CN-122027196-A - Multi-terminal adaptive social co-treatment information autonomous reporting system based on micro-service architecture

CN122027196ACN 122027196 ACN122027196 ACN 122027196ACN-122027196-A

Abstract

The multi-terminal adaptive social co-treatment information autonomous reporting system based on the micro-service architecture comprises a multi-terminal data acquisition and transmission module, a data preprocessing and hash generation module, a federal learning module, a federal gradient aggregation module, a data preprocessing and hash generation module, a data security enhancement module and a federal learning module, wherein the precision of a device twin model is called as an IoT device data weight, the historical reliability of a query user is used as an artificial data weight, the fusion confidence is calculated through a formula, the artificial data reliability and the IoT device precision are quantized, the problem of low data acquisition reliability is solved, the data preprocessing and hash generation module is used for accurately identifying abnormal data by extracting data to be checked and combining the number of target check rules and the evidence storage triggering weight through the formula, the abnormal data is accurately identified, the problem of inaccurate abnormal detection is solved, the federal learning module is used for obtaining a global gradient through the federal learning gradient aggregation formula, the training of a cross-domain model can be completed without data leaving a domain, and the data security is improved.

Inventors

LI GUANGYU
CUI JUAN
Yong Qijiang
SHI HUIYAN
LIU YUE
LIU SHUAI
ZHANG WEI
LIU LEI
GUO XIAOLEI
GU XIAOPENG
LIANG CHAO
WANG LEI

Assignees

中国电子科技集团有限公司电子科学研究院
合肥市公安局

Dates

Publication Date: 20260512
Application Date: 20251224

Claims (10)

1. The multi-terminal adaptive social co-treatment information autonomous reporting system based on the micro-service architecture is characterized by comprising the following components: the equipment twinning binding module is used for accessing the IoT equipment, generating an equipment twinning model and a data mapping rule, configuring a blockchain certificate triggering condition and generating a governance template; The multi-terminal data acquisition and transmission module is used for automatically acquiring treatment data according to the preset frequency of the equipment twin model and transmitting the treatment data to a system background through an encryption channel; the data preprocessing and hash generation module is used for preprocessing the treatment data, carrying out validity verification according to the intelligent verification rule of the treatment template, pushing the legal treatment data to the blockchain evidence storage micro-service according to the evidence storage triggering condition, and generating an evidence storage hash; The federal learning module is used for loading the local binding certificate hash treatment data, screening similar treatment data based on the industry label of the treatment template, initializing the local treatment model, carrying out gradient encryption transmission on the federal learning, polymerizing to generate a global treatment model, carrying out risk assessment on the treatment data, and generating a hierarchical early warning; and the data reflux module is used for calibrating the equipment twin model according to the hierarchical early warning, re-reporting the treatment data and triggering the secondary evidence storage to generate a new evidence storage hash, wherein the data is refluxed to the federal learning node, and the federal learning gradient calculation weight is optimized by combining the unit credit, so as to update the global treatment model.
2. The multi-terminal adaptive social co-therapy information autonomous reporting system based on the micro-service architecture as set forth in claim 1, wherein the device twinning binding module comprises: the form component configuration submodule is used for a social unit to drag the basic component through the low-code platform and set intelligent verification rules in combination with industry requirements; the device binding submodule is used for calling a compatibility detection interface of the access micro-service of the IoT device and verifying a device communication protocol; The twin-building mold module is used for generating a twin-device model based on the device data field and configuring the device data-form field data mapping rule; The certification rule configuration submodule is used for visually configuring the triggering condition of the blockchain certification and associating the blockchain certification micro-service interface; And the template generation submodule is used for integrating the intelligent verification rule and the equipment twin model to generate a unit exclusive governance template.
3. The system of claim 2, wherein the multi-terminal data acquisition and transmission module comprises: The multi-terminal data acquisition sub-module is used for the social units to fill in the manual data according to the management templates through the multi-terminal inlets, and associate the template ID and the unit ID; the automatic acquisition and data fusion sub-module is used for acquiring data according to the preset frequency of the twin model of the IoT device, transmitting the data to the system through an encryption protocol, classifying and fusing the artificial data and the IoT data according to the management template field, and generating unified management data.
4. The multi-terminal adaptive social co-therapy information autonomous reporting system based on the micro-service architecture of claim 3, wherein the data preprocessing and hash generation module comprises: the data cleaning and standardization sub-module is used for performing duplication removal, filtering and OCR (optical character recognition) on the treatment data and uniformly converting the treatment data into a JSON format; the intelligent rule checking sub-module is used for calling an intelligent checking rule preset in the management template, performing validity check on the management data and marking abnormal data fields and sources; and the block chain certificate and hash generation sub-module is used for generating a certificate hash and a block height by pushing the legal treatment data set to the block chain certificate micro-service through the RabbitMQ message queue according to the certificate triggering condition and starting PBFT a consensus mechanism.
5. The multi-terminal adaptive social co-therapy information autonomous reporting system based on the micro-service architecture as set forth in claim 4, wherein the federal learning module comprises: The regional data screening and model initializing sub-module is used for loading legal treatment data of the local binding certificate hash, screening classified treatment data based on an industry label of the treatment template, calling a federal learning node interface and initializing a local treatment model; The gradient encryption transmission and model aggregation sub-module is used for calculating a local treatment model federal learning gradient based on industry classification treatment data and encrypting the local treatment model federal learning gradient through a homomorphic encryption algorithm, wherein the encrypted gradient is transmitted to an aggregation node through a micro-service cluster security channel, and the multi-region gradient aggregation is completed by adopting FedAvg algorithm to generate a global treatment model; the risk assessment and early warning pushing sub-module is used for performing risk assessment on newly reported binding certificate hash data by the global treatment model, generating hierarchical early warning, associating abnormal data fields with correction suggestions, and pushing the abnormal data fields to a social unit through the multi-terminal access layer.
6. The multi-terminal adaptive social co-therapy information autonomous reporting system based on the micro-service architecture as set forth in claim 5, wherein the data reflow module comprises: The secondary evidence storage submodule is used for receiving early warning notice by a social unit through a multi-port, correcting abnormal data, correcting the acquisition rule of a twin model of equipment, reporting the corrected treatment data again, triggering a secondary evidence storage flow and generating a new evidence storage hash; the rectifying and modifying data reflux sub-module is used for refluxing the rectified and modified treatment data to the federal learning nodes of each region through the RabbitMQ message queue and synchronously updating the credit score of the unit; And the authority optimization sub-module is used for adjusting the federal learning gradient calculation weight based on the returned treatment data by the federal learning node, iterating the global treatment model and configuring authority hooks for the unit credit and the treatment template.
7. The multi-terminal adaptive social co-treatment information autonomous reporting system based on the micro-service architecture according to claim 2, wherein in the equipment twinning binding module, a social unit configures a basic form through a low code platform dragging component, industrial template data is called, component linkage probability is calculated, the adaptive entropy is calculated through a formula by combining the industrial label weight of a treatment template and the configured intelligent verification rule number, and when the output adaptive entropy is larger than a threshold value, the form configuration is valid and is synchronized to a template generation sub-module; the component linkage probability calculation formula is as follows: ; the adaptive entropy calculation formula is as follows: ; Wherein, the The entropy is adapted for the linkage of the components, In order to govern the total number of components in the template, Is a component Assembly Is used for the linkage probability of the (a), To govern the industry label weight of the template, In order to check the rule redundancy penalty coefficients, Is a component Is used for checking the number of rules of the intelligent verification system, The number of rules is checked for industry standards.
8. The multi-terminal adaptive social co-treatment information autonomous reporting system based on the micro-service architecture as set forth in claim 3, wherein the automatic acquisition and data fusion sub-module receives data and manual data acquired by an IoT device twinning model, invokes the accuracy of the device twinning model as IoT device data weight, queries user historical reliability as manual data weight, calculates fusion confidence through a formula, if the fusion confidence is greater than a threshold, the fusion data is valid, pushes the fusion data to a data preprocessing and hash generation module, if the fusion confidence is less than the threshold, triggers exception handling, namely pushes a prompt of low data reliability to a multi-terminal access layer, and guides a user to check the manual data or calibrate the IoT device; the fusion confidence coefficient calculation formula is as follows: ; Wherein, the For the IoT device data weight, For the weight of the artificial data, As an IoT device data average value, Is the mean value of the artificial data, The number of samples of data is collected for the IoT device, For the number of fields of the artificial data, Is the first The data samples are collected by the individual IoT, For the data sample mean of the IoT acquisition, As the standard deviation of the IoT data samples, Is the first The value of the personal data field, As the mean value of the field of the artificial data, Is the standard deviation of the artificial data field.
9. The multi-terminal adaptive social co-therapy information autonomous reporting system based on the micro-service architecture as claimed in claim 4, wherein: after the intelligent rule checking sub-module receives the standardized data set, extracting data to be checked, calling an intelligent checking rule to acquire a data average value of the same field in the same industry, a standard deviation of the same field in the same industry and a standard checking rule number of the same field in the same industry, combining a management template to calculate abnormal probability according to a formula aiming at a target checking rule number of the field and a certificate storing triggering weight corresponding to a certificate storing triggering condition; the abnormal probability calculation formula is as follows: ; Wherein, the For the data to be verified, the data, Is the data average value of the same industry and field, Is the standard deviation of the data in the same field of the same industry, To govern the number of target check rules for that field in the template, For the standard check rule number of this field in the same industry, In order to store the trigger weight for the certificate, For the anomaly penalty coefficient, To indicate a function.
10. The multi-terminal adaptive social co-treatment information autonomous reporting system based on the micro-service architecture of claim 5, wherein the federation learning module calculates local model gradients through legal treatment data based on binding certificate hashes of nodes in each region, obtains gradient encryption results through homomorphic encryption, calculates average credit, data contribution degree and hash consistency of the nodes, obtains global gradients through a federation learning gradient aggregation formula, and is used for updating a global treatment model; The federal learning gradient aggregation formula is as follows: ; Wherein, the As a global gradient of the gradient, The number of regional nodes that are federally learned, Is the first Average credit score for social units within individual regional nodes, Is the first The degree of data contribution of the individual region nodes, Is the first The local model gradient of the individual region nodes, For the purpose of the hash consistency weight, Is the first Average forensic hash consistency for individual regional nodes, Hash consistency is verified for the average of all nodes.

Description

Multi-terminal adaptive social co-treatment information autonomous reporting system based on micro-service architecture Technical Field The invention relates to the technical field of autonomous reporting of social co-treatment information, in particular to a multi-terminal adaptive autonomous reporting system of social co-treatment information based on a micro-service architecture. Background The social co-treatment information autonomous reporting system is a comprehensive information interaction platform which is built for the modern requirements of social treatment and integrates multi-element main body participation, full-flow digitization and intelligent co-treatment. The system takes breaking the information barriers among the treatment subjects as the core, focuses on core scenes such as public business management, civil service optimization, risk hidden danger prevention and control and the like, and provides a convenient, safe and efficient information self-main reporting and collaborative disposal solution for government departments, enterprise units, social organizations and the public. The social units in the existing autonomous report system of social co-treatment information need to rely on technicians to customize report forms, the operation is complex, the components are linked to conflict logically and check rule redundancy, the template configuration efficiency is low, the manually filled data is easy to falsify, the data reliability is low, intelligent analysis is lacking after data acquisition, abnormal data identification relies on manual work, the cross-domain treatment model training is not combined with the credit status of the units, and the prediction precision is low. Disclosure of Invention The multi-terminal adaptive social co-treatment information autonomous reporting system based on the micro-service architecture provided by the invention aims to solve the technical problems that social units in the social co-treatment information autonomous reporting system need to customize report forms by technicians, the operation is complex, the component linkage logic conflict and the verification rule redundancy are low in template configuration efficiency, the manually filled data is easy to falsify, the data reliability is low, intelligent analysis is lacking after data acquisition, abnormal data identification depends on manual work, the unit credit status is not combined in cross-domain treatment model training, and the prediction precision is low. In order to achieve the purpose, the invention adopts the following technical scheme that the multi-terminal adaptive social co-treatment information autonomous reporting system based on the micro-service architecture comprises: the equipment twinning binding module is used for accessing the IoT equipment, generating an equipment twinning model and a data mapping rule, configuring a blockchain certificate triggering condition and generating a governance template; The multi-terminal data acquisition and transmission module is used for automatically acquiring treatment data according to the preset frequency of the equipment twin model and transmitting the treatment data to a system background through an encryption channel; The data preprocessing and hash generation module is used for preprocessing the treatment data, carrying out validity verification according to the intelligent verification rule of the treatment template, pushing the legal treatment data to the blockchain evidence storage micro-service according to the evidence storage triggering condition, and generating an evidence storage hash; The federal learning module is used for loading the local binding certificate hash treatment data, screening similar treatment data based on the industry label of the treatment template, initializing the local treatment model, carrying out gradient encryption transmission on the federal learning, polymerizing to generate a global treatment model, carrying out risk assessment on the treatment data, and generating a hierarchical early warning; and the data reflux module is used for calibrating the equipment twin model according to the hierarchical early warning, re-reporting the treatment data and triggering the secondary evidence storage to generate a new evidence storage hash, wherein the data is refluxed to the federal learning node, and the federal learning gradient calculation weight is optimized by combining the unit credit, so as to update the global treatment model. Preferably, the device twinning binding module comprises: the form component configuration submodule is used for a social unit to drag the basic component through the low-code platform and set intelligent verification rules in combination with industry requirements; the device binding submodule is used for calling a compatibility detection interface of the access micro-service of the IoT device and verifying a device communication protocol; The twin-building mold module is use