Intelligent distinguishing method for sensitive information of micro-map text content

Abstract

The invention discloses an intelligent distinguishing method for sensitive information of micro-map text content, relates to the field of map cartography and geographic information engineering of surveying and mapping science and technology, and aims to solve the problems of intelligent identification and compliance auditing of the sensitive information in the micro-map text content. The method comprises the following steps: collecting map works, and extracting map text content; constructing a micro-map text content sensitive data set according to relevant national policies and regulations and industrial standards; building a micro map text content sensitive information discrimination model based on deep learning, and training on the data set to obtain a model with optimal performance; and building an online service platform, deploying the optimal model on the platform, and carrying out sensitive information judgment and auditing on the text content of the micro map in real time. The method effectively solves the problem of compliance of sensitive information in the micro-map text content, can remarkably reduce the workload of manual auditing, improves the auditing efficiency, and ensures that the micro-map text content meets the requirements of relevant laws and regulations and policies such as privacy protection and safety supervision.

Description

Technical Field

[0001] The present invention relates to the field of cartography and geographic information engineering of surveying and mapping science and technology, and in particular to a method for intelligently distinguishing sensitive information of text content in a micromap. Background Art

[0002] As a special form of map product, the production and use of We-Map must comply with national map management laws and regulations. Since We-Map is dominated by the general public, its user group covers a wide range of non-professional cartographers, including amateurs and ordinary users. This diverse user background leads to uneven map quality, which may result in "problematic We-Maps", such as the marking of sensitive information or the leakage of confidential information. The marking or leakage of such sensitive information may not only cause ethical and legal issues, but also threaten social stability and national security. In addition, We-Map has a strong self-media communication characteristic, and its communication method needs to be efficient, flexible and convenient to meet the personalized needs of users. Therefore, before the release of We-Map, potential problems must be identified and resolved through an effective review mechanism to avoid any negative impact on society or users.

[0003] National invention patent CN202311125148.3 discloses a micro-map review and discrimination method based on deep learning, which aims to quickly identify micro-maps that do not need to be reviewed through an intelligent screening mechanism so that they can be efficiently disseminated, and avoid the generation of "problematic micro-maps" during the dissemination of micro-maps. Although this invention can accurately determine whether a micro-map needs to be reviewed, it does not consider whether the content on the map, especially the text content of the micro-map, complies with national laws and regulations on map management.

[0004] National invention patent CN202110139298.4 discloses a KDTree-based image database data processing method, which uses map annotation location information in the database to solve the sensitive information detection of map annotation content. However, the technical problem of how to intelligently identify sensitive information in text information on micro-maps drawn by users remains unsolved.

[0005] In summary, whether it is the public micro-map review and identification method or the sensitivity detection of the annotated location information in the database, they have not identified the sensitive information of the micro-map text content drawn by the user. Therefore, how to provide an intelligent method for identifying sensitive information in the micro-map text content is an urgent problem that technicians in this field need to solve. Summary of the invention

[0006] In view of this, the present invention provides a method for intelligently distinguishing sensitive information in micro-map text content. On the basis of fully considering national map management laws and regulations and industry standards, based on deep learning text classification technology, it solves the technical gap in the review and identification of sensitive information in micro-map text content, significantly improves the automation and efficiency of micro-map text content review, ensures compliance while reducing the burden of manual review.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] In the first aspect, the method for intelligently distinguishing sensitive information in the text content of a micromap applied for in this application comprises the following steps:

[0009] S1: Collect map works and extract map text content;

[0010] The collection of map works mentioned above includes manual collection;

[0011] The map text content includes but is not limited to the names of geographical elements marked on the map, annotation text, legend description, unit information, and other textual or numerical information elements;

[0012] S2: Construct a micro-map text content sensitive data set based on relevant national policies, regulations and industry standards;

[0013] S3: Build a micro-map text content sensitive information discrimination model based on deep learning, train it on the above dataset, and obtain the model with the best performance;

[0014] S4: Build an online service platform, deploy the optimal model on the platform, and identify and review sensitive information in the micro-map text content in real time.

[0015] Based on the first aspect, in the above step S2, a micromap text content sensitive data set is constructed according to relevant national policies, regulations and industry standards.

[0016] The relevant national policies, regulations and industry standards include but are not limited to the "Map Management Regulations", "Basic Requirements for Safe Processing Technology of Navigation Electronic Maps", "Publishing Management Measures", "Supplementary Provisions on the Representation of Public Map Content (Trial)", "Surveying and Mapping Results Management Regulations", "Surveying and Mapping Law" and "Map Review Management Regulations".

[0017] The sensitive data in the micro-map text content includes, but is not limited to, confidential military facilities, national security confidentiality agency information; information on special groups of people, dangerous goods, national economic lifeline facilities, and large-scale material storage; information involving racial, gender, and regional discrimination and illegal information on the Internet.

[0018] Furthermore, the above step S3 includes the following steps:

[0019] S3-1. Divide the micromap text content sensitive dataset into a training set, a validation set, and a test set;

[0020] S3-2. Use the training set to train the sensitive information recognition model based on the BERT model;

[0021] S3-3. By tuning the hyperparameters of the trained BERT model, we obtain the model with the best performance.

[0022] It can be seen from the above technical solution that, compared with the prior art, the present invention discloses a method for intelligently distinguishing sensitive information in micro-map text content. The method is easy to implement, and can accurately distinguish sensitive information in micro-map text content by fully learning relevant national policies, regulations and industry standards. The optimal model for distinguishing sensitive information in micro-map text is deployed using an online service platform, which can realize efficient and rapid dissemination of micro-maps. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0024] Figure 1 A flow chart of intelligent identification of sensitive information in the text content of a micro-map provided by the present invention;

[0025] Figure 2 A specific flow chart of the optimal model for identifying sensitive information in optimal micro-map text provided by the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] See also Figure 1 The embodiment of the present invention discloses a method for intelligently distinguishing sensitive information in micro-map text content, comprising the following steps:

[0028] S1: Collect map works and extract map text content;

[0029] The collection of map works in step S1 is to obtain all map works from the standard map service website, and also to search Baidu images by keywords, screen relevant map works, and collect user-made micro-map works. The OCR (optical character recognition) technology is used to extract the text content of the collected map images, including but not limited to the names of geographical elements marked on the map, annotation text, legend description, unit information, and other textual or digital information elements.

[0030] S2: Construct a micro-map text content sensitive data set based on relevant national policies, regulations and industry standards;

[0031] In step S2, according to the relevant national policies and regulations and industry standards such as the "Map Management Regulations", "Basic Requirements for Navigation Electronic Map Security Processing Technology", "Publishing Management Measures", "Supplementary Provisions on Public Map Content Representation (Trial)", "Surveying and Mapping Achievement Management Regulations", "Surveying and Mapping Law" and "Map Review Management Regulations", a sensitive information dataset of the micro-map text content is constructed. This dataset includes but is not limited to confidential military facilities and national security confidentiality agency information; information on special groups, dangerous goods, national economic lifeline facilities, and large-scale material storage; information involving racial, gender, and regional discrimination and illegal Internet information. Through manual annotation, the dataset is divided into two categories: sensitive text and normal text, and a sensitive dataset of micro-map text content is constructed.

[0032] S3: Build a micro-map text content sensitive information discrimination model based on deep learning, train it on the above dataset, and obtain the model with the best performance;

[0033] like Figure 2 As shown, step S3 includes the following steps:

[0034] S3-1. Divide the micromap text content sensitive dataset into a training set, a validation set, and a test set;

[0035] For example, the micromap text content sensitive dataset is divided into a training set, a validation set, and a test set in a ratio of 8:1:1. The training set is used for model training, the validation set is used to tune the model parameters, and the test set is used to evaluate the actual effect and accuracy of the model.

[0036] S3-2. Use the training set to train the sensitive information recognition model based on the BERT model;

[0037] For demonstration, we use the BERT model to train the training set. BERT is a powerful natural language processing model. Through training, the model can automatically extract the features of sensitive information from the micromap text to provide support for subsequent sensitive information identification.

[0038] S3-3. By tuning the hyperparameters of the trained BERT model, we obtain the model with the best performance.

[0039] Demonstration, based on the BERT model training, we tuned the model's hyperparameters to further improve the model's discrimination accuracy. The trained model was evaluated using the validation set, and adjustments were made based on the evaluation results to ultimately obtain the model with the best performance.

[0040] S4: Build an online service platform, deploy the optimal model on the platform, and identify and review sensitive information in the micro-map text content in real time.

[0041] In step S4, a web online platform was built using Vue.js and Django frameworks, and the optimal model was deployed on the platform. Users can upload micro-map files through the browser for real-time sensitive information identification. The platform calls the deep learning model for review and displays the identification results in real time to ensure that the micro-map complies with relevant regulatory requirements.

[0042] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0043] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined by the scheme 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 in the scheme, but will conform to the widest scope consistent with the principles and novel features disclosed in the present invention.

Claims

1. A method for intelligently identifying sensitive information in micro-map text content, characterized in that: The aim is to solve the problem of intelligent identification and compliance review of sensitive information in micro-map text content, significantly improve the automation and efficiency of micro-map text content review, ensure compliance while reducing the burden of manual review, including the following steps: S1: Collect map works and extract map text content; The collection of map works mentioned above includes manual collection; The map text content includes but is not limited to the names of geographical elements marked on the map, annotation text, legend description, unit information, and other textual or numerical information elements; S2: Construct a micro-map text content sensitive data set based on relevant national policies, regulations and industry standards; S3: Build a micro-map text content sensitive information discrimination model based on deep learning, train it on the above dataset, and obtain the model with the best performance; S4: Build an online service platform, deploy the optimal model on the platform, and identify and review sensitive information in the micro-map text content in real time.

2. According to claim 1, a method for intelligently distinguishing sensitive information in micro-map text content is characterized in that: In step S2, it is also characterized in that: The relevant national policies, regulations and industry standards include but are not limited to the Regulations on Map Management, Basic Technical Requirements for the Safe Processing of Navigation Electronic Maps, Regulations on Publishing Management, Supplementary Provisions on the Representation of Public Map Content (Trial Implementation), Regulations on the Management of Surveying and Mapping Results, Surveying and Mapping Law and Regulations on the Management of Map Review; The sensitive data in the micro-map text content includes, but is not limited to, confidential military facilities, national security confidentiality agency information; information on special groups of people, dangerous goods, national economic lifeline facilities, and large-scale material storage; information involving racial, gender, and regional discrimination and illegal information on the Internet.

3. According to claim 1, the method for intelligently distinguishing sensitive information in micro-map text content is characterized in that: Step S3 includes the following steps: S3-1. Divide the micromap text content sensitive dataset into a training set, a validation set, and a test set; S3-2. Use the training set to train the sensitive information recognition model based on the BERT model; S3-3. Perform hyperparameter tuning on the trained BERT model to obtain the model with the best performance.

Images