CN-121981111-A - Distributed proposition error correction and dynamic updating system and method based on edge AI

CN121981111ACN 121981111 ACN121981111 ACN 121981111ACN-121981111-A

Abstract

The application relates to the technical field of edge computing, in particular to a distributed proposition error correction and dynamic updating system and method based on an edge AI, comprising a distributed edge node layer, a distributed proposition processing system and a distributed proposition updating system, wherein a plurality of edge computing nodes are deployed on the distributed edge node layer, and each edge computing node is respectively in communication connection with a proposition acquisition terminal and is used for receiving proposition original data; the system comprises an edge AI processing layer, a cloud coordination layer, a terminal interaction layer and a local proposition data synchronization layer, wherein the edge AI processing layer is used for performing data processing and error correction analysis on proposition original data to obtain structural proposition features and primary error correction results, the cloud coordination layer is used for aggregating the primary error correction results of all edge computing nodes and optimizing the primary error correction results to obtain optimized error correction results, the cloud coordination layer is also used for generating proposition dynamic update instructions based on the optimized error correction results, and the terminal interaction layer is used for receiving the optimized error correction results to perform manual verification and synchronously update the local proposition data. The data transmission pressure and the processing delay are obviously reduced, and the defects of high bandwidth dependence and high delay in the prior art are overcome.

Inventors

MA HE
NI XIAOMING
Guo nanming
DU YULIN
Hong Qiankai
Cui Haosong
JIANG YIXUAN

Assignees

网才科技(广州)集团股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260122

Claims (10)

1. An edge AI-based distributed proposition error correction and dynamic updating system, comprising: the distributed edge node layer is provided with a plurality of edge computing nodes, and each edge computing node is respectively in communication connection with the proposition acquisition terminal and is used for receiving proposition original data; The edge AI processing layer is integrated with each edge computing node and is used for performing data processing and error correction analysis on the proposition original data to obtain a structured proposition characteristic and a preliminary error correction result, and storing the proposition original data, the structured proposition characteristic and the preliminary error correction result; The cloud coordination layer is in communication connection with the distributed edge node layer, aggregates the preliminary error correction results of the edge computing nodes, and optimizes the preliminary error correction results to obtain optimized error correction results; the terminal interaction layer is respectively in communication connection with the distributed edge node layer and the cloud coordination layer and is used for receiving the optimized error correction result for manual verification and receiving the proposition dynamic update instruction to synchronously update the local proposition data; The edge AI processing layer and the cloud cooperation layer support an offline online cooperation mode, and the edge computing nodes independently execute proposition error correction when offline and synchronize data with the cloud cooperation layer after network recovery.
2. The system of claim 1, wherein the edge AI processing layer is specifically configured to: performing format conversion and feature extraction on the proposition raw data to obtain the structural proposition features; Detecting proposition errors according to the structural proposition features through a preset AI error correction model, generating the preliminary error correction result, and temporarily storing the proposition original data, the structural proposition features and the preliminary error correction result through a local buffer unit.
3. The system of claim 1, further comprising a security protection module disposed at the edge computing node and the cloud coordination layer, respectively, for performing data transmission encryption, integrity verification, and access identity authentication on the proposition raw data.
4. The system of claim 1, wherein the terminal interaction layer comprises a proposition auditing terminal and a user access terminal, the proposition auditing terminal is configured with an error correction result labeling function, and the proposition auditing terminal is used for synchronously feeding back labeling data to the cloud coordination layer.
5. The system of claim 1, wherein the proposition raw data comprises one or more of text-like data, formula-like data, and graphic-like data.
6. The system of claim 1, wherein the edge computing nodes of the distributed edge node layer comprise one or more of edge servers, edge gateway devices, and support low bandwidth data transmission.
7. The distributed proposition error correction and dynamic updating method based on the edge AI is characterized by comprising the following steps: s1, receiving proposition original data uploaded by a proposition acquisition terminal through each edge computing node of a distributed edge node layer; S2, performing data processing on the proposition original data through an edge AI processing layer integrated in each edge computing node to obtain structural proposition characteristics, and performing error correction analysis based on the structural proposition characteristics to generate a primary error correction result; S3, uploading the preliminary error correction result to a cloud cooperative layer through each edge computing node so that the cloud cooperative layer aggregates the preliminary error correction result of each edge computing node, and optimizing the preliminary error correction result to obtain an optimized error correction result; S4, generating a proposition dynamic update instruction based on the optimized error correction result through the cloud cooperative layer, pushing the proposition dynamic update instruction and the optimized error correction result to a terminal interaction layer, enabling the terminal interaction layer to receive the optimized error correction result for manual verification, and receiving the proposition dynamic update instruction to synchronously update local proposition data.
8. The method of claim 7, wherein the step S2 of the edge AI processing layer performing data processing on the proposition raw data comprises converting the proposition raw data into structured data and extracting semantic, logical or topological features corresponding to the structured data.
9. The method of claim 7, wherein the proposition dynamic update command in step S4 includes proposition identification, error location information, correction content and update time information, and the edge computing nodes synchronize proposition data by incremental update.
10. The method of claim 7, wherein the edge computing nodes of the distributed edge node layer comprise one or more of edge servers, edge gateway devices, and support low bandwidth data transmission.

Description

Distributed proposition error correction and dynamic updating system and method based on edge AI Technical Field The application relates to the technical field of edge computing, in particular to a distributed proposition error correction and dynamic updating system and method based on an edge AI. Background In propositional scenes such as educational examination, professional evaluation, enterprise training and the like, the propositional quality directly determines the evaluation effectiveness and fairness. The traditional proposition error correction relies on manual question-by-question auditing, and has the problems of low efficiency, high error leakage rate, untimely updating, difficult inter-regional collaboration and the like, and the problems of excessive bandwidth occupation, obvious delay and the like easily caused by massive proposition data transmission to a cloud processing. Disclosure of Invention The application provides a distributed proposition error correction and dynamic updating method and system based on an edge AI, which remarkably reduce data transmission pressure and processing delay and solve the defects of high bandwidth dependence and high delay in the prior art. In order to achieve the above purpose, the application adopts the following technical scheme: in a first aspect, an edge AI-based distributed proposition error correction and dynamic update system is provided, including: the distributed edge node layer is provided with a plurality of edge computing nodes, and each edge computing node is respectively in communication connection with the proposition acquisition terminal and is used for receiving proposition original data; The edge AI processing layer is integrated with each edge computing node and is used for performing data processing and error correction analysis on the proposition original data to obtain a structured proposition characteristic and a preliminary error correction result, and storing the proposition original data, the structured proposition characteristic and the preliminary error correction result; The cloud coordination layer is in communication connection with the distributed edge node layer, aggregates the preliminary error correction results of the edge computing nodes, and optimizes the preliminary error correction results to obtain optimized error correction results; the terminal interaction layer is respectively in communication connection with the distributed edge node layer and the cloud coordination layer and is used for receiving the optimized error correction result for manual verification and receiving the proposition dynamic update instruction to synchronously update the local proposition data; The edge AI processing layer and the cloud cooperation layer support an offline online cooperation mode, and the edge computing nodes independently execute proposition error correction when offline and synchronize data with the cloud cooperation layer after network recovery. In a second aspect, a distributed proposition error correction and dynamic update method based on an edge AI is provided, including: s1, receiving proposition original data uploaded by a proposition acquisition terminal through each edge computing node of a distributed edge node layer; S2, performing data processing on the proposition original data through an edge AI processing layer integrated in each edge computing node to obtain structural proposition characteristics, and performing error correction analysis based on the structural proposition characteristics to generate a primary error correction result; S3, uploading the preliminary error correction result to a cloud cooperative layer through each edge computing node so that the cloud cooperative layer aggregates the preliminary error correction result of each edge computing node, and optimizing the preliminary error correction result to obtain an optimized error correction result; S4, generating a proposition dynamic update instruction based on the optimized error correction result through the cloud cooperative layer, pushing the proposition dynamic update instruction and the optimized error correction result to a terminal interaction layer, enabling the terminal interaction layer to receive the optimized error correction result for manual verification, and receiving the proposition dynamic update instruction to synchronously update local proposition data. The system completes data processing and preliminary error correction locally through the edge node, only needs to upload the preliminary result subsequently, obviously reduces data transmission pressure and processing delay, and solves the defects of high bandwidth dependence and high delay in the prior art. And the multi-node data cross verification is realized by aggregating and optimizing the preliminary results of the multi-edge nodes by the cloud, so that the misjudgment risk of a single node is reduced, and the error correction precision is remarkably improved. The