Distributed proposition error correction and dynamic updating system and method based on edge AI
By using an edge AI distributed proposition correction system, which combines local and cloud-based collaborative correction methods, the problems of low efficiency and high latency in traditional proposition correction are solved, achieving efficient and accurate proposition correction and dynamic updates.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 网才科技(广州)集团股份有限公司
- Filing Date
- 2026-01-22
- Publication Date
- 2026-05-05
AI Technical Summary
Traditional error correction relies on manual review, which is inefficient, has a high error rate, and is not updated in a timely manner. In addition, the massive amount of data transmitted to the cloud results in excessive bandwidth consumption, significant latency, and difficulty in cross-regional collaboration.
It adopts a distributed proposition correction and dynamic update system based on edge AI. Through the collaborative work of the distributed edge node layer, edge AI processing layer, cloud collaboration layer and terminal interaction layer, it can achieve local preliminary error correction and cloud optimization, support offline and online collaborative modes, and reduce data transmission pressure and latency.
It significantly reduces data transmission pressure and processing latency, improves error correction accuracy, ensures that data is not lost when offline, enables multi-node data cross-validation and dynamic updates, and improves the applicability and reliability of the system.
Smart Images
Figure FT_1 
Figure FT_2 
Figure FT_3
Abstract
Description
Technical Field
[0001] This application relates to the field of edge computing technology, and in particular to a distributed proposition error correction and dynamic update system and method based on edge AI. Background Technology
[0002] In educational examinations, professional assessments, and corporate training, the quality of test questions directly determines the effectiveness and fairness of the assessment. Traditional test question correction relies on manual review of each question, which suffers from low efficiency, high error rate, untimely updates, and difficulties in cross-regional collaboration. Furthermore, transmitting massive amounts of test question data to the cloud for processing can easily lead to excessive bandwidth consumption and significant latency. Summary of the Invention
[0003] This application provides a distributed proposition error correction and dynamic update method and system based on edge AI, which significantly reduces data transmission pressure and processing latency, and solves the defects of high bandwidth dependence and high latency in existing technologies.
[0004] To achieve the above objectives, this application adopts the following technical solution: Firstly, a distributed proposition correction and dynamic update system based on edge AI is provided, including: The distributed edge node layer is equipped with multiple edge computing nodes, each of which is connected to the question collection terminal to receive the original question data. An edge AI processing layer, integrated into each edge computing node, is used to perform data processing and error correction analysis on the original proposition data to obtain structured proposition features and preliminary error correction results, and to store the original proposition data, the structured proposition features and the preliminary error correction results; The cloud collaboration layer communicates with the distributed edge node layer, aggregates the preliminary error correction results of each edge computing node, and optimizes the results to obtain optimized error correction results. The cloud collaboration layer is also used to generate dynamic update instructions for propositions based on the optimized error correction results. The terminal interaction layer is communicatively connected to the distributed edge node layer and the cloud collaboration layer, respectively, and is used to receive the optimized error correction results for manual verification, and to receive the dynamic update instruction for the proposition to synchronously update the local proposition data. The edge AI processing layer and the cloud collaboration layer support offline and online collaboration modes. When the edge computing node is offline, it independently performs propositional error correction and synchronizes data with the cloud collaboration layer after the network is restored.
[0005] Secondly, a distributed propositional error correction and dynamic update method based on edge AI is provided, including: S1: Receive the original question data uploaded by the question collection terminal through each edge computing node of the distributed edge node layer; S2: The original proposition data is processed by the edge AI processing layer integrated into each edge computing node to obtain structured proposition features, and then error correction analysis is performed based on the structured proposition features to generate preliminary error correction results; S3: The preliminary error correction results are uploaded to the cloud collaboration layer through each edge computing node, so that the cloud collaboration layer aggregates the preliminary error correction results of each edge computing node and optimizes them based on the preliminary error correction results to obtain the optimized error correction results; S4: The cloud collaboration layer generates a dynamic update instruction for the proposition based on the optimized error correction result, and pushes the dynamic update instruction and the optimized error correction result to the terminal interaction layer so that the terminal interaction layer receives the optimized error correction result for manual verification and receives the dynamic update instruction to synchronously update the local proposition data.
[0006] This system completes data processing and initial error correction locally at edge nodes, requiring only subsequent uploads of preliminary results. This significantly reduces data transmission pressure and processing latency, addressing the high bandwidth dependence and high latency shortcomings of existing technologies. By aggregating and optimizing preliminary results from multiple edge nodes in the cloud, it achieves cross-validation of data from multiple nodes, reducing the risk of misjudgment from a single node and significantly improving error correction accuracy. Dynamic update commands are generated in the cloud, ensuring precise synchronous updates at the terminal interaction layer, avoiding latency and redundancy in full data transmission. Layered data management, combining edge node temporary storage and cloud storage, ensures data is not lost when edge nodes are offline, while centralized management is achieved through unified cloud storage of correct data.
[0007] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description
[0008] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiments below. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a structural block diagram of the distributed proposition error correction and dynamic update system based on edge AI provided in the embodiments of this application; Figure 2This is a flowchart illustrating the distributed proposition error correction and dynamic update method based on edge AI provided in the embodiments of this application; Figure 3 A flowchart illustrating the workflow and inter-module interactions of a distributed proposition correction and dynamic update system based on edge AI; Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation
[0009] The embodiments of the technical solutions of this application will now be described in detail with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of this application, and are therefore merely examples and should not be used to limit the scope of protection of this application. When the following description relates to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. Various changes, modifications, and equivalents of the methods, apparatus, and / or systems described herein will become apparent upon understanding this disclosure. For example, the order of operations described herein is merely illustrative and is not limited to those orders set forth herein, but can be changed as will become apparent upon understanding this disclosure, except for operations that must be performed in a specific order. Furthermore, for clarity and conciseness, descriptions of features known in the art may be omitted.
[0010] The embodiments described in the following examples of this disclosure are not representative of all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.
[0011] See Figure 1 This is a schematic diagram of the structure of the distributed proposition error correction and dynamic update system based on edge AI provided in the first embodiment of this application. Figure 1 As shown, the distributed proposition error correction and dynamic update system 100 based on edge AI may include: The distributed edge node layer 110 is equipped with multiple edge computing nodes, and each edge computing node is connected to the question collection terminal to receive the original question data. The edge AI processing layer 120 is integrated into each edge computing node. It is used to perform data processing and error correction analysis on the original proposition data, obtain structured proposition features and preliminary error correction results, and store the original proposition data, structured proposition features and preliminary error correction results. The cloud collaboration layer 130 communicates and connects with the distributed edge node layer, aggregates the preliminary error correction results of each edge computing node, and optimizes them based on the preliminary error correction results to obtain optimized error correction results; the cloud collaboration layer is also used to generate dynamic update instructions for propositions based on the optimized error correction results. The terminal interaction layer 140 communicates with the distributed edge node layer and the cloud collaboration layer respectively, and is used to receive the optimized error correction results for manual verification, and to receive the dynamic update command of the question to synchronously update the local question data. Among them, the edge AI processing layer and the cloud collaboration layer support offline and online collaboration modes. When the edge computing node is offline, it independently performs propositional error correction and synchronizes data with the cloud collaboration layer after the network is restored.
[0012] The distributed edge node layer deploys multiple edge computing nodes, including one or more edge servers and edge gateway devices. Edge servers possess powerful computing capabilities, enabling them to handle complex computational tasks; edge gateway devices are primarily responsible for initial data processing and transmission, connecting different devices to the network. Furthermore, this layer supports low-bandwidth data transmission, ensuring stable data transmission even under limited network bandwidth conditions, guaranteeing normal system operation. Each edge computing node communicates with a question acquisition terminal. The question acquisition terminal is the source of raw question data, receiving it from various channels. This data includes one or more of the following: textual data (such as textual descriptions of question content), formulaic data (formulas from mathematics, physics, etc.), and graphical data (geometric figures, charts, etc.). The main function of the distributed edge node layer is to receive this raw question data and pass it to the next layer for processing.
[0013] The edge AI processing layer, integrated into various edge computing nodes, is the core component of the system for initial data processing and error correction. It performs format conversion and feature extraction operations on the received raw proposition data. Format conversion unifies proposition data in different formats into a standardized format that the system can process, such as converting text data from different document formats into a unified text encoding format. Feature extraction extracts representative and discriminative features from the proposition data, forming structured proposition features. These structured proposition features more accurately reflect the key information of the proposition, providing a foundation for subsequent error correction analysis. A pre-built AI error correction model detects proposition errors based on the structured proposition features and generates preliminary error correction results. The AI error correction model is an intelligent model trained on a large amount of data, capable of identifying potential grammatical errors, logical errors, and knowledge errors in the proposition. After generating the preliminary error correction results, the system temporarily stores the raw proposition data, structured proposition features, and preliminary error correction results in a local cache unit for subsequent querying and analysis. The edge AI processing layer and the cloud collaboration layer support offline and online collaborative modes. When an edge computing node is offline, it can independently perform propositional error correction tasks, processing data and models stored locally. After the network is restored, the edge computing node will synchronize data with the cloud collaboration layer to ensure the consistency and integrity of system data.
[0014] The cloud-based collaboration layer communicates with the distributed edge node layer, aggregating preliminary error correction results from various edge computing nodes. By collecting data from different edge nodes, the cloud-based collaboration layer obtains more comprehensive error correction information. Based on the aggregated preliminary error correction results, the cloud-based collaboration layer performs optimization processing to obtain optimized error correction results. The optimization process may comprehensively consider error correction suggestions from multiple edge nodes, historical error correction data, and a more comprehensive knowledge base to improve the accuracy and reliability of error correction. Furthermore, the cloud-based collaboration layer is also used to generate dynamic error correction update instructions based on the optimized error correction results, guiding terminals to update error correction data. The terminal interaction layer communicates with both the distributed edge node layer and the cloud-based collaboration layer, serving as a bridge between the system and the user. The terminal interaction layer includes an error correction review terminal and a user access terminal. The error correction review terminal is equipped with an error correction result annotation function, allowing reviewers to manually verify the optimized error correction results and synchronously feed the annotation data back to the cloud-based collaboration layer. Manual verification further ensures the accuracy of the error correction results, while the annotation data helps the system continuously optimize and improve. The user access terminal is mainly used by ordinary users to query and access updated proposition data, meeting users' needs.
[0015] Optionally, the edge AI processing layer is specifically used for: Perform format conversion and feature extraction on the original proposition data to obtain structured proposition features; The system uses a pre-built AI error correction model to detect errors in propositions based on structured proposition features and generates preliminary error correction results. The system also uses a local cache unit to temporarily store the original proposition data, structured proposition features, and preliminary error correction results. Specifically, the edge AI processing layer first performs format conversion and feature extraction operations on the original proposition data. Format conversion aims to unify proposition data from different formats into a standardized format that the system can process efficiently. For example, for text-based proposition data from different document formats (such as Word and PDF), the system converts it into a unified text encoding format to ensure data consistency and readability, facilitating subsequent processing and analysis. Feature extraction extracts representative and discriminative features from the format-converted data, forming structured proposition features. The extracted features differ depending on the type of proposition data. For text-based data, semantic features, such as keywords and themes, are mainly extracted, reflecting the core content and semantic information of the text. For formula-based data, logical features, such as the structure and operational relationships of the formula, are extracted to help accurately understand the meaning and logic of the formula. For graphical data, topological features, such as the shape and connectivity of the graph, are extracted to describe the spatial structure and topological properties of the graph. These structured proposition features provide strong support for subsequent error correction analysis.
[0016] Secondly, a pre-built AI error correction model is used to detect errors in structured propositions based on their characteristics and generate preliminary error correction results. This AI error correction model is trained on a large amount of proposition data and error correction rules, exhibiting high accuracy and intelligence. It can perform in-depth analysis of the characteristics of structured propositions, identify potential errors, and generate corresponding preliminary error correction results according to pre-defined rules.
[0017] Finally, the original proposition data, structured proposition features, and preliminary error correction results are temporarily stored in a local cache unit. This local cache unit, serving as a temporary storage location, enables rapid data storage and retrieval, improving system processing efficiency. Simultaneously, it facilitates subsequent data querying, analysis, and backtracking, ensuring timely access to relevant data information when needed.
[0018] Optionally, the system also includes a security protection module, deployed on edge computing nodes and the cloud collaboration layer, respectively, for encrypting data transmission, verifying integrity, and authenticating access identity for the original question data. Specifically, the security protection modules are deployed on edge computing nodes and the cloud collaboration layer, providing comprehensive protection for the system's data security. The original data is encrypted during transmission to prevent theft or tampering. Advanced encryption algorithms are employed to ensure data security and confidentiality. Integrity checks are performed on transmitted data to detect loss or damage during transmission. If incomplete data is detected, the system will promptly request retransmission to ensure accuracy and integrity. Devices and users accessing the system are authenticated; only authorized devices and users can access system resources. This strict authentication mechanism prevents unauthorized access and ensures system security.
[0019] Optionally, the terminal interaction layer includes a question-setting review terminal and a user access terminal. The question-setting review terminal is configured with an error correction result annotation function, and the question-setting review terminal is used to synchronously feed the annotation data back to the cloud collaboration layer.
[0020] Optionally, the original data for the proposition may include one or more of the following: text data, formula data, and graph data.
[0021] Optionally, the edge computing nodes in the distributed edge node layer include one or more of edge servers and edge gateway devices, and support low-bandwidth data transmission.
[0022] The terminal interaction layer is an important link in the interaction between the distributed proposition correction and dynamic update system based on edge AI and the user. It communicates and connects with the distributed edge node layer and the cloud collaboration layer, and mainly includes the proposition review terminal and the user access terminal, each terminal having different functions.
[0023] The exam question review terminal is a professional device for reviewers, equipped with error correction result annotation capabilities. After receiving the optimized error correction results pushed by the cloud-based collaboration layer, reviewers can manually verify them. During verification, if inaccuracies or incompleteness are found, reviewers can use the error correction result annotation function to detail the location and type of the error, as well as the correct corrections. After annotation, the exam question review terminal synchronously feeds the annotated data back to the cloud-based collaboration layer. Based on this feedback data, the cloud-based collaboration layer further optimizes the error correction model and algorithm, improving the accuracy and reliability of subsequent exam question corrections. This combination of human and AI effectively compensates for the shortcomings of relying solely on AI for error correction, ensuring the quality of exam question correction.
[0024] The user access terminal is geared towards ordinary users, providing them with a convenient way to query and access updated exam question data. Users can log in to the system through the user access terminal and, according to their needs and permissions, query relevant exam question information, including the question content, answers, explanations, and error corrections. Simultaneously, the user access terminal also supports user feedback and evaluation of the exam question data. Based on user feedback, the system can assess and improve the quality and applicability of the exam questions, continuously enhancing the user experience.
[0025] In this distributed proposition correction and dynamic update system based on edge AI, the original proposition data has diverse types, mainly including one or more of text data, formula data, and graphical data.
[0026] Text-based data is one of the most common forms of question designation, encompassing various text-based questions, such as reading comprehension and essay questions in Chinese, and grammar and cloze tests in English. This text-based data contains rich semantic information, expressing the content, requirements, and conditions of the question through text.
[0027] Formula-based data primarily appears in disciplines such as mathematics, physics, and chemistry, including various mathematical formulas, physical laws, and chemical equations. Formula-based data possesses a rigorous logical structure and operational relationships, enabling it to accurately describe the quantitative relationships and patterns of change within these disciplines.
[0028] Graphical data includes geometric figures, statistical charts, and experimental setup diagrams. Geometric figures are used to describe spatial shapes and positional relationships, statistical charts are used to display the distribution and trends of data, and experimental setup diagrams are used to illustrate experimental equipment and operating procedures. Graphical data presents propositional information in an intuitive visual form, helping users to better understand and analyze the propositions.
[0029] Different types of original data for propositions have their own characteristics and processing requirements. This system, through the collaborative work of the edge AI processing layer and the cloud collaboration layer, can effectively process and correct these diverse data, ensuring the accuracy and quality of the propositions.
[0030] The distributed edge node layer is a crucial component of this edge AI-based distributed proposition correction and dynamic update system. The edge computing nodes within this layer include one or more edge servers and edge gateway devices, and are characterized by their ability to support low-bandwidth data transmission. Edge servers, as powerful edge computing devices, possess high computing power and storage capacity. They can independently handle relatively complex proposition data processing and correction tasks, rapidly processing and analyzing the received raw proposition data. Edge servers are typically deployed close to the proposition acquisition terminals, enabling timely responses to data upload requests and reducing data transmission latency. Edge gateway devices primarily function as data aggregators and forwarders. They can connect multiple proposition acquisition terminals, collecting and organizing the raw proposition data uploaded by the terminals, and then forwarding it to edge servers or other edge computing nodes for processing. Edge gateway devices have low cost and small size, making them easy to deploy in various environments.
[0031] The edge computing nodes of this system support low-bandwidth data transmission, adapting to data transmission needs in various network environments. In remote areas or locations with poor network coverage, network bandwidth may be low, limiting traditional data transmission methods. However, the edge computing nodes of this system, by employing optimized data compression algorithms and transmission protocols, can achieve efficient data transmission under low-bandwidth conditions. For example, the original propositional data is compressed to reduce the data volume before transmission; reliable transmission protocols are used to ensure the integrity and accuracy of data during transmission. This characteristic enables the system to operate stably in a wider range of network environments, improving the system's applicability and reliability.
[0032] This system completes data processing and initial error correction locally at edge nodes, requiring only subsequent uploads of preliminary results. This significantly reduces data transmission pressure and processing latency, addressing the high bandwidth dependence and high latency shortcomings of existing technologies. By aggregating and optimizing preliminary results from multiple edge nodes in the cloud, it achieves cross-validation of data from multiple nodes, reducing the risk of misjudgment from a single node and significantly improving error correction accuracy. Dynamic update commands are generated in the cloud, ensuring precise synchronous updates at the terminal interaction layer, avoiding latency and redundancy in full data transmission. Layered data management, combining edge node temporary storage and cloud storage, ensures data is not lost when edge nodes are offline, while centralized management is achieved through unified cloud storage of correct data.
[0033] like Figure 2 As shown, Figure 2This is a flowchart illustrating a distributed propositional error correction and dynamic update method based on edge AI. The method includes: S1: Receive the original question data uploaded by the question collection terminal through each edge computing node of the distributed edge node layer; S2: The original proposition data is processed by the edge AI processing layer integrated into each edge computing node to obtain structured proposition features, and then error correction analysis is performed based on the structured proposition features to generate preliminary error correction results; S3: The preliminary error correction results are uploaded to the cloud collaboration layer through each edge computing node, so that the cloud collaboration layer aggregates the preliminary error correction results of each edge computing node and optimizes them based on the preliminary error correction results to obtain the optimized error correction results; S4: The cloud collaboration layer generates a dynamic update instruction for the proposition based on the optimized error correction result, and pushes the dynamic update instruction and the optimized error correction result to the terminal interaction layer so that the terminal interaction layer receives the optimized error correction result for manual verification and receives the dynamic update instruction to synchronously update the local proposition data. Optionally, the process by which the edge AI processing layer performs data processing on the original proposition data includes: converting the original proposition data into structured data, and extracting the semantic features, logical features, or topological features corresponding to the structured data. Optionally, the dynamic update instruction for the proposition includes the proposition identifier, error location information, correction content, and update time information, and each edge computing node synchronizes the proposition data through incremental update. Optionally, the edge computing nodes of the distributed edge node layer include one or more of edge servers and edge gateway devices, and support low-bandwidth data transmission.
[0034] The question collection terminal, as the source of question data, can collect raw question data from different channels and in various forms, including text data, formula data, and graphical data. Edge computing nodes, distributed at the edge of the network, can receive this data in a timely manner, preparing it for subsequent processing.
[0035] The edge AI processing layer first converts the format of the original proposition data, unifying data from different formats into a standardized format that the system can process. Then, it extracts features from the data to form structured proposition features, which accurately describe the key information of the proposition. Next, using a pre-built AI error correction model, it detects errors in the proposition based on the structured proposition features and generates preliminary error correction results.
[0036] As the central coordination and optimization center of the system, the cloud-based collaboration layer receives preliminary error correction results from multiple edge computing nodes and performs comprehensive analysis and comparison of these results. By employing more complex algorithms and models, the preliminary error correction results are optimized to improve the accuracy and reliability of error correction.
[0037] The terminal interaction layer includes a question review terminal and a user access terminal. The question review terminal allows reviewers to manually verify the optimized error correction results to ensure the accuracy of the corrections. The user access terminal, based on the question dynamic update instructions, synchronously updates the local question data, enabling users to obtain the latest and most accurate question information in a timely manner.
[0038] In step S2, the edge AI processing layer performs data processing on the original proposition data, which is a key part of the entire method and mainly includes the following steps: First, the raw proposition data is converted into structured data. Since the raw proposition data may come from different sources and have different formats and structures, it needs to be standardized. For example, for text-based proposition data, redundant information and special symbols are removed, and it is converted into a uniform text format; for formula-based proposition data, it is converted into a standard mathematical expression format; for graphical proposition data, key information such as shape and coordinates is extracted and converted into structured graphical data. This conversion ensures that the data has a consistent format and structure, facilitating subsequent feature extraction and analysis.
[0039] Then, semantic, logical, or topological features are extracted from the structured data. For text-based structured data, semantic features are primarily extracted, such as keywords, themes, and semantic relationships. These semantic features reflect the core content and semantic information of the text, helping to understand the intent and requirements of the proposition. For formula-based structured data, logical features are extracted, such as the structure of the formula, operational relationships, and logical connections between variables. These logical features help to accurately understand the meaning and logic of the formula and detect potential errors. For graphical structured data, topological features are extracted, such as the shape, connectivity, and spatial layout of the graph. These topological features describe the spatial structure and topological properties of the graph, which is crucial for judging the correctness and rationality of the graph.
[0040] Through the above data processing, the edge AI processing layer can transform the original proposition data into structured data with clear characteristics, providing an accurate and reliable foundation for subsequent error correction analysis.
[0041] In step S4, the proposition dynamic update instruction generated by the cloud collaboration layer based on the optimized error correction results contains several key pieces of information. Each edge computing node synchronizes the proposition data through incremental updates. The specific content is as follows: The dynamic update command for a proposition includes a proposition identifier, error location information, correction details, and update time information. The proposition identifier uniquely identifies the proposition requiring an update, ensuring the accuracy and relevance of the update operation. Error location information precisely pinpoints the exact location of the error within the proposition, such as a paragraph or sentence in text, a variable or operator in a formula, or a specific part of a graph. Correction details provide the correct solution to restore the proposition to its accurate state. Update time information records the date the proposition was updated, facilitating the tracking of proposition change history and the management of proposition versions.
[0042] Each edge computing node synchronizes the test data using incremental updates. Incremental updates refer to transmitting only the changed parts of the test data, rather than transmitting the entire test data. Compared to full updates, incremental updates significantly reduce data transmission volume and improve data synchronization efficiency. When the cloud collaboration layer generates a dynamic update instruction for the test data, it sends the data to be updated and the corresponding update instructions to each edge computing node. Upon receiving this information, the edge computing nodes perform partial updates to the locally stored test data according to the instructions, modifying only the changed parts while keeping other unchanged data unchanged. This incremental update method can minimize data transmission overhead while ensuring data consistency, thereby improving the overall performance and response speed of the system.
[0043] In this distributed proposition correction and dynamic update method based on edge AI, the edge computing nodes in the distributed edge node layer include one or more of edge servers and edge gateway devices, and have the characteristic of supporting low-bandwidth data transmission, as detailed below: Edge servers, as powerful edge computing devices, possess high computing power and storage capacity. They can independently handle relatively complex proposition data processing and error correction tasks, rapidly processing and analyzing the received raw proposition data. During method execution, edge servers can run relevant algorithms and models of the edge AI processing layer, performing operations such as format conversion, feature extraction, and error correction analysis on the proposition data. Simultaneously, edge servers can also store large amounts of proposition data and processing results, supporting subsequent data querying and analysis.
[0044] Edge gateway devices primarily function as data aggregation and forwarding systems. They connect multiple data acquisition terminals, collect and organize the raw data uploaded by these terminals, and then forward it to edge servers or other edge computing nodes for processing. Edge gateway devices are cost-effective and compact, making them easy to deploy in various environments. In this method, edge gateway devices ensure smooth data transmission between data acquisition terminals and edge servers, improving data collection efficiency.
[0045] This method completes data processing and initial error correction locally at edge nodes, requiring only subsequent uploading of preliminary results. This significantly reduces data transmission pressure and processing latency, addressing the shortcomings of existing technologies such as high bandwidth dependence and high latency. By aggregating and optimizing preliminary results from multiple edge nodes in the cloud, cross-validation of multi-node data is achieved, reducing the risk of misjudgment from a single node and significantly improving error correction accuracy. Dynamic update commands are generated in the cloud, and the terminal interaction layer accurately synchronizes updates, avoiding latency and redundancy in full data transmission. Through layered data management of edge node temporary storage and cloud storage, data is not lost when edge nodes are offline, while centralized management is achieved through unified cloud storage of correct proposition data.
[0046] Figure 3 The workflow and inter-module interactions of a distributed proposition correction and dynamic update system based on edge AI are demonstrated. The functions and data flow logic of each module are as follows: Question collection terminal: As the data entry point, it is responsible for collecting the original question data (text, formulas, graphs, etc.) and transmitting it to the distributed edge node layer.
[0047] Distributed edge node layer: Deploy multiple edge computing nodes (such as edge servers and edge gateways) to receive raw data from the question collection terminal on the one hand, and provide a running platform for the edge AI processing layer on the other hand. At the same time, it supports offline / online data synchronization with the cloud collaboration layer when the network environment changes.
[0048] Edge AI processing layer: Integrated into edge computing nodes, it performs format conversion and feature extraction on the original data of the proposition, and generates preliminary error correction results through a pre-built AI error correction model. At the same time, it temporarily stores the original data of the proposition, structured features and preliminary results in a local cache unit.
[0049] Cloud Collaboration Layer: Aggregates the preliminary error correction results from various edge nodes, performs multi-source verification and optimization, and generates optimized error correction results; at the same time, it generates dynamic update instructions for propositions based on the results, and supports receiving manually labeled data from the terminal interaction layer to continuously iterate the model.
[0050] Terminal interaction layer: It is divided into question review terminal and user access terminal. On the one hand, it receives the optimization and error correction results from the cloud for manual verification and feeds back the labeled data to the cloud; on the other hand, it receives the question dynamic update instructions and updates the local question data synchronously.
[0051] Security Protection Module: It runs through the "distributed edge node layer" and the "cloud collaboration layer", providing end-to-end security protection such as data transmission encryption, integrity verification, and access identity authentication to prevent data leakage, tampering, or unauthorized access.
[0052] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of this application. For example... Figure 4 As shown, the electronic device 5 of this embodiment includes: at least one processor 50 ( Figure 4 (Only one is shown in the diagram) a processor, a memory 51, and a computer program 52 stored in the memory 51 and executable on the at least one processor 50, wherein the processor 50 executes the computer program 52 to implement the steps in any of the above embodiments of the distributed propositional error correction and dynamic update method based on edge AI.
[0053] The electronic device may be a desktop computer, laptop, handheld computer, or cloud server, etc. This electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that... Figure 4 This is merely an example of electronic device 5 and does not constitute a limitation on electronic device 5. It may include more or fewer components than shown in the figure, or combine certain components, or different components. For example, it may also include input / output devices, network access devices, etc.
[0054] The processor 50 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.
[0055] In some embodiments, the memory 51 may be an internal storage unit of the electronic device 5, such as a hard disk or memory of the electronic device 5. In other embodiments, the memory 51 may be an external storage device of the electronic device 5, such as a plug-in hard disk, SmartMediaCard (SMC), SecureDigital (SD) card, or FlashCard equipped on the electronic device 5. Furthermore, the memory 51 may include both internal and external storage units of the electronic device 5. The memory 51 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 51 can also be used to temporarily store data that has been output or will be output.
[0056] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps in the above-described method embodiments.
[0057] This application provides a computer program product that, when run on an electronic device, enables the electronic device to implement the steps described in the various method embodiments above.
[0058] If the integrated unit is implemented as a software functional unit and used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a device / electronic device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks.
[0059] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0060] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0061] In the embodiments provided in this application, it should be understood that the disclosed devices / electronic devices and methods can be implemented in other ways. For example, the device / electronic device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings or direct couplings or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.
[0062] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0063] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.
[0064] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.
[0065] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0066] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.
[0067] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).
[0068] In the description of the embodiments of this application, the technical terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of this application and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of this application.
[0069] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.
[0070] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.
Claims
1. A distributed proposition error correction and dynamic update system based on edge AI, characterized in that, include: The distributed edge node layer is equipped with multiple edge computing nodes, each of which is connected to the question collection terminal to receive the original question data. An edge AI processing layer, integrated into each edge computing node, is used to perform data processing and error correction analysis on the original proposition data to obtain structured proposition features and preliminary error correction results, and to store the original proposition data, the structured proposition features and the preliminary error correction results; The cloud collaboration layer communicates with the distributed edge node layer, aggregates the preliminary error correction results of each edge computing node, and optimizes the results based on the preliminary error correction results to obtain the optimized error correction results. The cloud-based collaboration layer is also used to generate dynamic update instructions for propositions based on the optimized error correction results; The terminal interaction layer is communicatively connected to the distributed edge node layer and the cloud collaboration layer, respectively, and is used to receive the optimized error correction results for manual verification, and to receive the dynamic update instruction for the proposition to synchronously update the local proposition data. The edge AI processing layer and the cloud collaboration layer support offline and online collaboration modes. When the edge computing node is offline, it independently performs propositional error correction and synchronizes data with the cloud collaboration layer after the network is restored.
2. The system according to claim 1, characterized in that, The edge AI processing layer is specifically used for: Perform format conversion and feature extraction on the original proposition data to obtain the structured proposition features; Using a pre-built AI error correction model, proposition errors are detected based on the structured proposition features, and preliminary error correction results are generated. The original proposition data, the structured proposition features, and the preliminary error correction results are temporarily stored in a local cache unit.
3. The system according to claim 1, characterized in that, It also includes a security protection module, which is deployed on the edge computing node and the cloud collaboration layer respectively, for encrypting data transmission, verifying integrity, and authenticating access identity for the original data of the proposition.
4. The system according to claim 1, characterized in that, The terminal interaction layer includes a question-setting review terminal and a user access terminal. The question-setting review terminal is configured with an error correction result annotation function, and the question-setting review terminal is used to synchronously feed back the annotation data to the cloud collaboration layer.
5. The system according to claim 1, characterized in that, The original data for the proposition includes one or more of the following: text data, formula data, and graphical data.
6. The system according to claim 1, characterized in that, The edge computing nodes of the distributed edge node layer include one or more of edge servers and edge gateway devices, and support low-bandwidth data transmission.
7. A distributed proposition error correction and dynamic update method based on edge AI, characterized in that, Includes the following steps: S1: Receive the original question data uploaded by the question collection terminal through each edge computing node of the distributed edge node layer; S2: The original proposition data is processed by the edge AI processing layer integrated into each edge computing node to obtain structured proposition features, and then error correction analysis is performed based on the structured proposition features to generate preliminary error correction results; S3: The preliminary error correction results are uploaded to the cloud collaboration layer through each edge computing node, so that the cloud collaboration layer aggregates the preliminary error correction results of each edge computing node and optimizes them based on the preliminary error correction results to obtain the optimized error correction results; S4: The cloud collaboration layer generates a dynamic update instruction for the proposition based on the optimized error correction result, and pushes the dynamic update instruction and the optimized error correction result to the terminal interaction layer so that the terminal interaction layer receives the optimized error correction result for manual verification and receives the dynamic update instruction to synchronously update the local proposition data.
8. The method according to claim 7, characterized in that, The process of the edge AI processing layer performing data processing on the original proposition data in step S2 includes: converting the original proposition data into structured data, and extracting the semantic features, logical features or topological features corresponding to the structured data.
9. The method according to claim 7, characterized in that, The dynamic update instruction for the proposition in step S4 includes the proposition identifier, error location information, correction content, and update time information. Each edge computing node synchronizes the proposition data through incremental update.
10. The method according to claim 7, characterized in that, The edge computing nodes of the distributed edge node layer include one or more of edge servers and edge gateway devices, and support low-bandwidth data transmission.