross-modal data interaction architecture according to the interactive parallel framework to obtain a cross-modal data interaction performance optimized architecture.

[0144] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0145] Step S31: constructing a cross-modal data interaction network for the cross-modal data interaction graph to obtain a cross-modal data interaction network;

[0146] In an embodiment of the present invention, cross-modal data is collected and represented in the form of an interaction graph, in which nodes represent data of different modalities and edges represent the interaction relationships between them. Based on the collected data, a cross-modal data interaction network is constructed using a graph theory algorithm, and the interaction relationships between the modalities are mapped to connection relationships in the network.

[0147] Step S32: Designing an interaction architecture according to the cross-modal data interaction network and the cross-modal data interaction graph to obtain a cross-modal data interaction architecture;

[0148] In an embodiment of the present invention, an interaction architecture is designed based on the constructed cross-modal data interaction network to determine how to use these interaction relationships to support various application scenarios. According to the cross-modal data interaction graph, the structure of the interaction architecture is designed, including the functions and connection methods of the nodes, to achieve effective cross-modal data interaction.

[0149] Step S33: Designing a parallel framework for the cross-modal data interaction architecture to obtain an interactive parallel framework;

[0150] In an embodiment of the present invention, a parallel framework is designed to realize parallel processing of a cross-modal data interaction architecture, improve the operating efficiency and performance of the system, and consider using parallel computing technologies, such as GPU acceleration or distributed computing, to process large-scale cross-modal data and increase computing speed.

[0151] Step S34: Optimize the architecture performance of the cross-modal data interaction architecture according to the interactive parallel framework to obtain a cross-modal data interaction performance optimized architecture.

[0152] In an embodiment of the present invention, the cross-modal data interaction architecture is performance optimized based on the interactive parallel framework to improve the efficiency and response speed of the system. The architecture performance is optimized by adjusting algorithms, optimizing data flows, reducing communication costs, etc., to ensure that the system can maintain efficient and stable operation when processing large-scale cross-modal data.

[0153] In this step of the present invention, by constructing a cross-modal data interaction network, the system can effectively integrate and interact data of different modes. The establishment of such a network helps to achieve cross-modal association and information transmission between data, thereby improving the utilization efficiency and overall performance of data. By designing an interactive architecture based on a cross-modal data interaction network and a data interaction graph, the system can better organize and manage the relationship between data and improve the scalability and flexibility of the system. This architecture design can effectively support the interaction and integration between different modal data, laying the foundation for subsequent parallel processing and performance optimization. By designing a parallel framework, the system can make full use of multi-core processors or distributed computing resources to achieve efficient processing and calculation of the cross-modal data interaction architecture. This parallel framework design helps to improve the processing speed and throughput of the system, thereby responding to user requests and providing services more quickly. By performing performance optimization processing on the interactive parallel framework, the system can further improve the performance and efficiency of the cross-modal data interaction architecture. This includes optimization measures in terms of optimizing algorithms, adjusting parameters, reducing computing and communication overhead, etc., to ensure that the system can still operate efficiently in large-scale data and complex task scenarios.

[0154] Preferably, step S32 includes the following steps:

[0155] Step S321: performing a node topology structure analysis on the cross-modal data interaction graph to obtain the interaction graph node topology data; constructing a node connectivity graph on the interaction graph node topology data to obtain the interaction graph node connectivity graph;

[0156] Step S322: performing a depth search on the interaction graph node connectivity graph to obtain node connectivity graph depth data; performing a node depth-first search on the interaction graph node connectivity graph according to the node connectivity graph depth data to obtain the interaction graph node connectivity data;

[0157] Step S323: Perform structural sparse distribution law analysis on the interaction graph node connectivity data to obtain structural sparse distribution data; perform degree distribution structure analysis on the structural sparse distribution data to obtain structural sparse degree distribution data;

[0158] Step S324: establishing a data connection channel for the structure sparsity distribution data according to the cross-modal data interaction network to obtain a cross-modal data connection channel;

[0159] Step S325: performing bandwidth matching on the cross-modal data connection channel according to the structural sparsity distribution data to obtain connection bandwidth matching data;

[0160] Step S326: Designing a connection route for the cross-modal data connection channel according to the connection bandwidth matching data and the structure sparsity distribution data to obtain cross-modal structure routing data;

[0161] Step S327: Design an interaction architecture according to the cross-modal data connection channel, the connection bandwidth matching data, and the cross-modal structure routing data to obtain a cross-modal data interaction architecture.

[0162] In an embodiment of the present invention, a graph theory algorithm such as a shortest path algorithm or a minimum spanning tree algorithm is used to analyze the node topology structure of a cross-modal data interaction graph. According to the analysis results, the interaction graph node topology data is obtained, including the connection relationship and topological structure information between the nodes. The node connectivity graph is constructed using the topological data, wherein each node represents a data modality and the edge represents the interaction relationship between the modalities. A depth-first search (DFS) algorithm or a breadth-first search (BFS) algorithm is used to perform a deep search on the interaction graph node connectivity graph. According to the deep search results, the depth data of the node connectivity graph is obtained, including the depth information and search path of each node. The node depth-first search algorithm is used to traverse the interaction graph node connectivity graph to obtain the connection relationship and path information between the nodes. A statistical method or a machine learning algorithm is used to analyze the structural sparse distribution law of the interaction graph node connectivity data. According to the analysis results, the structural sparse distribution data is obtained, including the sparse connection status and distribution law between the nodes. The degree distribution structure analysis of the structural sparse distribution data is performed to calculate each node. degree and degree distribution to understand the connection density and structural characteristics between nodes; determine the nodes and channel paths for establishing data connection channels based on the cross-modal data interaction network and structural sparsity distribution data, use graph theory algorithms or network modeling methods to establish data connection channels in the cross-modal data interaction network to ensure that cross-modal data can be effectively transmitted and interacted, evaluate the bandwidth requirements of each data connection channel based on the structural sparsity distribution data, perform bandwidth matching on the cross-modal data connection channels to ensure that the bandwidth of each channel meets the corresponding data transmission requirements, design a routing solution for the cross-modal data connection channel based on the connection bandwidth matching data and the structural sparsity distribution data, determine the optimal data transmission path based on the channel bandwidth and the connection relationship between nodes to maximize the optimization of data interaction efficiency; design a cross-modal data interaction architecture based on the cross-modal data connection channel, the connection bandwidth matching data and the cross-modal structural routing data, and take into account the channel bandwidth, data transmission path and the connection relationship between nodes during the design process to ensure that the designed interaction architecture can meet the data interaction requirements in various application scenarios.

[0163] The present invention obtains the node topology data of the interaction graph through the node topology structure analysis of the cross-modal data interaction graph, and constructs the node connectivity graph, so that the system can better understand the relationship and connection mode between different data. This helps to identify the important nodes and key paths between the data, thereby providing an important reference for subsequent data interaction and processing. By performing a deep search on the node connectivity graph of the interaction graph and performing a node depth-first search, the system can deeply explore the connection structure and hierarchical relationship between the data. This helps to discover the potential associations hidden between the data and provide support for the effectiveness and accuracy of data interaction. By performing a structural sparse distribution law analysis and a degree distribution structure analysis on the node connectivity data of the interaction graph, the system can understand the distribution law and characteristics between the data. This helps to identify the structural sparsity in the data interaction, thereby providing guidance for the establishment and optimization of subsequent data connection channels. According to the cross-modal data interaction network, a data connection channel is established so that data of different modes can be effectively interacted and transmitted. This helps to achieve rapid communication and information exchange between data, and improves data utilization and overall performance. By matching the bandwidth of the structured sparsity distribution data, the system can better configure the bandwidth resources of the connection channel to ensure efficient transmission and communication during data interaction. This helps to avoid data congestion and transmission delays, and improve the system's response speed and user experience. By connecting bandwidth matching data and structured sparsity distribution data for connection routing design, the system can optimize the data transmission path and achieve the best data interaction effect. This helps to reduce the path length and transmission delay of data transmission, and improve the efficiency and reliability of data interaction. By comprehensively considering cross-modal data connection channels, connection bandwidth matching data, and cross-modal structure routing data, the system can design an interaction architecture to achieve efficient interaction and integration of cross-modal data. This helps to build an optimized data interaction architecture, improve the performance and scalability of the system, and provide users with a better service experience.

[0164] Preferably, designing a connection route for a cross-modal data connection channel includes:

[0165] According to the connection bandwidth matching data and the structure sparsity distribution data, the structure demand capacity is evaluated to obtain the structure connection demand capacity data;

[0166] Based on the structural connection demand capacity data, the structural sparsity distribution data is sorted for connection structure demand to obtain connection structure demand sorting data; according to the connection structure demand sorting data, the structural sparsity distribution data is sorted and extracted to obtain structure distribution rank data;

[0167] Sorting and indexing the structure distribution ranking data based on the hash algorithm to obtain the structure data distribution index;

[0168] Combining routing addresses according to the structural distribution ranking data and the structural data distribution index to obtain structural distribution routing address data;

[0169] The connection routing is designed for the cross-modal data connection channel according to the structural distribution routing address data to obtain the cross-modal structural routing data.

[0170] In an embodiment of the present invention, the structural demand capacity of various data types is evaluated according to the connection bandwidth, the data transmission rate and structural sparsity are considered, and the structural sparsity distribution data is sorted based on the structural demand capacity data to determine the priority of the connection structure demand, the data is sorted according to the connection structure demand, the sorting rank of the structural sparsity distribution data is extracted, the relative importance and priority of the data are determined, and the sorting rank data is indexed and generated using a hash algorithm to quickly access and retrieve the distribution of the structural data, and a valid routing address combination is generated in combination with the structural distribution rank data and the structural data distribution index to ensure that the data is transmitted according to the priority and demand, and the routing of the cross-modal data connection channel is designed according to the structural distribution routing address data to ensure the efficiency and stability of data transmission.

[0171] The present invention can accurately determine the structural connection capacity required by the cross-modal data connection channel by evaluating the connection bandwidth matching data and the structural sparsity distribution data. This helps to ensure that there will be no bottlenecks or delays in the data transmission process, and ensure the efficiency and reliability of data interaction. According to the structural connection demand capacity data, the system can sort the structural sparsity distribution data to determine the priority and demand of the connection channel. This helps to optimize the data transmission path and improve the efficiency and priority of data interaction. By extracting the position of the connection structure demand sorting data, the system can understand the priority and importance of different data connection channels, so as to better plan the data transmission path and resource allocation. Based on the hash algorithm to generate the sorting index of the structural distribution position data, the system can quickly locate and access the data connection channel, and improve the speed and efficiency of data interaction. According to the structural distribution position data and the structural data distribution index, the system can effectively combine the routing address and determine the best data transmission path and routing strategy. This helps to reduce the delay and energy consumption of data transmission and improve the overall performance and stability of the system. Based on the structural distribution routing address data, the system can design the optimal cross-modal data connection route to ensure efficient interaction and communication between different data modes. This helps improve the scalability and flexibility of the system and meet the needs of different data processing scenarios.

[0172] Preferably, step S33 includes the following steps:

[0173] Step S331: performing high-concurrency traffic simulation on the cross-modal data interaction architecture to obtain high-concurrency traffic simulation data;

[0174] Step S332: performing concurrent transmission channel configuration according to high concurrent traffic simulation data to obtain concurrent transmission channels;

[0175] Step S333: Designing a buffer mechanism for high concurrent traffic simulation data to obtain a concurrent traffic buffer mechanism;

[0176] Step S334: constructing a buffered transmission channel for the concurrent transmission channel according to the concurrent traffic buffering mechanism to obtain a concurrent buffered transmission channel;

[0177] Step S335: Optimizing the concurrent buffer structure of the Apache Flink framework based on the concurrent buffer transmission channel to obtain a structurally optimized Apache Flink framework;

[0178] Step S336: Design a parallel framework based on the structurally optimized Apache Flink framework to obtain an interactive parallel framework.

[0179] In an embodiment of the present invention, a high-concurrency traffic simulation is performed on a cross-modal data interaction architecture, and the interaction modes and frequencies of different types of data and possible high-concurrency situations are considered to generate high-concurrency traffic simulation data. Based on the high-concurrency traffic simulation data, an appropriate number and type of concurrent transmission channels are configured to meet the needs of high-concurrency data interaction. Considering factors such as bandwidth and delay of data transmission, the parameters and configuration of the transmission channel are reasonably set, the characteristics and volatility of the high-concurrency traffic simulation data are analyzed, and a suitable buffer mechanism is designed to balance the rate and processing capacity of data traffic. Considering the delay and processing capacity of data processing, a buffer mechanism is designed to adjust the transmission rate and processing speed of data. In combination with the designed buffer mechanism, a corresponding buffer transmission channel is constructed for the concurrent transmission channel to ensure smooth and efficient data transmission to meet the needs of high-concurrency traffic simulation data. Based on the designed concurrent buffer transmission channel, the Apache Flink framework is structurally optimized to meet the needs of high-concurrency data processing and transmission, and the data processing and transmission modules of the Apache Flink framework are optimized to improve its concurrent processing capability and data processing efficiency. According to the optimized Apache Flink framework, a parallel framework adapted to high-concurrency data processing is designed to improve the concurrency and efficiency of data processing. Based on the characteristics and optimized structure of the Flink framework, we design and implement an efficient parallel data processing framework to meet the needs of cross-modal data interaction.

[0180] The present invention simulates high-concurrency traffic in a cross-modal data interaction architecture. The system can simulate high-concurrency traffic conditions in various data interaction scenarios, including information such as data transmission rate and request response time. This helps the system evaluate its performance under high load conditions, discover potential bottlenecks and performance bottlenecks, and provide a reference for subsequent optimization work. According to the high-concurrency traffic simulation data, the system can configure concurrent transmission channels to meet the high-concurrency data transmission requirements. By reasonably configuring the transmission channels, the data transmission efficiency and stability can be improved, and the reliability and performance of the system under high load conditions can be ensured. A buffering mechanism is designed for the high-concurrency traffic simulation data. The system can achieve smooth flow of data transmission and avoid data congestion and loss. The design of the buffering mechanism helps to adjust the data transmission rate, balance the data interaction between producers and consumers, and improve the stability and throughput of the system. According to the concurrent traffic buffering mechanism, the system can construct a concurrent buffering transmission channel, combine the buffering mechanism with the transmission channel, and achieve efficient data transmission, which helps to reduce delays and fluctuations in the data transmission process and improve the reliability and stability of data transmission. Based on the concurrent buffering transmission channel, Apache Structural optimization of the Flink framework can improve the performance of the framework in high-concurrency scenarios. Optimizing the structure of the Apache Flink framework can accelerate the data processing and analysis process and improve the real-time and responsiveness of the system. Based on the structurally optimized Apache Flink framework, the system can design an interactive parallel framework to achieve parallel and concurrency in data interaction and processing, which helps to improve the system's processing power and efficiency, support more complex data processing tasks and scenarios, and meet users' real-time data interaction needs.

[0181] Preferably, the buffering mechanism design for high concurrent traffic simulation data includes:

[0182] Perform buffer pool capacity evaluation on high concurrent traffic simulation data to obtain buffer pool capacity evaluation data;

[0183] Memory space allocation is performed according to the buffer pool capacity assessment data to obtain memory space allocation data;

[0184] Dynamically adjust temporary memory space for high concurrent traffic simulation data according to buffer pool capacity assessment data and memory space allocation data to obtain temporary memory space adjustment data;

[0185] A buffer mechanism is designed for high concurrent traffic simulation data based on buffer pool capacity assessment data, memory space allocation data and temporary memory space adjustment data to obtain a concurrent traffic buffer mechanism.

[0186] In an embodiment of the present invention, high-concurrency traffic simulation data is analyzed to evaluate the amount of data generated and its volatility. According to the characteristics and requirements of the data traffic, a suitable buffer pool capacity is determined. According to the buffer pool capacity evaluation data, a memory space of appropriate size is allocated to accommodate the buffer data. Taking into account system resource limitations and data processing capabilities, the memory space is reasonably allocated to ensure efficient data buffering and processing. The actual traffic situation is monitored, and the size of the temporary memory space is flexibly adjusted according to the dynamic adjustment strategy. When the traffic increases or decreases sharply, the memory space is dynamically adjusted according to real-time needs to maintain the buffering effect and system stability. In combination with the buffer pool capacity evaluation data, the memory space allocation data and the temporary memory space adjustment data, a buffer mechanism suitable for high-concurrency traffic simulation data is designed. Taking into account the speed and delay of data processing, a suitable buffering strategy is designed to balance the speed of data transmission and processing.

[0187] The present invention evaluates the buffer pool capacity for high-concurrency flow simulation data, so that the system can understand the amount of data that needs to be buffered under high load conditions and the fluctuation of data flow, which helps to determine the appropriate buffer pool capacity and avoid data loss or system performance degradation caused by insufficient buffer pool capacity. According to the buffer pool capacity evaluation data, the system can allocate memory space to ensure that the system has enough memory for buffering high-concurrency flow simulation data. Correct memory space allocation can avoid system crash or performance degradation caused by memory overflow or insufficient memory, and ensure the stability and reliability of the system. According to the buffer pool capacity evaluation data and memory space allocation data, the system can dynamically adjust the temporary memory space for high-concurrency flow simulation data, which means that the system can dynamically adjust the memory allocation according to actual conditions to adapt to flow fluctuations or sudden increases in data volume, thereby maintaining the stability and efficiency of the system. In combination with the buffer pool capacity evaluation data, the memory space allocation data and the temporary memory space adjustment data, the system can design a buffer mechanism for high-concurrency flow, and this buffer mechanism can ensure stable transmission of data under high concurrency conditions, avoid data congestion or loss, and ensure the high availability and reliability of the system.

[0188] The present invention can capture multiple input methods of users, such as text, image, audio, etc., by extracting multimodal features of user operation input data. The advantage of doing so is that the content of user input can be understood more comprehensively, the system's recognition accuracy of user intentions can be improved, and the feature data of different modes can be aligned, which helps to eliminate data differences between different modes, making subsequent data processing more accurate and reliable, and improving the robustness and stability of the system; by matching multimodal feature alignment data, input data of different modes can be effectively associated, so as to better understand the user's operation intentions, which helps to improve the system's understanding and response capabilities to user needs; by performing intent recognition on cross-modal input matching data, the system can more accurately understand the user's intentions, and thus better provide services and support for users. This improves the interactive experience between the system and the user, and enhances the user's satisfaction with the system. Based on the identified human-computer interaction intention, building a cross-modal data interaction model helps the system better organize and manage data of different modalities, achieve more efficient data interaction and processing, and improve the overall performance and efficiency of the system. By building a cross-modal data interaction map, the data relationship of different modalities can be visualized, helping the system to better understand the association and connection between data, and providing more possibilities and flexibility for subsequent data analysis and application; by building a cross-modal data interaction network, the system can model the data interaction relationship of different modalities, so as to better understand and process the interaction relationship between multimodal data. This helps to improve the system's ability to understand complex data relationships and enhance the system's ability to integrate and utilize data. Based on the cross-modal data interaction network, a reasonable interaction architecture is designed, which can effectively organize and manage the various modules of the system and achieve efficient operation and management of data interaction. This helps to improve the maintainability and scalability of the system, and is also conducive to the optimization and upgrade of the subsequent system. Performance optimization of the interaction architecture can improve the system's operating efficiency and response speed, so as to better meet user needs. By optimizing the system's architecture design, the system's resource consumption can be reduced and the system's overall performance can be improved. Based on the optimized architecture and model, an efficient cross-modal human-computer interaction engine can be designed to achieve rapid processing and interactive response to multimodal data. This helps to improve the real-time and responsiveness of the system and provide users with a smoother interactive experience. By deploying the interactive engine on the cloud platform, human-computer interaction for intelligent data analysis can be achieved, allowing the system to process data and interact anytime and anywhere. This helps to improve the flexibility and availability of the system, while also better supporting distributed and cross-platform application scenarios.Therefore, the present invention provides a human-computer interaction method based on intelligent data analysis, which is an optimization processing of the traditional human-computer interaction method based on intelligent data analysis. It solves the problem that the traditional human-computer interaction method based on intelligent data analysis has low accuracy of multimodal data interaction, so that multimodal data interaction is difficult to maintain the comfort of user experience. It improves the accuracy of multimodal data interaction and ensures the consistency of multimodal data interaction and user experience comfort.

[0189] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0190] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A human-computer interaction method based on intelligent data analysis, It is characterized in that The following steps are involved: Step S1: Obtain user operation input data; Performing multimodal feature extraction on user operation input data to obtain user input multimodal feature data; Performing feature alignment on the multimodal feature data input by the user to obtain multimodal feature alignment data; Step S2: performing cross-modal input data matching on the user operation input data according to the multi-modal feature alignment data to obtain cross-modal input matching data; Perform human-computer interaction intention recognition on cross-modal input matching data to obtain human-computer interaction intention data; A cross-modal data interaction model is constructed according to the human-computer interaction intention data to obtain a cross-modal data interaction model; A cross-modal data interaction graph is constructed according to the cross-modal data interaction model to obtain a cross-modal data interaction graph; wherein step S2 includes: Step S21: performing cross-modal input data matching on the user operation input data according to the multi-modal feature alignment data to obtain cross-modal input matching data; Step S22: performing data logical association on the cross-modal input matching data to obtain cross-modal logical association data; Step S23: performing human-computer interaction intention recognition on the cross-modal input matching data according to the cross-modal logical association data to obtain human-computer interaction intention data; Step S24: constructing a cross-modal data interaction model according to the human-computer interaction intention data and the cross-modal logic association data to obtain a cross-modal data interaction model; wherein step S24 includes: Step S241: capturing dependency relationships of cross-modal logic association data based on a self-attention mechanism to obtain cross-modal logic dependency relationship data; Step S242: performing cross-modal causal relationship analysis on the cross-modal logic dependency relationship data according to the human-computer interaction intention data to obtain cross-modal interaction logic causal data; Step S243: performing interactive dependency transfer learning on the human-computer interaction intention data and the cross-modal logical dependency relationship data according to the interactive logic dependency causal data to obtain interactive dependency transfer learning data; Step S244: performing data volume incremental learning processing on the interactive dependency transfer learning data to obtain transfer learning incremental data; Step S245: constructing a cross-modal data interaction model for the transfer learning incremental data and the cross-modal interaction logic causal data according to the recursive neural network to obtain a cross-modal data interaction model; Step S25: constructing a cross-modal data interaction graph according to the cross-modal data interaction model to obtain a cross-modal data interaction graph; Step S3: constructing a cross-modal data interaction network for the cross-modal data interaction graph to obtain a cross-modal data interaction network; designing an interaction architecture based on the cross-modal data interaction network to obtain a cross-modal data interaction architecture; performing architecture performance optimization processing on the cross-modal data interaction architecture to obtain a cross-modal data interaction performance optimization architecture; wherein step S3 includes: Step S31: constructing a cross-modal data interaction network for the cross-modal data interaction graph to obtain a cross-modal data interaction network; Step S32: Designing an interaction architecture according to the cross-modal data interaction network and the cross-modal data interaction graph to obtain a cross-modal data interaction architecture; wherein step S32 includes: Step S321: performing a node topology structure analysis on the cross-modal data interaction graph to obtain the interaction graph node topology data; constructing a node connectivity graph on the interaction graph node topology data to obtain the interaction graph node connectivity graph; Step S322: performing a depth search on the interaction graph node connectivity graph to obtain node connectivity graph depth data; performing a node depth-first search on the interaction graph node connectivity graph according to the node connectivity graph depth data to obtain the interaction graph node connectivity data; Step S323: Perform structural sparse distribution law analysis on the interaction graph node connectivity data to obtain structural sparse distribution data; perform degree distribution structure analysis on the structural sparse distribution data to obtain structural sparse degree distribution data; Step S324: establishing a data connection channel for the structure sparsity distribution data according to the cross-modal data interaction network to obtain a cross-modal data connection channel; Step S325: performing bandwidth matching on the cross-modal data connection channel according to the structural sparsity distribution data to obtain connection bandwidth matching data; Step S326: Designing a connection route for the cross-modal data connection channel according to the connection bandwidth matching data and the structure sparsity distribution data to obtain cross-modal structure routing data; Step S327: Designing an interaction architecture according to the cross-modal data connection channel, the connection bandwidth matching data, and the cross-modal structure routing data to obtain a cross-modal data interaction architecture; Step S33: Designing a parallel framework for the cross-modal data interaction architecture to obtain an interactive parallel framework; Step S34: performing architecture performance optimization processing on the cross-modal data interaction architecture according to the interactive parallel framework to obtain a cross-modal data interaction performance optimization architecture; Step S4: Design an interaction engine based on the cross-modal data interaction performance optimization architecture and the cross-modal data interaction model to obtain a cross-modal human-computer interaction engine, and send the cross-modal human-computer interaction engine to the cloud platform to perform human-computer interaction for intelligent data analysis.

2. The human-computer interaction method based on intelligent data analysis according to claim 1, It is characterized in that Step S1 includes the following steps: Step S11: Obtain user operation input data; Step S12: classifying the input data type of the user operation input data to obtain input data type data; Step S13: extracting multimodal features of the user operation input data according to the input data type data to obtain user input multimodal feature data; Step S14: performing feature alignment on the multimodal feature data input by the user to obtain multimodal feature alignment data.

3. The human-computer interaction method based on intelligent data analysis according to claim 1, It is characterized in that Step S23 includes the following steps: Step S231: performing context-aware analysis on the cross-modal logical association data to obtain logical association context-aware data; Step S232: performing a perceptual multidimensional analysis on the cross-modal logical association data according to the logical association context perception data to obtain logical association multidimensional perception data; Step S233: performing cluster analysis on the logically associated multidimensional perception data to obtain logically associated multidimensional cluster data; performing multidimensional category structure analysis on the logically associated multidimensional perception data according to the logically associated multidimensional cluster data to obtain logically associated multidimensional category structure data; Step S234: performing structural linear logic adjustment on the logically associated multidimensional perception data according to the logically associated multidimensional category structure data to obtain logical structural linear adjustment data; Step S235: performing cross-modal input content linear correlation analysis on the cross-modal input matching data according to the logical structure linear adjustment data to obtain cross-modal content linear correlation data; performing content correlation time interval exponential distribution calculation on the cross-modal content linear correlation data to obtain content correlation exponential distribution data; Step S236: Based on the content-related exponential distribution data, a content distribution Poisson process simulation is performed on the cross-modal content linear correlation data to obtain content distribution Poisson process data; based on the content distribution Poisson process data, human-computer interaction intention recognition is performed on the cross-modal content linear correlation data to obtain human-computer interaction intention data.

4. The human-computer interaction method based on intelligent data analysis according to claim 1, It is characterized in that The connection routing design for the cross-modal data connection channel includes: According to the connection bandwidth matching data and the structure sparsity distribution data, the structure demand capacity is evaluated to obtain the structure connection demand capacity data; Based on the structural connection demand capacity data, the structural sparsity distribution data is sorted for connection structure demand to obtain connection structure demand sorting data; according to the connection structure demand sorting data, the structural sparsity distribution data is sorted and extracted to obtain structure distribution rank data; Sorting and indexing the structure distribution ranking data based on the hash algorithm to obtain the structure data distribution index; Combining routing addresses according to the structural distribution ranking data and the structural data distribution index to obtain structural distribution routing address data; The connection routing is designed for the cross-modal data connection channel according to the structural distribution routing address data to obtain the cross-modal structural routing data.

5. The human-computer interaction method based on intelligent data analysis according to claim 1, It is characterized in that Step S33 includes the following steps: Step S331: performing high-concurrency traffic simulation on the cross-modal data interaction architecture to obtain high-concurrency traffic simulation data; Step S332: performing concurrent transmission channel configuration according to high concurrent traffic simulation data to obtain concurrent transmission channels; Step S333: Designing a buffer mechanism for high concurrent traffic simulation data to obtain a concurrent traffic buffer mechanism; Step S334: constructing a buffered transmission channel for the concurrent transmission channel according to the concurrent traffic buffering mechanism to obtain a concurrent buffered transmission channel; Step S335: Optimizing the concurrent buffer structure of the Apache Flink framework based on the concurrent buffer transmission channel to obtain a structurally optimized Apache Flink framework; Step S336: Design a parallel framework based on the structurally optimized Apache Flink framework to obtain an interactive parallel framework.

6. The human-computer interaction method based on intelligent data analysis according to claim 1, It is characterized in that The buffering mechanism design for high concurrent traffic simulation data includes: Perform buffer pool capacity evaluation on high concurrent traffic simulation data to obtain buffer pool capacity evaluation data; Memory space allocation is performed according to the buffer pool capacity assessment data to obtain memory space allocation data; Dynamically adjust temporary memory space for high concurrent traffic simulation data according to buffer pool capacity assessment data and memory space allocation data to obtain temporary memory space adjustment data; A buffer mechanism is designed for high concurrent traffic simulation data based on buffer pool capacity assessment data, memory space allocation data and temporary memory space adjustment data to obtain a concurrent traffic buffer mechanism.